更新autogptadd

autogpt/__init__.py

@@ -0,0 +1,14 @@
+import os
+import random
+import sys
+
+from dotenv import load_dotenv
+
+if "pytest" in sys.argv or "pytest" in sys.modules or os.getenv("CI"):
+    print("Setting random seed to 42")
+    random.seed(42)
+
+# Load the users .env file into environment variables
+load_dotenv(verbose=True, override=True)
+
+del load_dotenv

autogpt/__main__.py

@@ -0,0 +1,5 @@
+"""Auto-GPT: A GPT powered AI Assistant"""
+import autogpt.cli
+
+if __name__ == "__main__":
+    autogpt.cli.main()

autogpt/agent/__init__.py

@@ -0,0 +1,4 @@
+from autogpt.agent.agent import Agent
+from autogpt.agent.agent_manager import AgentManager
+
+__all__ = ["Agent", "AgentManager"]

autogpt/agent/agent_manager.py

@@ -0,0 +1,141 @@
+"""Agent manager for managing GPT agents"""
+from __future__ import annotations
+
+from autogpt.config import Config
+from autogpt.llm.base import ChatSequence
+from autogpt.llm.chat import Message, create_chat_completion
+from autogpt.singleton import Singleton
+
+
+class AgentManager(metaclass=Singleton):
+    """Agent manager for managing GPT agents"""
+
+    def __init__(self):
+        self.next_key = 0
+        self.agents: dict[
+            int, tuple[str, list[Message], str]
+        ] = {}  # key, (task, full_message_history, model)
+        self.cfg = Config()
+
+    # Create new GPT agent
+    # TODO: Centralise use of create_chat_completion() to globally enforce token limit
+
+    def create_agent(
+        self, task: str, creation_prompt: str, model: str
+    ) -> tuple[int, str]:
+        """Create a new agent and return its key
+
+        Args:
+            task: The task to perform
+            creation_prompt: Prompt passed to the LLM at creation
+            model: The model to use to run this agent
+
+        Returns:
+            The key of the new agent
+        """
+        messages = ChatSequence.for_model(model, [Message("user", creation_prompt)])
+
+        for plugin in self.cfg.plugins:
+            if not plugin.can_handle_pre_instruction():
+                continue
+            if plugin_messages := plugin.pre_instruction(messages.raw()):
+                messages.extend([Message(**raw_msg) for raw_msg in plugin_messages])
+        # Start GPT instance
+        agent_reply = create_chat_completion(prompt=messages)
+
+        messages.add("assistant", agent_reply)
+
+        plugins_reply = ""
+        for i, plugin in enumerate(self.cfg.plugins):
+            if not plugin.can_handle_on_instruction():
+                continue
+            if plugin_result := plugin.on_instruction([m.raw() for m in messages]):
+                sep = "\n" if i else ""
+                plugins_reply = f"{plugins_reply}{sep}{plugin_result}"
+
+        if plugins_reply and plugins_reply != "":
+            messages.add("assistant", plugins_reply)
+        key = self.next_key
+        # This is done instead of len(agents) to make keys unique even if agents
+        # are deleted
+        self.next_key += 1
+
+        self.agents[key] = (task, list(messages), model)
+
+        for plugin in self.cfg.plugins:
+            if not plugin.can_handle_post_instruction():
+                continue
+            agent_reply = plugin.post_instruction(agent_reply)
+
+        return key, agent_reply
+
+    def message_agent(self, key: str | int, message: str) -> str:
+        """Send a message to an agent and return its response
+
+        Args:
+            key: The key of the agent to message
+            message: The message to send to the agent
+
+        Returns:
+            The agent's response
+        """
+        task, messages, model = self.agents[int(key)]
+
+        # Add user message to message history before sending to agent
+        messages = ChatSequence.for_model(model, messages)
+        messages.add("user", message)
+
+        for plugin in self.cfg.plugins:
+            if not plugin.can_handle_pre_instruction():
+                continue
+            if plugin_messages := plugin.pre_instruction([m.raw() for m in messages]):
+                messages.extend([Message(**raw_msg) for raw_msg in plugin_messages])
+
+        # Start GPT instance
+        agent_reply = create_chat_completion(prompt=messages)
+
+        messages.add("assistant", agent_reply)
+
+        plugins_reply = agent_reply
+        for i, plugin in enumerate(self.cfg.plugins):
+            if not plugin.can_handle_on_instruction():
+                continue
+            if plugin_result := plugin.on_instruction([m.raw() for m in messages]):
+                sep = "\n" if i else ""
+                plugins_reply = f"{plugins_reply}{sep}{plugin_result}"
+        # Update full message history
+        if plugins_reply and plugins_reply != "":
+            messages.add("assistant", plugins_reply)
+
+        for plugin in self.cfg.plugins:
+            if not plugin.can_handle_post_instruction():
+                continue
+            agent_reply = plugin.post_instruction(agent_reply)
+
+        return agent_reply
+
+    def list_agents(self) -> list[tuple[str | int, str]]:
+        """Return a list of all agents
+
+        Returns:
+            A list of tuples of the form (key, task)
+        """
+
+        # Return a list of agent keys and their tasks
+        return [(key, task) for key, (task, _, _) in self.agents.items()]
+
+    def delete_agent(self, key: str | int) -> bool:
+        """Delete an agent from the agent manager
+
+        Args:
+            key: The key of the agent to delete
+
+        Returns:
+            True if successful, False otherwise
+        """
+
+        try:
+            del self.agents[int(key)]
+            return True
+        except KeyError:
+            return False

autogpt/app.py

@@ -0,0 +1,242 @@
+""" Command and Control """
+import json
+from typing import Dict, List, Union
+
+from autogpt.agent.agent_manager import AgentManager
+from autogpt.commands.command import CommandRegistry, command
+from autogpt.commands.web_requests import scrape_links, scrape_text
+from autogpt.config import Config
+from autogpt.processing.text import summarize_text
+from autogpt.prompts.generator import PromptGenerator
+from autogpt.speech import say_text
+from autogpt.url_utils.validators import validate_url
+
+
+def is_valid_int(value: str) -> bool:
+    """Check if the value is a valid integer
+
+    Args:
+        value (str): The value to check
+
+    Returns:
+        bool: True if the value is a valid integer, False otherwise
+    """
+    try:
+        int(value)
+        return True
+    except ValueError:
+        return False
+
+
+def get_command(response_json: Dict):
+    """Parse the response and return the command name and arguments
+
+    Args:
+        response_json (json): The response from the AI
+
+    Returns:
+        tuple: The command name and arguments
+
+    Raises:
+        json.decoder.JSONDecodeError: If the response is not valid JSON
+
+        Exception: If any other error occurs
+    """
+    try:
+        if "command" not in response_json:
+            return "Error:", "Missing 'command' object in JSON"
+
+        if not isinstance(response_json, dict):
+            return "Error:", f"'response_json' object is not dictionary {response_json}"
+
+        command = response_json["command"]
+        if not isinstance(command, dict):
+            return "Error:", "'command' object is not a dictionary"
+
+        if "name" not in command:
+            return "Error:", "Missing 'name' field in 'command' object"
+
+        command_name = command["name"]
+
+        # Use an empty dictionary if 'args' field is not present in 'command' object
+        arguments = command.get("args", {})
+
+        return command_name, arguments
+    except json.decoder.JSONDecodeError:
+        return "Error:", "Invalid JSON"
+    # All other errors, return "Error: + error message"
+    except Exception as e:
+        return "Error:", str(e)
+
+
+def map_command_synonyms(command_name: str):
+    """Takes the original command name given by the AI, and checks if the
+    string matches a list of common/known hallucinations
+    """
+    synonyms = [
+        ("write_file", "write_to_file"),
+        ("create_file", "write_to_file"),
+        ("search", "google"),
+    ]
+    for seen_command, actual_command_name in synonyms:
+        if command_name == seen_command:
+            return actual_command_name
+    return command_name
+
+
+def execute_command(
+    command_registry: CommandRegistry,
+    command_name: str,
+    arguments,
+    prompt: PromptGenerator,
+    config: Config,
+):
+    """Execute the command and return the result
+
+    Args:
+        command_name (str): The name of the command to execute
+        arguments (dict): The arguments for the command
+
+    Returns:
+        str: The result of the command
+    """
+    try:
+        cmd = command_registry.commands.get(command_name)
+
+        # If the command is found, call it with the provided arguments
+        if cmd:
+            return cmd(**arguments, config=config)
+
+        # TODO: Remove commands below after they are moved to the command registry.
+        command_name = map_command_synonyms(command_name.lower())
+
+        # TODO: Change these to take in a file rather than pasted code, if
+        # non-file is given, return instructions "Input should be a python
+        # filepath, write your code to file and try again
+        for command in prompt.commands:
+            if (
+                command_name == command["label"].lower()
+                or command_name == command["name"].lower()
+            ):
+                return command["function"](**arguments)
+        return (
+            f"Unknown command '{command_name}'. Please refer to the 'COMMANDS'"
+            " list for available commands and only respond in the specified JSON"
+            " format."
+        )
+    except Exception as e:
+        return f"Error: {str(e)}"
+
+
+@command(
+    "get_text_summary", "Get text summary", '"url": "<url>", "question": "<question>"'
+)
+@validate_url
+def get_text_summary(url: str, question: str, config: Config) -> str:
+    """Get the text summary of a webpage
+
+    Args:
+        url (str): The url to scrape
+        question (str): The question to summarize the text for
+
+    Returns:
+        str: The summary of the text
+    """
+    text = scrape_text(url)
+    summary, _ = summarize_text(text, question=question)
+
+    return f""" "Result" : {summary}"""
+
+
+@command("get_hyperlinks", "Get hyperlinks", '"url": "<url>"')
+@validate_url
+def get_hyperlinks(url: str, config: Config) -> Union[str, List[str]]:
+    """Get all hyperlinks on a webpage
+
+    Args:
+        url (str): The url to scrape
+
+    Returns:
+        str or list: The hyperlinks on the page
+    """
+    return scrape_links(url, config)
+
+
+@command(
+    "start_agent",
+    "Start GPT Agent",
+    '"name": "<name>", "task": "<short_task_desc>", "prompt": "<prompt>"',
+)
+def start_agent(name: str, task: str, prompt: str, config: Config, model=None) -> str:
+    """Start an agent with a given name, task, and prompt
+
+    Args:
+        name (str): The name of the agent
+        task (str): The task of the agent
+        prompt (str): The prompt for the agent
+        model (str): The model to use for the agent
+
+    Returns:
+        str: The response of the agent
+    """
+    agent_manager = AgentManager()
+
+    # Remove underscores from name
+    voice_name = name.replace("_", " ")
+
+    first_message = f"""You are {name}.  Respond with: "Acknowledged"."""
+    agent_intro = f"{voice_name} here, Reporting for duty!"
+
+    # Create agent
+    if config.speak_mode:
+        say_text(agent_intro, 1)
+    key, ack = agent_manager.create_agent(task, first_message, model)
+
+    if config.speak_mode:
+        say_text(f"Hello {voice_name}. Your task is as follows. {task}.")
+
+    # Assign task (prompt), get response
+    agent_response = agent_manager.message_agent(key, prompt)
+
+    return f"Agent {name} created with key {key}. First response: {agent_response}"
+
+
+@command("message_agent", "Message GPT Agent", '"key": "<key>", "message": "<message>"')
+def message_agent(key: str, message: str, config: Config) -> str:
+    """Message an agent with a given key and message"""
+    # Check if the key is a valid integer
+    if is_valid_int(key):
+        agent_response = AgentManager().message_agent(int(key), message)
+    else:
+        return "Invalid key, must be an integer."
+
+    # Speak response
+    if config.speak_mode:
+        say_text(agent_response, 1)
+    return agent_response
+
+
+@command("list_agents", "List GPT Agents", "() -> str")
+def list_agents(config: Config) -> str:
+    """List all agents
+
+    Returns:
+        str: A list of all agents
+    """
+    return "List of agents:\n" + "\n".join(
+        [str(x[0]) + ": " + x[1] for x in AgentManager().list_agents()]
+    )
+
+
+@command("delete_agent", "Delete GPT Agent", '"key": "<key>"')
+def delete_agent(key: str, config: Config) -> str:
+    """Delete an agent with a given key
+
+    Args:
+        key (str): The key of the agent to delete
+
+    Returns:
+        str: A message indicating whether the agent was deleted or not
+    """
+    result = AgentManager().delete_agent(key)
+    return f"Agent {key} deleted." if result else f"Agent {key} does not exist."

autogpt/cli.py

@@ -0,0 +1,116 @@
+"""Main script for the autogpt package."""
+import click
+
+
+@click.group(invoke_without_command=True)
+@click.option("-c", "--continuous", is_flag=True, help="Enable Continuous Mode")
+@click.option(
+    "--skip-reprompt",
+    "-y",
+    is_flag=True,
+    help="Skips the re-prompting messages at the beginning of the script",
+)
+@click.option(
+    "--ai-settings",
+    "-C",
+    help="Specifies which ai_settings.yaml file to use, will also automatically skip the re-prompt.",
+)
+@click.option(
+    "--prompt-settings",
+    "-P",
+    help="Specifies which prompt_settings.yaml file to use.",
+)
+@click.option(
+    "-l",
+    "--continuous-limit",
+    type=int,
+    help="Defines the number of times to run in continuous mode",
+)
+@click.option("--speak", is_flag=True, help="Enable Speak Mode")
+@click.option("--debug", is_flag=True, help="Enable Debug Mode")
+@click.option("--gpt3only", is_flag=True, help="Enable GPT3.5 Only Mode")
+@click.option("--gpt4only", is_flag=True, help="Enable GPT4 Only Mode")
+@click.option(
+    "--use-memory",
+    "-m",
+    "memory_type",
+    type=str,
+    help="Defines which Memory backend to use",
+)
+@click.option(
+    "-b",
+    "--browser-name",
+    help="Specifies which web-browser to use when using selenium to scrape the web.",
+)
+@click.option(
+    "--allow-downloads",
+    is_flag=True,
+    help="Dangerous: Allows Auto-GPT to download files natively.",
+)
+@click.option(
+    "--skip-news",
+    is_flag=True,
+    help="Specifies whether to suppress the output of latest news on startup.",
+)
+@click.option(
+    # TODO: this is a hidden option for now, necessary for integration testing.
+    #   We should make this public once we're ready to roll out agent specific workspaces.
+    "--workspace-directory",
+    "-w",
+    type=click.Path(),
+    hidden=True,
+)
+@click.option(
+    "--install-plugin-deps",
+    is_flag=True,
+    help="Installs external dependencies for 3rd party plugins.",
+)
+@click.pass_context
+def main(
+    ctx: click.Context,
+    continuous: bool,
+    continuous_limit: int,
+    ai_settings: str,
+    prompt_settings: str,
+    skip_reprompt: bool,
+    speak: bool,
+    debug: bool,
+    gpt3only: bool,
+    gpt4only: bool,
+    memory_type: str,
+    browser_name: str,
+    allow_downloads: bool,
+    skip_news: bool,
+    workspace_directory: str,
+    install_plugin_deps: bool,
+) -> None:
+    """
+    Welcome to AutoGPT an experimental open-source application showcasing the capabilities of the GPT-4 pushing the boundaries of AI.
+
+    Start an Auto-GPT assistant.
+    """
+    # Put imports inside function to avoid importing everything when starting the CLI
+    from autogpt.main import run_auto_gpt
+
+    if ctx.invoked_subcommand is None:
+        run_auto_gpt(
+            continuous,
+            continuous_limit,
+            ai_settings,
+            prompt_settings,
+            skip_reprompt,
+            speak,
+            debug,
+            gpt3only,
+            gpt4only,
+            memory_type,
+            browser_name,
+            allow_downloads,
+            skip_news,
+            workspace_directory,
+            install_plugin_deps,
+        )
+
+
+if __name__ == "__main__":
+    main()

autogpt/commands/analyze_code.py

@@ -0,0 +1,36 @@
+"""Code evaluation module."""
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+from autogpt.commands.command import command
+from autogpt.llm.utils import call_ai_function
+
+if TYPE_CHECKING:
+    from autogpt.config import Config
+
+
+@command(
+    "analyze_code",
+    "Analyze Code",
+    '"code": "<full_code_string>"',
+)
+def analyze_code(code: str, config: Config) -> list[str]:
+    """
+    A function that takes in a string and returns a response from create chat
+      completion api call.
+
+    Parameters:
+        code (str): Code to be evaluated.
+    Returns:
+        A result string from create chat completion. A list of suggestions to
+            improve the code.
+    """
+
+    function_string = "def analyze_code(code: str) -> list[str]:"
+    args = [code]
+    description_string = (
+        "Analyzes the given code and returns a list of suggestions for improvements."
+    )
+
+    return call_ai_function(function_string, args, description_string, config=config)

autogpt/commands/audio_text.py

@@ -0,0 +1,64 @@
+"""Commands for converting audio to text."""
+import json
+from typing import TYPE_CHECKING
+
+import requests
+
+from autogpt.commands.command import command
+from autogpt.config import Config
+
+if TYPE_CHECKING:
+    from autogpt.config import Config
+
+
+@command(
+    "read_audio_from_file",
+    "Convert Audio to text",
+    '"filename": "<filename>"',
+    lambda config: config.huggingface_audio_to_text_model
+    and config.huggingface_api_token,
+    "Configure huggingface_audio_to_text_model and Hugging Face api token.",
+)
+def read_audio_from_file(filename: str, config: Config) -> str:
+    """
+    Convert audio to text.
+
+    Args:
+        filename (str): The path to the audio file
+
+    Returns:
+        str: The text from the audio
+    """
+    with open(filename, "rb") as audio_file:
+        audio = audio_file.read()
+    return read_audio(audio, config)
+
+
+def read_audio(audio: bytes, config: Config) -> str:
+    """
+    Convert audio to text.
+
+    Args:
+        audio (bytes): The audio to convert
+
+    Returns:
+        str: The text from the audio
+    """
+    model = config.huggingface_audio_to_text_model
+    api_url = f"https://api-inference.huggingface.co/models/{model}"
+    api_token = config.huggingface_api_token
+    headers = {"Authorization": f"Bearer {api_token}"}
+
+    if api_token is None:
+        raise ValueError(
+            "You need to set your Hugging Face API token in the config file."
+        )
+
+    response = requests.post(
+        api_url,
+        headers=headers,
+        data=audio,
+    )
+
+    text = json.loads(response.content.decode("utf-8"))["text"]
+    return f"The audio says: {text}"

autogpt/commands/command.py

@@ -0,0 +1,177 @@
+import functools
+import importlib
+import inspect
+from typing import Any, Callable, Optional
+
+from autogpt.config import Config
+from autogpt.logs import logger
+
+# Unique identifier for auto-gpt commands
+AUTO_GPT_COMMAND_IDENTIFIER = "auto_gpt_command"
+
+
+class Command:
+    """A class representing a command.
+
+    Attributes:
+        name (str): The name of the command.
+        description (str): A brief description of what the command does.
+        signature (str): The signature of the function that the command executes. Defaults to None.
+    """
+
+    def __init__(
+        self,
+        name: str,
+        description: str,
+        method: Callable[..., Any],
+        signature: str = "",
+        enabled: bool | Callable[[Config], bool] = True,
+        disabled_reason: Optional[str] = None,
+    ):
+        self.name = name
+        self.description = description
+        self.method = method
+        self.signature = signature
+        self.enabled = enabled
+        self.disabled_reason = disabled_reason
+
+    def __call__(self, *args, **kwargs) -> Any:
+        if hasattr(kwargs, "config") and callable(self.enabled):
+            self.enabled = self.enabled(kwargs["config"])
+        if not self.enabled:
+            if self.disabled_reason:
+                return f"Command '{self.name}' is disabled: {self.disabled_reason}"
+            return f"Command '{self.name}' is disabled"
+        return self.method(*args, **kwargs)
+
+    def __str__(self) -> str:
+        return f"{self.name}: {self.description}, args: {self.signature}"
+
+
+class CommandRegistry:
+    """
+    The CommandRegistry class is a manager for a collection of Command objects.
+    It allows the registration, modification, and retrieval of Command objects,
+    as well as the scanning and loading of command plugins from a specified
+    directory.
+    """
+
+    def __init__(self):
+        self.commands = {}
+
+    def _import_module(self, module_name: str) -> Any:
+        return importlib.import_module(module_name)
+
+    def _reload_module(self, module: Any) -> Any:
+        return importlib.reload(module)
+
+    def register(self, cmd: Command) -> None:
+        if cmd.name in self.commands:
+            logger.warn(
+                f"Command '{cmd.name}' already registered and will be overwritten!"
+            )
+        self.commands[cmd.name] = cmd
+
+    def unregister(self, command_name: str):
+        if command_name in self.commands:
+            del self.commands[command_name]
+        else:
+            raise KeyError(f"Command '{command_name}' not found in registry.")
+
+    def reload_commands(self) -> None:
+        """Reloads all loaded command plugins."""
+        for cmd_name in self.commands:
+            cmd = self.commands[cmd_name]
+            module = self._import_module(cmd.__module__)
+            reloaded_module = self._reload_module(module)
+            if hasattr(reloaded_module, "register"):
+                reloaded_module.register(self)
+
+    def get_command(self, name: str) -> Callable[..., Any]:
+        return self.commands[name]
+
+    def call(self, command_name: str, **kwargs) -> Any:
+        if command_name not in self.commands:
+            raise KeyError(f"Command '{command_name}' not found in registry.")
+        command = self.commands[command_name]
+        return command(**kwargs)
+
+    def command_prompt(self) -> str:
+        """
+        Returns a string representation of all registered `Command` objects for use in a prompt
+        """
+        commands_list = [
+            f"{idx + 1}. {str(cmd)}" for idx, cmd in enumerate(self.commands.values())
+        ]
+        return "\n".join(commands_list)
+
+    def import_commands(self, module_name: str) -> None:
+        """
+        Imports the specified Python module containing command plugins.
+
+        This method imports the associated module and registers any functions or
+        classes that are decorated with the `AUTO_GPT_COMMAND_IDENTIFIER` attribute
+        as `Command` objects. The registered `Command` objects are then added to the
+        `commands` dictionary of the `CommandRegistry` object.
+
+        Args:
+            module_name (str): The name of the module to import for command plugins.
+        """
+
+        module = importlib.import_module(module_name)
+
+        for attr_name in dir(module):
+            attr = getattr(module, attr_name)
+            # Register decorated functions
+            if hasattr(attr, AUTO_GPT_COMMAND_IDENTIFIER) and getattr(
+                attr, AUTO_GPT_COMMAND_IDENTIFIER
+            ):
+                self.register(attr.command)
+            # Register command classes
+            elif (
+                inspect.isclass(attr) and issubclass(attr, Command) and attr != Command
+            ):
+                cmd_instance = attr()
+                self.register(cmd_instance)
+
+
+def command(
+    name: str,
+    description: str,
+    signature: str,
+    enabled: bool | Callable[[Config], bool] = True,
+    disabled_reason: Optional[str] = None,
+) -> Callable[..., Any]:
+    """The command decorator is used to create Command objects from ordinary functions."""
+
+    # TODO: Remove this in favor of better command management
+    CFG = Config()
+
+    if callable(enabled):
+        enabled = enabled(CFG)
+    if not enabled:
+        if disabled_reason is not None:
+            logger.debug(f"Command '{name}' is disabled: {disabled_reason}")
+        return lambda func: func
+
+    def decorator(func: Callable[..., Any]) -> Command:
+        cmd = Command(
+            name=name,
+            description=description,
+            method=func,
+            signature=signature,
+            enabled=enabled,
+            disabled_reason=disabled_reason,
+        )
+
+        @functools.wraps(func)
+        def wrapper(*args, **kwargs) -> Any:
+            return func(*args, **kwargs)
+
+        wrapper.command = cmd
+
+        setattr(wrapper, AUTO_GPT_COMMAND_IDENTIFIER, True)
+
+        return wrapper
+
+    return decorator

autogpt/commands/execute_code.py

@@ -0,0 +1,214 @@
+"""Execute code in a Docker container"""
+import os
+import subprocess
+from pathlib import Path
+
+import docker
+from docker.errors import ImageNotFound
+
+from autogpt.commands.command import command
+from autogpt.config import Config
+from autogpt.logs import logger
+
+
+@command("execute_python_file", "Execute Python File", '"filename": "<filename>"')
+def execute_python_file(filename: str, config: Config) -> str:
+    """Execute a Python file in a Docker container and return the output
+
+    Args:
+        filename (str): The name of the file to execute
+
+    Returns:
+        str: The output of the file
+    """
+    logger.info(f"Executing file '{filename}'")
+
+    if not filename.endswith(".py"):
+        return "Error: Invalid file type. Only .py files are allowed."
+
+    if not os.path.isfile(filename):
+        return f"Error: File '{filename}' does not exist."
+
+    if we_are_running_in_a_docker_container():
+        result = subprocess.run(
+            ["python", filename], capture_output=True, encoding="utf8"
+        )
+        if result.returncode == 0:
+            return result.stdout
+        else:
+            return f"Error: {result.stderr}"
+
+    try:
+        client = docker.from_env()
+        # You can replace this with the desired Python image/version
+        # You can find available Python images on Docker Hub:
+        # https://hub.docker.com/_/python
+        image_name = "python:3-alpine"
+        try:
+            client.images.get(image_name)
+            logger.warn(f"Image '{image_name}' found locally")
+        except ImageNotFound:
+            logger.info(
+                f"Image '{image_name}' not found locally, pulling from Docker Hub"
+            )
+            # Use the low-level API to stream the pull response
+            low_level_client = docker.APIClient()
+            for line in low_level_client.pull(image_name, stream=True, decode=True):
+                # Print the status and progress, if available
+                status = line.get("status")
+                progress = line.get("progress")
+                if status and progress:
+                    logger.info(f"{status}: {progress}")
+                elif status:
+                    logger.info(status)
+        container = client.containers.run(
+            image_name,
+            ["python", str(Path(filename).relative_to(config.workspace_path))],
+            volumes={
+                config.workspace_path: {
+                    "bind": "/workspace",
+                    "mode": "ro",
+                }
+            },
+            working_dir="/workspace",
+            stderr=True,
+            stdout=True,
+            detach=True,
+        )
+
+        container.wait()
+        logs = container.logs().decode("utf-8")
+        container.remove()
+
+        # print(f"Execution complete. Output: {output}")
+        # print(f"Logs: {logs}")
+
+        return logs
+
+    except docker.errors.DockerException as e:
+        logger.warn(
+            "Could not run the script in a container. If you haven't already, please install Docker https://docs.docker.com/get-docker/"
+        )
+        return f"Error: {str(e)}"
+
+    except Exception as e:
+        return f"Error: {str(e)}"
+
+
+def validate_command(command: str, config: Config) -> bool:
+    """Validate a command to ensure it is allowed
+
+    Args:
+        command (str): The command to validate
+
+    Returns:
+        bool: True if the command is allowed, False otherwise
+    """
+    tokens = command.split()
+
+    if not tokens:
+        return False
+
+    if config.deny_commands and tokens[0] not in config.deny_commands:
+        return False
+
+    for keyword in config.allow_commands:
+        if keyword in tokens:
+            return True
+    if config.allow_commands:
+        return False
+
+    return True
+
+
+@command(
+    "execute_shell",
+    "Execute Shell Command, non-interactive commands only",
+    '"command_line": "<command_line>"',
+    lambda cfg: cfg.execute_local_commands,
+    "You are not allowed to run local shell commands. To execute"
+    " shell commands, EXECUTE_LOCAL_COMMANDS must be set to 'True' "
+    "in your config file: .env - do not attempt to bypass the restriction.",
+)
+def execute_shell(command_line: str, config: Config) -> str:
+    """Execute a shell command and return the output
+
+    Args:
+        command_line (str): The command line to execute
+
+    Returns:
+        str: The output of the command
+    """
+    if not validate_command(command_line, config):
+        logger.info(f"Command '{command_line}' not allowed")
+        return "Error: This Shell Command is not allowed."
+
+    current_dir = Path.cwd()
+    # Change dir into workspace if necessary
+    if not current_dir.is_relative_to(config.workspace_path):
+        os.chdir(config.workspace_path)
+
+    logger.info(
+        f"Executing command '{command_line}' in working directory '{os.getcwd()}'"
+    )
+
+    result = subprocess.run(command_line, capture_output=True, shell=True)
+    output = f"STDOUT:\n{result.stdout}\nSTDERR:\n{result.stderr}"
+
+    # Change back to whatever the prior working dir was
+
+    os.chdir(current_dir)
+    return output
+
+
+@command(
+    "execute_shell_popen",
+    "Execute Shell Command, non-interactive commands only",
+    '"command_line": "<command_line>"',
+    lambda config: config.execute_local_commands,
+    "You are not allowed to run local shell commands. To execute"
+    " shell commands, EXECUTE_LOCAL_COMMANDS must be set to 'True' "
+    "in your config. Do not attempt to bypass the restriction.",
+)
+def execute_shell_popen(command_line, config: Config) -> str:
+    """Execute a shell command with Popen and returns an english description
+    of the event and the process id
+
+    Args:
+        command_line (str): The command line to execute
+
+    Returns:
+        str: Description of the fact that the process started and its id
+    """
+    if not validate_command(command_line, config):
+        logger.info(f"Command '{command_line}' not allowed")
+        return "Error: This Shell Command is not allowed."
+
+    current_dir = os.getcwd()
+    # Change dir into workspace if necessary
+    if config.workspace_path not in current_dir:
+        os.chdir(config.workspace_path)
+
+    logger.info(
+        f"Executing command '{command_line}' in working directory '{os.getcwd()}'"
+    )
+
+    do_not_show_output = subprocess.DEVNULL
+    process = subprocess.Popen(
+        command_line, shell=True, stdout=do_not_show_output, stderr=do_not_show_output
+    )
+
+    # Change back to whatever the prior working dir was
+
+    os.chdir(current_dir)
+
+    return f"Subprocess started with PID:'{str(process.pid)}'"
+
+
+def we_are_running_in_a_docker_container() -> bool:
+    """Check if we are running in a Docker container
+
+    Returns:
+        bool: True if we are running in a Docker container, False otherwise
+    """
+    return os.path.exists("/.dockerenv")

autogpt/commands/file_operations.py

@@ -0,0 +1,348 @@
+"""File operations for AutoGPT"""
+from __future__ import annotations
+
+import hashlib
+import os
+import os.path
+from typing import TYPE_CHECKING, Generator, Literal
+
+import requests
+from colorama import Back, Fore
+from requests.adapters import HTTPAdapter, Retry
+
+from autogpt.commands.command import command
+from autogpt.commands.file_operations_utils import read_textual_file
+from autogpt.logs import logger
+from autogpt.memory.vector import MemoryItem, VectorMemory
+from autogpt.spinner import Spinner
+from autogpt.utils import readable_file_size
+
+if TYPE_CHECKING:
+    from autogpt.config import Config
+
+
+Operation = Literal["write", "append", "delete"]
+
+
+def text_checksum(text: str) -> str:
+    """Get the hex checksum for the given text."""
+    return hashlib.md5(text.encode("utf-8")).hexdigest()
+
+
+def operations_from_log(
+    log_path: str,
+) -> Generator[tuple[Operation, str, str | None], None, None]:
+    """Parse the file operations log and return a tuple containing the log entries"""
+    try:
+        log = open(log_path, "r", encoding="utf-8")
+    except FileNotFoundError:
+        return
+
+    for line in log:
+        line = line.replace("File Operation Logger", "").strip()
+        if not line:
+            continue
+        operation, tail = line.split(": ", maxsplit=1)
+        operation = operation.strip()
+        if operation in ("write", "append"):
+            try:
+                path, checksum = (x.strip() for x in tail.rsplit(" #", maxsplit=1))
+            except ValueError:
+                logger.warn(f"File log entry lacks checksum: '{line}'")
+                path, checksum = tail.strip(), None
+            yield (operation, path, checksum)
+        elif operation == "delete":
+            yield (operation, tail.strip(), None)
+
+    log.close()
+
+
+def file_operations_state(log_path: str) -> dict[str, str]:
+    """Iterates over the operations log and returns the expected state.
+
+    Parses a log file at config.file_logger_path to construct a dictionary that maps
+    each file path written or appended to its checksum. Deleted files are removed
+    from the dictionary.
+
+    Returns:
+        A dictionary mapping file paths to their checksums.
+
+    Raises:
+        FileNotFoundError: If config.file_logger_path is not found.
+        ValueError: If the log file content is not in the expected format.
+    """
+    state = {}
+    for operation, path, checksum in operations_from_log(log_path):
+        if operation in ("write", "append"):
+            state[path] = checksum
+        elif operation == "delete":
+            del state[path]
+    return state
+
+
+def is_duplicate_operation(
+    operation: Operation, filename: str, config: Config, checksum: str | None = None
+) -> bool:
+    """Check if the operation has already been performed
+
+    Args:
+        operation: The operation to check for
+        filename: The name of the file to check for
+        checksum: The checksum of the contents to be written
+
+    Returns:
+        True if the operation has already been performed on the file
+    """
+    state = file_operations_state(config.file_logger_path)
+    if operation == "delete" and filename not in state:
+        return True
+    if operation == "write" and state.get(filename) == checksum:
+        return True
+    return False
+
+
+def log_operation(
+    operation: str, filename: str, config: Config, checksum: str | None = None
+) -> None:
+    """Log the file operation to the file_logger.txt
+
+    Args:
+        operation: The operation to log
+        filename: The name of the file the operation was performed on
+        checksum: The checksum of the contents to be written
+    """
+    log_entry = f"{operation}: {filename}"
+    if checksum is not None:
+        log_entry += f" #{checksum}"
+    logger.debug(f"Logging file operation: {log_entry}")
+    append_to_file(config.file_logger_path, f"{log_entry}\n", config, should_log=False)
+
+
+def split_file(
+    content: str, max_length: int = 4000, overlap: int = 0
+) -> Generator[str, None, None]:
+    """
+    Split text into chunks of a specified maximum length with a specified overlap
+    between chunks.
+
+    :param content: The input text to be split into chunks
+    :param max_length: The maximum length of each chunk,
+        default is 4000 (about 1k token)
+    :param overlap: The number of overlapping characters between chunks,
+        default is no overlap
+    :return: A generator yielding chunks of text
+    """
+    start = 0
+    content_length = len(content)
+
+    while start < content_length:
+        end = start + max_length
+        if end + overlap < content_length:
+            chunk = content[start : end + max(overlap - 1, 0)]
+        else:
+            chunk = content[start:content_length]
+
+            # Account for the case where the last chunk is shorter than the overlap, so it has already been consumed
+            if len(chunk) <= overlap:
+                break
+
+        yield chunk
+        start += max_length - overlap
+
+
+@command("read_file", "Read a file", '"filename": "<filename>"')
+def read_file(filename: str, config: Config) -> str:
+    """Read a file and return the contents
+
+    Args:
+        filename (str): The name of the file to read
+
+    Returns:
+        str: The contents of the file
+    """
+    try:
+        content = read_textual_file(filename, logger)
+
+        # TODO: invalidate/update memory when file is edited
+        file_memory = MemoryItem.from_text_file(content, filename)
+        if len(file_memory.chunks) > 1:
+            return file_memory.summary
+
+        return content
+    except Exception as e:
+        return f"Error: {str(e)}"
+
+
+def ingest_file(
+    filename: str,
+    memory: VectorMemory,
+) -> None:
+    """
+    Ingest a file by reading its content, splitting it into chunks with a specified
+    maximum length and overlap, and adding the chunks to the memory storage.
+
+    Args:
+        filename: The name of the file to ingest
+        memory: An object with an add() method to store the chunks in memory
+    """
+    try:
+        logger.info(f"Ingesting file {filename}")
+        content = read_file(filename)
+
+        # TODO: differentiate between different types of files
+        file_memory = MemoryItem.from_text_file(content, filename)
+        logger.debug(f"Created memory: {file_memory.dump()}")
+        memory.add(file_memory)
+
+        logger.info(f"Ingested {len(file_memory.e_chunks)} chunks from {filename}")
+    except Exception as err:
+        logger.warn(f"Error while ingesting file '{filename}': {err}")
+
+
+@command("write_to_file", "Write to file", '"filename": "<filename>", "text": "<text>"')
+def write_to_file(filename: str, text: str, config: Config) -> str:
+    """Write text to a file
+
+    Args:
+        filename (str): The name of the file to write to
+        text (str): The text to write to the file
+
+    Returns:
+        str: A message indicating success or failure
+    """
+    checksum = text_checksum(text)
+    if is_duplicate_operation("write", filename, config, checksum):
+        return "Error: File has already been updated."
+    try:
+        directory = os.path.dirname(filename)
+        os.makedirs(directory, exist_ok=True)
+        with open(filename, "w", encoding="utf-8") as f:
+            f.write(text)
+        log_operation("write", filename, config, checksum)
+        return "File written to successfully."
+    except Exception as err:
+        return f"Error: {err}"
+
+
+@command(
+    "append_to_file", "Append to file", '"filename": "<filename>", "text": "<text>"'
+)
+def append_to_file(
+    filename: str, text: str, config: Config, should_log: bool = True
+) -> str:
+    """Append text to a file
+
+    Args:
+        filename (str): The name of the file to append to
+        text (str): The text to append to the file
+        should_log (bool): Should log output
+
+    Returns:
+        str: A message indicating success or failure
+    """
+    try:
+        directory = os.path.dirname(filename)
+        os.makedirs(directory, exist_ok=True)
+        with open(filename, "a", encoding="utf-8") as f:
+            f.write(text)
+
+        if should_log:
+            with open(filename, "r", encoding="utf-8") as f:
+                checksum = text_checksum(f.read())
+            log_operation("append", filename, config, checksum=checksum)
+
+        return "Text appended successfully."
+    except Exception as err:
+        return f"Error: {err}"
+
+
+@command("delete_file", "Delete file", '"filename": "<filename>"')
+def delete_file(filename: str, config: Config) -> str:
+    """Delete a file
+
+    Args:
+        filename (str): The name of the file to delete
+
+    Returns:
+        str: A message indicating success or failure
+    """
+    if is_duplicate_operation("delete", filename, config):
+        return "Error: File has already been deleted."
+    try:
+        os.remove(filename)
+        log_operation("delete", filename, config)
+        return "File deleted successfully."
+    except Exception as err:
+        return f"Error: {err}"
+
+
+@command("list_files", "List Files in Directory", '"directory": "<directory>"')
+def list_files(directory: str, config: Config) -> list[str]:
+    """lists files in a directory recursively
+
+    Args:
+        directory (str): The directory to search in
+
+    Returns:
+        list[str]: A list of files found in the directory
+    """
+    found_files = []
+
+    for root, _, files in os.walk(directory):
+        for file in files:
+            if file.startswith("."):
+                continue
+            relative_path = os.path.relpath(
+                os.path.join(root, file), config.workspace_path
+            )
+            found_files.append(relative_path)
+
+    return found_files
+
+
+@command(
+    "download_file",
+    "Download File",
+    '"url": "<url>", "filename": "<filename>"',
+    lambda config: config.allow_downloads,
+    "Error: You do not have user authorization to download files locally.",
+)
+def download_file(url, filename, config: Config):
+    """Downloads a file
+    Args:
+        url (str): URL of the file to download
+        filename (str): Filename to save the file as
+    """
+    try:
+        directory = os.path.dirname(filename)
+        os.makedirs(directory, exist_ok=True)
+        message = f"{Fore.YELLOW}Downloading file from {Back.LIGHTBLUE_EX}{url}{Back.RESET}{Fore.RESET}"
+        with Spinner(message, plain_output=config.plain_output) as spinner:
+            session = requests.Session()
+            retry = Retry(total=3, backoff_factor=1, status_forcelist=[502, 503, 504])
+            adapter = HTTPAdapter(max_retries=retry)
+            session.mount("http://", adapter)
+            session.mount("https://", adapter)
+
+            total_size = 0
+            downloaded_size = 0
+
+            with session.get(url, allow_redirects=True, stream=True) as r:
+                r.raise_for_status()
+                total_size = int(r.headers.get("Content-Length", 0))
+                downloaded_size = 0
+
+                with open(filename, "wb") as f:
+                    for chunk in r.iter_content(chunk_size=8192):
+                        f.write(chunk)
+                        downloaded_size += len(chunk)
+
+                        # Update the progress message
+                        progress = f"{readable_file_size(downloaded_size)} / {readable_file_size(total_size)}"
+                        spinner.update_message(f"{message} {progress}")
+
+            return f'Successfully downloaded and locally stored file: "{filename}"! (Size: {readable_file_size(downloaded_size)})'
+    except requests.HTTPError as err:
+        return f"Got an HTTP Error whilst trying to download file: {err}"
+    except Exception as err:
+        return f"Error: {err}"

autogpt/commands/file_operations_utils.py

@@ -0,0 +1,159 @@
+import json
+import os
+
+import charset_normalizer
+import docx
+import markdown
+import PyPDF2
+import yaml
+from bs4 import BeautifulSoup
+from pylatexenc.latex2text import LatexNodes2Text
+
+from autogpt import logs
+from autogpt.logs import logger
+
+
+class ParserStrategy:
+    def read(self, file_path: str) -> str:
+        raise NotImplementedError
+
+
+# Basic text file reading
+class TXTParser(ParserStrategy):
+    def read(self, file_path: str) -> str:
+        charset_match = charset_normalizer.from_path(file_path).best()
+        logger.debug(f"Reading '{file_path}' with encoding '{charset_match.encoding}'")
+        return str(charset_match)
+
+
+# Reading text from binary file using pdf parser
+class PDFParser(ParserStrategy):
+    def read(self, file_path: str) -> str:
+        parser = PyPDF2.PdfReader(file_path)
+        text = ""
+        for page_idx in range(len(parser.pages)):
+            text += parser.pages[page_idx].extract_text()
+        return text
+
+
+# Reading text from binary file using docs parser
+class DOCXParser(ParserStrategy):
+    def read(self, file_path: str) -> str:
+        doc_file = docx.Document(file_path)
+        text = ""
+        for para in doc_file.paragraphs:
+            text += para.text
+        return text
+
+
+# Reading as dictionary and returning string format
+class JSONParser(ParserStrategy):
+    def read(self, file_path: str) -> str:
+        with open(file_path, "r") as f:
+            data = json.load(f)
+            text = str(data)
+        return text
+
+
+class XMLParser(ParserStrategy):
+    def read(self, file_path: str) -> str:
+        with open(file_path, "r") as f:
+            soup = BeautifulSoup(f, "xml")
+            text = soup.get_text()
+        return text
+
+
+# Reading as dictionary and returning string format
+class YAMLParser(ParserStrategy):
+    def read(self, file_path: str) -> str:
+        with open(file_path, "r") as f:
+            data = yaml.load(f, Loader=yaml.FullLoader)
+            text = str(data)
+        return text
+
+
+class HTMLParser(ParserStrategy):
+    def read(self, file_path: str) -> str:
+        with open(file_path, "r") as f:
+            soup = BeautifulSoup(f, "html.parser")
+            text = soup.get_text()
+        return text
+
+
+class MarkdownParser(ParserStrategy):
+    def read(self, file_path: str) -> str:
+        with open(file_path, "r") as f:
+            html = markdown.markdown(f.read())
+            text = "".join(BeautifulSoup(html, "html.parser").findAll(string=True))
+        return text
+
+
+class LaTeXParser(ParserStrategy):
+    def read(self, file_path: str) -> str:
+        with open(file_path, "r") as f:
+            latex = f.read()
+        text = LatexNodes2Text().latex_to_text(latex)
+        return text
+
+
+class FileContext:
+    def __init__(self, parser: ParserStrategy, logger: logs.Logger):
+        self.parser = parser
+        self.logger = logger
+
+    def set_parser(self, parser: ParserStrategy) -> None:
+        self.logger.debug(f"Setting Context Parser to {parser}")
+        self.parser = parser
+
+    def read_file(self, file_path) -> str:
+        self.logger.debug(f"Reading file {file_path} with parser {self.parser}")
+        return self.parser.read(file_path)
+
+
+extension_to_parser = {
+    ".txt": TXTParser(),
+    ".csv": TXTParser(),
+    ".pdf": PDFParser(),
+    ".docx": DOCXParser(),
+    ".json": JSONParser(),
+    ".xml": XMLParser(),
+    ".yaml": YAMLParser(),
+    ".yml": YAMLParser(),
+    ".html": HTMLParser(),
+    ".htm": HTMLParser(),
+    ".xhtml": HTMLParser(),
+    ".md": MarkdownParser(),
+    ".markdown": MarkdownParser(),
+    ".tex": LaTeXParser(),
+}
+
+
+def is_file_binary_fn(file_path: str):
+    """Given a file path load all its content and checks if the null bytes is present
+
+    Args:
+        file_path (_type_): _description_
+
+    Returns:
+        bool: is_binary
+    """
+    with open(file_path, "rb") as f:
+        file_data = f.read()
+    if b"\x00" in file_data:
+        return True
+    return False
+
+
+def read_textual_file(file_path: str, logger: logs.Logger) -> str:
+    if not os.path.isfile(file_path):
+        raise FileNotFoundError(f"{file_path} not found!")
+    is_binary = is_file_binary_fn(file_path)
+    file_extension = os.path.splitext(file_path)[1].lower()
+    parser = extension_to_parser.get(file_extension)
+    if not parser:
+        if is_binary:
+            raise ValueError(f"Unsupported binary file format: {file_extension}")
+        # fallback to txt file parser (to support script and code files loading)
+        parser = TXTParser()
+    file_context = FileContext(parser, logger)
+    return file_context.read_file(file_path)

autogpt/commands/git_operations.py

@@ -0,0 +1,40 @@
+"""Git operations for autogpt"""
+from typing import TYPE_CHECKING
+
+from git.repo import Repo
+
+from autogpt.commands.command import command
+from autogpt.config import Config
+from autogpt.url_utils.validators import validate_url
+
+if TYPE_CHECKING:
+    from autogpt.config import Config
+
+
+@command(
+    "clone_repository",
+    "Clone Repository",
+    '"url": "<repository_url>", "clone_path": "<clone_path>"',
+    lambda config: config.github_username and config.github_api_key,
+    "Configure github_username and github_api_key.",
+)
+@validate_url
+def clone_repository(url: str, clone_path: str, config: Config) -> str:
+    """Clone a GitHub repository locally.
+
+    Args:
+        url (str): The URL of the repository to clone.
+        clone_path (str): The path to clone the repository to.
+
+    Returns:
+        str: The result of the clone operation.
+    """
+    split_url = url.split("//")
+    auth_repo_url = f"//{config.github_username}:{config.github_api_key}@".join(
+        split_url
+    )
+    try:
+        Repo.clone_from(url=auth_repo_url, to_path=clone_path)
+        return f"""Cloned {url} to {clone_path}"""
+    except Exception as e:
+        return f"Error: {str(e)}"

autogpt/commands/google_search.py

@@ -0,0 +1,132 @@
+"""Google search command for Autogpt."""
+from __future__ import annotations
+
+import json
+from itertools import islice
+from typing import TYPE_CHECKING
+
+from duckduckgo_search import DDGS
+
+from autogpt.commands.command import command
+
+if TYPE_CHECKING:
+    from autogpt.config import Config
+
+
+@command(
+    "google",
+    "Google Search",
+    '"query": "<query>"',
+    lambda config: not config.google_api_key,
+)
+def google_search(query: str, config: Config, num_results: int = 8) -> str:
+    """Return the results of a Google search
+
+    Args:
+        query (str): The search query.
+        num_results (int): The number of results to return.
+
+    Returns:
+        str: The results of the search.
+    """
+    search_results = []
+    if not query:
+        return json.dumps(search_results)
+
+    results = DDGS().text(query)
+    if not results:
+        return json.dumps(search_results)
+
+    for item in islice(results, num_results):
+        search_results.append(item)
+
+    results = json.dumps(search_results, ensure_ascii=False, indent=4)
+    return safe_google_results(results)
+
+
+@command(
+    "google",
+    "Google Search",
+    '"query": "<query>"',
+    lambda config: bool(config.google_api_key) and bool(config.custom_search_engine_id),
+    "Configure google_api_key and custom_search_engine_id.",
+)
+def google_official_search(
+    query: str, config: Config, num_results: int = 8
+) -> str | list[str]:
+    """Return the results of a Google search using the official Google API
+
+    Args:
+        query (str): The search query.
+        num_results (int): The number of results to return.
+
+    Returns:
+        str: The results of the search.
+    """
+
+    from googleapiclient.discovery import build
+    from googleapiclient.errors import HttpError
+
+    try:
+        # Get the Google API key and Custom Search Engine ID from the config file
+        api_key = config.google_api_key
+        custom_search_engine_id = config.custom_search_engine_id
+
+        # Initialize the Custom Search API service
+        service = build("customsearch", "v1", developerKey=api_key)
+
+        # Send the search query and retrieve the results
+        result = (
+            service.cse()
+            .list(q=query, cx=custom_search_engine_id, num=num_results)
+            .execute()
+        )
+
+        # Extract the search result items from the response
+        search_results = result.get("items", [])
+
+        # Create a list of only the URLs from the search results
+        search_results_links = [item["link"] for item in search_results]
+
+    except HttpError as e:
+        # Handle errors in the API call
+        error_details = json.loads(e.content.decode())
+
+        # Check if the error is related to an invalid or missing API key
+        if error_details.get("error", {}).get(
+            "code"
+        ) == 403 and "invalid API key" in error_details.get("error", {}).get(
+            "message", ""
+        ):
+            return "Error: The provided Google API key is invalid or missing."
+        else:
+            return f"Error: {e}"
+    # google_result can be a list or a string depending on the search results
+
+    # Return the list of search result URLs
+    return safe_google_results(search_results_links)
+
+
+def safe_google_results(results: str | list) -> str:
+    """
+        Return the results of a google search in a safe format.
+
+    Args:
+        results (str | list): The search results.
+
+    Returns:
+        str: The results of the search.
+    """
+    if isinstance(results, list):
+        safe_message = json.dumps(
+            [result.encode("utf-8", "ignore").decode("utf-8") for result in results]
+        )
+    else:
+        safe_message = results.encode("utf-8", "ignore").decode("utf-8")
+    return safe_message
+
+
+if __name__ == '__main__':
+    print(google_search('你是谁？'))
+    results = ddg('你是谁', max_results=8)
+    print(results)

autogpt/commands/image_gen.py

@@ -0,0 +1,200 @@
+""" Image Generation Module for AutoGPT."""
+import io
+import json
+import time
+import uuid
+from base64 import b64decode
+from typing import TYPE_CHECKING
+
+import openai
+import requests
+from PIL import Image
+
+from autogpt.commands.command import command
+from autogpt.config import Config
+from autogpt.logs import logger
+
+if TYPE_CHECKING:
+    from autogpt.config import Config
+
+
+@command(
+    "generate_image",
+    "Generate Image",
+    '"prompt": "<prompt>"',
+    lambda config: config.image_provider,
+    "Requires a image provider to be set.",
+)
+def generate_image(prompt: str, config: Config, size: int = 256) -> str:
+    """Generate an image from a prompt.
+
+    Args:
+        prompt (str): The prompt to use
+        size (int, optional): The size of the image. Defaults to 256. (Not supported by HuggingFace)
+
+    Returns:
+        str: The filename of the image
+    """
+    filename = f"{config.workspace_path}/{str(uuid.uuid4())}.jpg"
+
+    # DALL-E
+    if config.image_provider == "dalle":
+        return generate_image_with_dalle(prompt, filename, size, config)
+    # HuggingFace
+    elif config.image_provider == "huggingface":
+        return generate_image_with_hf(prompt, filename, config)
+    # SD WebUI
+    elif config.image_provider == "sdwebui":
+        return generate_image_with_sd_webui(prompt, filename, config, size)
+    return "No Image Provider Set"
+
+
+def generate_image_with_hf(prompt: str, filename: str, config: Config) -> str:
+    """Generate an image with HuggingFace's API.
+
+    Args:
+        prompt (str): The prompt to use
+        filename (str): The filename to save the image to
+
+    Returns:
+        str: The filename of the image
+    """
+    API_URL = (
+        f"https://api-inference.huggingface.co/models/{config.huggingface_image_model}"
+    )
+    if config.huggingface_api_token is None:
+        raise ValueError(
+            "You need to set your Hugging Face API token in the config file."
+        )
+    headers = {
+        "Authorization": f"Bearer {config.huggingface_api_token}",
+        "X-Use-Cache": "false",
+    }
+
+    retry_count = 0
+    while retry_count < 10:
+        response = requests.post(
+            API_URL,
+            headers=headers,
+            json={
+                "inputs": prompt,
+            },
+        )
+
+        if response.ok:
+            try:
+                image = Image.open(io.BytesIO(response.content))
+                logger.info(f"Image Generated for prompt:{prompt}")
+                image.save(filename)
+                return f"Saved to disk:{filename}"
+            except Exception as e:
+                logger.error(e)
+                break
+        else:
+            try:
+                error = json.loads(response.text)
+                if "estimated_time" in error:
+                    delay = error["estimated_time"]
+                    logger.debug(response.text)
+                    logger.info("Retrying in", delay)
+                    time.sleep(delay)
+                else:
+                    break
+            except Exception as e:
+                logger.error(e)
+                break
+
+        retry_count += 1
+
+    return f"Error creating image."
+
+
+def generate_image_with_dalle(
+    prompt: str, filename: str, size: int, config: Config
+) -> str:
+    """Generate an image with DALL-E.
+
+    Args:
+        prompt (str): The prompt to use
+        filename (str): The filename to save the image to
+        size (int): The size of the image
+
+    Returns:
+        str: The filename of the image
+    """
+
+    # Check for supported image sizes
+    if size not in [256, 512, 1024]:
+        closest = min([256, 512, 1024], key=lambda x: abs(x - size))
+        logger.info(
+            f"DALL-E only supports image sizes of 256x256, 512x512, or 1024x1024. Setting to {closest}, was {size}."
+        )
+        size = closest
+
+    response = openai.Image.create(
+        prompt=prompt,
+        n=1,
+        size=f"{size}x{size}",
+        response_format="b64_json",
+        api_key=config.openai_api_key,
+    )
+
+    logger.info(f"Image Generated for prompt:{prompt}")
+
+    image_data = b64decode(response["data"][0]["b64_json"])
+
+    with open(filename, mode="wb") as png:
+        png.write(image_data)
+
+    return f"Saved to disk:{filename}"
+
+
+def generate_image_with_sd_webui(
+    prompt: str,
+    filename: str,
+    config: Config,
+    size: int = 512,
+    negative_prompt: str = "",
+    extra: dict = {},
+) -> str:
+    """Generate an image with Stable Diffusion webui.
+    Args:
+        prompt (str): The prompt to use
+        filename (str): The filename to save the image to
+        size (int, optional): The size of the image. Defaults to 256.
+        negative_prompt (str, optional): The negative prompt to use. Defaults to "".
+        extra (dict, optional): Extra parameters to pass to the API. Defaults to {}.
+    Returns:
+        str: The filename of the image
+    """
+    # Create a session and set the basic auth if needed
+    s = requests.Session()
+    if config.sd_webui_auth:
+        username, password = config.sd_webui_auth.split(":")
+        s.auth = (username, password or "")
+
+    # Generate the images
+    response = requests.post(
+        f"{config.sd_webui_url}/sdapi/v1/txt2img",
+        json={
+            "prompt": prompt,
+            "negative_prompt": negative_prompt,
+            "sampler_index": "DDIM",
+            "steps": 20,
+            "cfg_scale": 7.0,
+            "width": size,
+            "height": size,
+            "n_iter": 1,
+            **extra,
+        },
+    )
+
+    logger.info(f"Image Generated for prompt:{prompt}")
+
+    # Save the image to disk
+    response = response.json()
+    b64 = b64decode(response["images"][0].split(",", 1)[0])
+    image = Image.open(io.BytesIO(b64))
+    image.save(filename)
+
+    return f"Saved to disk:{filename}"

autogpt/commands/improve_code.py

@@ -0,0 +1,39 @@
+from __future__ import annotations
+
+import json
+from typing import TYPE_CHECKING
+
+from autogpt.commands.command import command
+from autogpt.llm.utils import call_ai_function
+
+if TYPE_CHECKING:
+    from autogpt.config import Config
+
+
+@command(
+    "improve_code",
+    "Get Improved Code",
+    '"suggestions": "<list_of_suggestions>", "code": "<full_code_string>"',
+)
+def improve_code(suggestions: list[str], code: str, config: Config) -> str:
+    """
+    A function that takes in code and suggestions and returns a response from create
+      chat completion api call.
+
+    Parameters:
+        suggestions (list): A list of suggestions around what needs to be improved.
+        code (str): Code to be improved.
+    Returns:
+        A result string from create chat completion. Improved code in response.
+    """
+
+    function_string = (
+        "def generate_improved_code(suggestions: list[str], code: str) -> str:"
+    )
+    args = [json.dumps(suggestions), code]
+    description_string = (
+        "Improves the provided code based on the suggestions"
+        " provided, making no other changes."
+    )
+
+    return call_ai_function(function_string, args, description_string, config=config)

autogpt/commands/task_statuses.py

@@ -0,0 +1,29 @@
+"""Task Statuses module."""
+from __future__ import annotations
+
+from typing import TYPE_CHECKING, NoReturn
+
+from autogpt.commands.command import command
+from autogpt.logs import logger
+
+if TYPE_CHECKING:
+    from autogpt.config import Config
+
+
+@command(
+    "task_complete",
+    "Task Complete (Shutdown)",
+    '"reason": "<reason>"',
+)
+def task_complete(reason: str, config: Config) -> NoReturn:
+    """
+    A function that takes in a string and exits the program
+
+    Parameters:
+        reason (str): The reason for shutting down.
+    Returns:
+        A result string from create chat completion. A list of suggestions to
+            improve the code.
+    """
+    logger.info(title="Shutting down...\n", message=reason)
+    quit()

autogpt/commands/times.py

@@ -0,0 +1,10 @@
+from datetime import datetime
+
+
+def get_datetime() -> str:
+    """Return the current date and time
+
+    Returns:
+        str: The current date and time
+    """
+    return "Current date and time: " + datetime.now().strftime("%Y-%m-%d %H:%M:%S")

autogpt/commands/web_playwright.py

@@ -0,0 +1,82 @@
+"""Web scraping commands using Playwright"""
+from __future__ import annotations
+
+from autogpt.logs import logger
+
+try:
+    from playwright.sync_api import sync_playwright
+except ImportError:
+    logger.info(
+        "Playwright not installed. Please install it with 'pip install playwright' to use."
+    )
+from bs4 import BeautifulSoup
+
+from autogpt.processing.html import extract_hyperlinks, format_hyperlinks
+
+
+def scrape_text(url: str) -> str:
+    """Scrape text from a webpage
+
+    Args:
+        url (str): The URL to scrape text from
+
+    Returns:
+        str: The scraped text
+    """
+    with sync_playwright() as p:
+        browser = p.chromium.launch()
+        page = browser.new_page()
+
+        try:
+            page.goto(url)
+            html_content = page.content()
+            soup = BeautifulSoup(html_content, "html.parser")
+
+            for script in soup(["script", "style"]):
+                script.extract()
+
+            text = soup.get_text()
+            lines = (line.strip() for line in text.splitlines())
+            chunks = (phrase.strip() for line in lines for phrase in line.split("  "))
+            text = "\n".join(chunk for chunk in chunks if chunk)
+
+        except Exception as e:
+            text = f"Error: {str(e)}"
+
+        finally:
+            browser.close()
+
+    return text
+
+
+def scrape_links(url: str) -> str | list[str]:
+    """Scrape links from a webpage
+
+    Args:
+        url (str): The URL to scrape links from
+
+    Returns:
+        Union[str, List[str]]: The scraped links
+    """
+    with sync_playwright() as p:
+        browser = p.chromium.launch()
+        page = browser.new_page()
+
+        try:
+            page.goto(url)
+            html_content = page.content()
+            soup = BeautifulSoup(html_content, "html.parser")
+
+            for script in soup(["script", "style"]):
+                script.extract()
+
+            hyperlinks = extract_hyperlinks(soup, url)
+            formatted_links = format_hyperlinks(hyperlinks)
+
+        except Exception as e:
+            formatted_links = f"Error: {str(e)}"
+
+        finally:
+            browser.close()
+
+    return formatted_links

autogpt/commands/web_requests.py

@@ -0,0 +1,100 @@
+"""Browse a webpage and summarize it using the LLM model"""
+from __future__ import annotations
+
+import requests
+from bs4 import BeautifulSoup
+from requests import Response
+
+from autogpt.config import Config
+from autogpt.processing.html import extract_hyperlinks, format_hyperlinks
+from autogpt.url_utils.validators import validate_url
+
+session = requests.Session()
+
+
+@validate_url
+def get_response(
+    url: str, config: Config, timeout: int = 10
+) -> tuple[None, str] | tuple[Response, None]:
+    """Get the response from a URL
+
+    Args:
+        url (str): The URL to get the response from
+        timeout (int): The timeout for the HTTP request
+
+    Returns:
+        tuple[None, str] | tuple[Response, None]: The response and error message
+
+    Raises:
+        ValueError: If the URL is invalid
+        requests.exceptions.RequestException: If the HTTP request fails
+    """
+    try:
+        session.headers.update({"User-Agent": config.user_agent})
+        response = session.get(url, timeout=timeout)
+
+        # Check if the response contains an HTTP error
+        if response.status_code >= 400:
+            return None, f"Error: HTTP {str(response.status_code)} error"
+
+        return response, None
+    except ValueError as ve:
+        # Handle invalid URL format
+        return None, f"Error: {str(ve)}"
+
+    except requests.exceptions.RequestException as re:
+        # Handle exceptions related to the HTTP request
+        #  (e.g., connection errors, timeouts, etc.)
+        return None, f"Error: {str(re)}"
+
+
+def scrape_text(url: str, config: Config) -> str:
+    """Scrape text from a webpage
+
+    Args:
+        url (str): The URL to scrape text from
+
+    Returns:
+        str: The scraped text
+    """
+    response, error_message = get_response(url, config)
+    if error_message:
+        return error_message
+    if not response:
+        return "Error: Could not get response"
+
+    soup = BeautifulSoup(response.text, "html.parser")
+
+    for script in soup(["script", "style"]):
+        script.extract()
+
+    text = soup.get_text()
+    lines = (line.strip() for line in text.splitlines())
+    chunks = (phrase.strip() for line in lines for phrase in line.split("  "))
+    text = "\n".join(chunk for chunk in chunks if chunk)
+
+    return text
+
+
+def scrape_links(url: str, config: Config) -> str | list[str]:
+    """Scrape links from a webpage
+
+    Args:
+        url (str): The URL to scrape links from
+
+    Returns:
+       str | list[str]: The scraped links
+    """
+    response, error_message = get_response(url, config)
+    if error_message:
+        return error_message
+    if not response:
+        return "Error: Could not get response"
+    soup = BeautifulSoup(response.text, "html.parser")
+
+    for script in soup(["script", "style"]):
+        script.extract()
+
+    hyperlinks = extract_hyperlinks(soup, url)
+
+    return format_hyperlinks(hyperlinks)

autogpt/commands/web_selenium.py

@@ -0,0 +1,232 @@
+"""Selenium web scraping module."""
+from __future__ import annotations
+
+import logging
+from pathlib import Path
+from sys import platform
+from typing import TYPE_CHECKING, Optional, Type
+
+from bs4 import BeautifulSoup
+from selenium.common.exceptions import WebDriverException
+from selenium.webdriver.chrome.options import Options as ChromeOptions
+from selenium.webdriver.chrome.service import Service as ChromeDriverService
+from selenium.webdriver.chrome.webdriver import WebDriver as ChromeDriver
+from selenium.webdriver.common.by import By
+from selenium.webdriver.edge.options import Options as EdgeOptions
+from selenium.webdriver.edge.service import Service as EdgeDriverService
+from selenium.webdriver.edge.webdriver import WebDriver as EdgeDriver
+from selenium.webdriver.firefox.options import Options as FirefoxOptions
+from selenium.webdriver.firefox.service import Service as GeckoDriverService
+from selenium.webdriver.firefox.webdriver import WebDriver as FirefoxDriver
+from selenium.webdriver.remote.webdriver import WebDriver
+from selenium.webdriver.safari.options import Options as SafariOptions
+from selenium.webdriver.safari.webdriver import WebDriver as SafariDriver
+from selenium.webdriver.support import expected_conditions as EC
+from selenium.webdriver.support.wait import WebDriverWait
+from webdriver_manager.chrome import ChromeDriverManager
+from webdriver_manager.firefox import GeckoDriverManager
+from webdriver_manager.microsoft import EdgeChromiumDriverManager as EdgeDriverManager
+
+from autogpt.commands.command import command
+from autogpt.logs import logger
+from autogpt.memory.vector import MemoryItem, get_memory
+from autogpt.processing.html import extract_hyperlinks, format_hyperlinks
+from autogpt.url_utils.validators import validate_url
+
+if TYPE_CHECKING:
+    from autogpt.config import Config
+
+BrowserOptions = ChromeOptions | EdgeOptions | FirefoxOptions | SafariOptions
+
+FILE_DIR = Path(__file__).parent.parent
+
+
+@command(
+    "browse_website",
+    "Browse Website",
+    '"url": "<url>", "question": "<what_you_want_to_find_on_website>"',
+)
+@validate_url
+def browse_website(url: str, question: str, config: Config) -> str:
+    """Browse a website and return the answer and links to the user
+
+    Args:
+        url (str): The url of the website to browse
+        question (str): The question asked by the user
+
+    Returns:
+        Tuple[str, WebDriver]: The answer and links to the user and the webdriver
+    """
+    try:
+        driver, text = scrape_text_with_selenium(url, config)
+    except WebDriverException as e:
+        # These errors are often quite long and include lots of context.
+        # Just grab the first line.
+        msg = e.msg.split("\n")[0]
+        return f"Error: {msg}"
+
+    add_header(driver)
+    summary = summarize_memorize_webpage(url, text, question, config, driver)
+    links = scrape_links_with_selenium(driver, url)
+
+    # Limit links to 5
+    if len(links) > 5:
+        links = links[:5]
+    close_browser(driver)
+    return f"Answer gathered from website: {summary}\n\nLinks: {links}"
+
+
+def scrape_text_with_selenium(url: str, config: Config) -> tuple[WebDriver, str]:
+    """Scrape text from a website using selenium
+
+    Args:
+        url (str): The url of the website to scrape
+
+    Returns:
+        Tuple[WebDriver, str]: The webdriver and the text scraped from the website
+    """
+    logging.getLogger("selenium").setLevel(logging.CRITICAL)
+
+    options_available: dict[str, Type[BrowserOptions]] = {
+        "chrome": ChromeOptions,
+        "edge": EdgeOptions,
+        "firefox": FirefoxOptions,
+        "safari": SafariOptions,
+    }
+
+    options: BrowserOptions = options_available[config.selenium_web_browser]()
+    options.add_argument(
+        "user-agent=Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/112.0.5615.49 Safari/537.36"
+    )
+
+    if config.selenium_web_browser == "firefox":
+        if config.selenium_headless:
+            options.headless = True
+            options.add_argument("--disable-gpu")
+        driver = FirefoxDriver(
+            service=GeckoDriverService(GeckoDriverManager().install()), options=options
+        )
+    elif config.selenium_web_browser == "edge":
+        driver = EdgeDriver(
+            service=EdgeDriverService(EdgeDriverManager().install()), options=options
+        )
+    elif config.selenium_web_browser == "safari":
+        # Requires a bit more setup on the users end
+        # See https://developer.apple.com/documentation/webkit/testing_with_webdriver_in_safari
+        driver = SafariDriver(options=options)
+    else:
+        if platform == "linux" or platform == "linux2":
+            options.add_argument("--disable-dev-shm-usage")
+            options.add_argument("--remote-debugging-port=9222")
+
+        options.add_argument("--no-sandbox")
+        if config.selenium_headless:
+            options.add_argument("--headless=new")
+            options.add_argument("--disable-gpu")
+
+        chromium_driver_path = Path("/usr/bin/chromedriver")
+
+        driver = ChromeDriver(
+            service=ChromeDriverService(str(chromium_driver_path))
+            if chromium_driver_path.exists()
+            else ChromeDriverService(ChromeDriverManager().install()),
+            options=options,
+        )
+    driver.get(url)
+
+    WebDriverWait(driver, 10).until(
+        EC.presence_of_element_located((By.TAG_NAME, "body"))
+    )
+
+    # Get the HTML content directly from the browser's DOM
+    page_source = driver.execute_script("return document.body.outerHTML;")
+    soup = BeautifulSoup(page_source, "html.parser")
+
+    for script in soup(["script", "style"]):
+        script.extract()
+
+    text = soup.get_text()
+    lines = (line.strip() for line in text.splitlines())
+    chunks = (phrase.strip() for line in lines for phrase in line.split("  "))
+    text = "\n".join(chunk for chunk in chunks if chunk)
+    return driver, text
+
+
+def scrape_links_with_selenium(driver: WebDriver, url: str) -> list[str]:
+    """Scrape links from a website using selenium
+
+    Args:
+        driver (WebDriver): The webdriver to use to scrape the links
+
+    Returns:
+        List[str]: The links scraped from the website
+    """
+    page_source = driver.page_source
+    soup = BeautifulSoup(page_source, "html.parser")
+
+    for script in soup(["script", "style"]):
+        script.extract()
+
+    hyperlinks = extract_hyperlinks(soup, url)
+
+    return format_hyperlinks(hyperlinks)
+
+
+def close_browser(driver: WebDriver) -> None:
+    """Close the browser
+
+    Args:
+        driver (WebDriver): The webdriver to close
+
+    Returns:
+        None
+    """
+    driver.quit()
+
+
+def add_header(driver: WebDriver) -> None:
+    """Add a header to the website
+
+    Args:
+        driver (WebDriver): The webdriver to use to add the header
+
+    Returns:
+        None
+    """
+    try:
+        with open(f"{FILE_DIR}/js/overlay.js", "r") as overlay_file:
+            overlay_script = overlay_file.read()
+        driver.execute_script(overlay_script)
+    except Exception as e:
+        print(f"Error executing overlay.js: {e}")
+
+
+def summarize_memorize_webpage(
+    url: str,
+    text: str,
+    question: str,
+    config: Config,
+    driver: Optional[WebDriver] = None,
+) -> str:
+    """Summarize text using the OpenAI API
+
+    Args:
+        url (str): The url of the text
+        text (str): The text to summarize
+        question (str): The question to ask the model
+        driver (WebDriver): The webdriver to use to scroll the page
+
+    Returns:
+        str: The summary of the text
+    """
+    if not text:
+        return "Error: No text to summarize"
+
+    text_length = len(text)
+    logger.info(f"Text length: {text_length} characters")
+
+    memory = get_memory(config)
+
+    new_memory = MemoryItem.from_webpage(text, url, question=question)
+    memory.add(new_memory)
+    return new_memory.summary

autogpt/commands/write_tests.py

@@ -0,0 +1,41 @@
+"""A module that contains a function to generate test cases for the submitted code."""
+from __future__ import annotations
+
+import json
+from typing import TYPE_CHECKING
+
+from autogpt.commands.command import command
+from autogpt.llm.utils import call_ai_function
+
+if TYPE_CHECKING:
+    from autogpt.config import Config
+
+
+@command(
+    "write_tests",
+    "Write Tests",
+    '"code": "<full_code_string>", "focus": "<list_of_focus_areas>"',
+)
+def write_tests(code: str, focus: list[str], config: Config) -> str:
+    """
+    A function that takes in code and focus topics and returns a response from create
+      chat completion api call.
+
+    Parameters:
+        focus (list): A list of suggestions around what needs to be improved.
+        code (str): Code for test cases to be generated against.
+    Returns:
+        A result string from create chat completion. Test cases for the submitted code
+          in response.
+    """
+
+    function_string = (
+        "def create_test_cases(code: str, focus: Optional[str] = None) -> str:"
+    )
+    args = [code, json.dumps(focus)]
+    description_string = (
+        "Generates test cases for the existing code, focusing on"
+        " specific areas if required."
+    )
+
+    return call_ai_function(function_string, args, description_string, config=config)

autogpt/config/__init__.py

@@ -0,0 +1,11 @@
+"""
+This module contains the configuration classes for AutoGPT.
+"""
+from autogpt.config.ai_config import AIConfig
+from autogpt.config.config import Config, check_openai_api_key
+
+__all__ = [
+    "check_openai_api_key",
+    "AIConfig",
+    "Config",
+]

autogpt/config/ai_config.py

@@ -0,0 +1,170 @@
+# sourcery skip: do-not-use-staticmethod
+"""
+A module that contains the AIConfig class object that contains the configuration
+"""
+from __future__ import annotations
+
+import os
+import platform
+from pathlib import Path
+from typing import TYPE_CHECKING, Optional
+
+import distro
+import yaml
+
+if TYPE_CHECKING:
+    from autogpt.commands.command import CommandRegistry
+    from autogpt.prompts.generator import PromptGenerator
+
+# Soon this will go in a folder where it remembers more stuff about the run(s)
+SAVE_FILE = str(Path(os.getcwd()) / "ai_settings.yaml")
+
+
+class AIConfig:
+    """
+    A class object that contains the configuration information for the AI
+
+    Attributes:
+        ai_name (str): The name of the AI.
+        ai_role (str): The description of the AI's role.
+        ai_goals (list): The list of objectives the AI is supposed to complete.
+        api_budget (float): The maximum dollar value for API calls (0.0 means infinite)
+    """
+
+    def __init__(
+        self,
+        ai_name: str = "",
+        ai_role: str = "",
+        ai_goals: list | None = None,
+        api_budget: float = 0.0,
+    ) -> None:
+        """
+        Initialize a class instance
+
+        Parameters:
+            ai_name (str): The name of the AI.
+            ai_role (str): The description of the AI's role.
+            ai_goals (list): The list of objectives the AI is supposed to complete.
+            api_budget (float): The maximum dollar value for API calls (0.0 means infinite)
+        Returns:
+            None
+        """
+        if ai_goals is None:
+            ai_goals = []
+        self.ai_name = ai_name
+        self.ai_role = ai_role
+        self.ai_goals = ai_goals
+        self.api_budget = api_budget
+        self.prompt_generator: PromptGenerator | None = None
+        self.command_registry: CommandRegistry | None = None
+
+    @staticmethod
+    def load(config_file: str = SAVE_FILE) -> "AIConfig":
+        """
+        Returns class object with parameters (ai_name, ai_role, ai_goals, api_budget) loaded from
+          yaml file if yaml file exists,
+        else returns class with no parameters.
+
+        Parameters:
+           config_file (int): The path to the config yaml file.
+             DEFAULT: "../ai_settings.yaml"
+
+        Returns:
+            cls (object): An instance of given cls object
+        """
+
+        try:
+            with open(config_file, encoding="utf-8") as file:
+                config_params = yaml.load(file, Loader=yaml.FullLoader) or {}
+        except FileNotFoundError:
+            config_params = {}
+
+        ai_name = config_params.get("ai_name", "")
+        ai_role = config_params.get("ai_role", "")
+        ai_goals = [
+            str(goal).strip("{}").replace("'", "").replace('"', "")
+            if isinstance(goal, dict)
+            else str(goal)
+            for goal in config_params.get("ai_goals", [])
+        ]
+        api_budget = config_params.get("api_budget", 0.0)
+        # type: Type[AIConfig]
+        return AIConfig(ai_name, ai_role, ai_goals, api_budget)
+
+    def save(self, config_file: str = SAVE_FILE) -> None:
+        """
+        Saves the class parameters to the specified file yaml file path as a yaml file.
+
+        Parameters:
+            config_file(str): The path to the config yaml file.
+              DEFAULT: "../ai_settings.yaml"
+
+        Returns:
+            None
+        """
+
+        config = {
+            "ai_name": self.ai_name,
+            "ai_role": self.ai_role,
+            "ai_goals": self.ai_goals,
+            "api_budget": self.api_budget,
+        }
+        with open(config_file, "w", encoding="utf-8") as file:
+            yaml.dump(config, file, allow_unicode=True)
+
+    def construct_full_prompt(
+        self, prompt_generator: Optional[PromptGenerator] = None
+    ) -> str:
+        """
+        Returns a prompt to the user with the class information in an organized fashion.
+
+        Parameters:
+            None
+
+        Returns:
+            full_prompt (str): A string containing the initial prompt for the user
+              including the ai_name, ai_role, ai_goals, and api_budget.
+        """
+
+        prompt_start = (
+            "Your decisions must always be made independently without"
+            " seeking user assistance. Play to your strengths as an LLM and pursue"
+            " simple strategies with no legal complications."
+            ""
+        )
+
+        from autogpt.config import Config
+        from autogpt.prompts.prompt import build_default_prompt_generator
+
+        cfg = Config()
+        if prompt_generator is None:
+            prompt_generator = build_default_prompt_generator()
+        prompt_generator.goals = self.ai_goals
+        prompt_generator.name = self.ai_name
+        prompt_generator.role = self.ai_role
+        prompt_generator.command_registry = self.command_registry
+        for plugin in cfg.plugins:
+            if not plugin.can_handle_post_prompt():
+                continue
+            prompt_generator = plugin.post_prompt(prompt_generator)
+
+        if cfg.execute_local_commands:
+            # add OS info to prompt
+            os_name = platform.system()
+            os_info = (
+                platform.platform(terse=True)
+                if os_name != "Linux"
+                else distro.name(pretty=True)
+            )
+
+            prompt_start += f"\nThe OS you are running on is: {os_info}"
+
+        # Construct full prompt
+        full_prompt = f"You are {prompt_generator.name}, {prompt_generator.role}\n{prompt_start}\n\nGOALS:\n\n"
+        for i, goal in enumerate(self.ai_goals):
+            full_prompt += f"{i+1}. {goal}\n"
+        if self.api_budget > 0.0:
+            full_prompt += f"\nIt takes money to let you run. Your API budget is ${self.api_budget:.3f}"
+        self.prompt_generator = prompt_generator
+        full_prompt += f"\n\n{prompt_generator.generate_prompt_string()}"
+        return full_prompt

autogpt/config/config.py

@@ -0,0 +1,279 @@
+"""Configuration class to store the state of bools for different scripts access."""
+import os
+from typing import List
+
+import openai
+import yaml
+from auto_gpt_plugin_template import AutoGPTPluginTemplate
+from colorama import Fore
+
+from autogpt.singleton import Singleton
+
+
+class Config(metaclass=Singleton):
+    """
+    Configuration class to store the state of bools for different scripts access.
+    """
+
+    def __init__(self) -> None:
+        """Initialize the Config class"""
+        self.workspace_path: str = None
+        self.file_logger_path: str = None
+
+        self.debug_mode = False
+        self.continuous_mode = False
+        self.continuous_limit = 0
+        self.speak_mode = False
+        self.skip_reprompt = False
+        self.allow_downloads = False
+        self.skip_news = False
+
+        self.authorise_key = os.getenv("AUTHORISE_COMMAND_KEY", "y")
+        self.exit_key = os.getenv("EXIT_KEY", "n")
+        self.plain_output = os.getenv("PLAIN_OUTPUT", "False") == "True"
+
+        disabled_command_categories = os.getenv("DISABLED_COMMAND_CATEGORIES")
+        if disabled_command_categories:
+            self.disabled_command_categories = disabled_command_categories.split(",")
+        else:
+            self.disabled_command_categories = []
+
+        deny_commands = os.getenv("DENY_COMMANDS")
+        if deny_commands:
+            self.deny_commands = deny_commands.split(",")
+        else:
+            self.deny_commands = []
+
+        allow_commands = os.getenv("ALLOW_COMMANDS")
+        if allow_commands:
+            self.allow_commands = allow_commands.split(",")
+        else:
+            self.allow_commands = []
+
+        self.ai_settings_file = os.getenv("AI_SETTINGS_FILE", "ai_settings.yaml")
+        self.prompt_settings_file = os.getenv(
+            "PROMPT_SETTINGS_FILE", "prompt_settings.yaml"
+        )
+        self.fast_llm_model = os.getenv("FAST_LLM_MODEL", "gpt-3.5-turbo")
+        self.smart_llm_model = os.getenv("SMART_LLM_MODEL", "gpt-4")
+        self.fast_token_limit = int(os.getenv("FAST_TOKEN_LIMIT", 4000))
+        self.smart_token_limit = int(os.getenv("SMART_TOKEN_LIMIT", 8000))
+        self.embedding_model = os.getenv("EMBEDDING_MODEL", "text-embedding-ada-002")
+        self.browse_spacy_language_model = os.getenv(
+            "BROWSE_SPACY_LANGUAGE_MODEL", "en_core_web_sm"
+        )
+
+        self.openai_api_key = os.getenv("OPENAI_API_KEY")
+        self.temperature = float(os.getenv("TEMPERATURE", "0"))
+        self.use_azure = os.getenv("USE_AZURE") == "True"
+        self.execute_local_commands = (
+            os.getenv("EXECUTE_LOCAL_COMMANDS", "False") == "True"
+        )
+        self.restrict_to_workspace = (
+            os.getenv("RESTRICT_TO_WORKSPACE", "True") == "True"
+        )
+
+        if self.use_azure:
+            self.load_azure_config()
+            openai.api_type = self.openai_api_type
+            openai.api_base = self.openai_api_base
+            openai.api_version = self.openai_api_version
+
+        self.elevenlabs_api_key = os.getenv("ELEVENLABS_API_KEY")
+        self.elevenlabs_voice_1_id = os.getenv("ELEVENLABS_VOICE_1_ID")
+        self.elevenlabs_voice_2_id = os.getenv("ELEVENLABS_VOICE_2_ID")
+
+        self.use_mac_os_tts = False
+        self.use_mac_os_tts = os.getenv("USE_MAC_OS_TTS")
+
+        self.chat_messages_enabled = os.getenv("CHAT_MESSAGES_ENABLED") == "True"
+
+        self.use_brian_tts = False
+        self.use_brian_tts = os.getenv("USE_BRIAN_TTS")
+
+        self.github_api_key = os.getenv("GITHUB_API_KEY")
+        self.github_username = os.getenv("GITHUB_USERNAME")
+
+        self.google_api_key = os.getenv("GOOGLE_API_KEY")
+        self.custom_search_engine_id = os.getenv("CUSTOM_SEARCH_ENGINE_ID")
+
+        self.image_provider = os.getenv("IMAGE_PROVIDER")
+        self.image_size = int(os.getenv("IMAGE_SIZE", 256))
+        self.huggingface_api_token = os.getenv("HUGGINGFACE_API_TOKEN")
+        self.huggingface_image_model = os.getenv(
+            "HUGGINGFACE_IMAGE_MODEL", "CompVis/stable-diffusion-v1-4"
+        )
+        self.huggingface_audio_to_text_model = os.getenv(
+            "HUGGINGFACE_AUDIO_TO_TEXT_MODEL"
+        )
+        self.sd_webui_url = os.getenv("SD_WEBUI_URL", "http://localhost:7860")
+        self.sd_webui_auth = os.getenv("SD_WEBUI_AUTH")
+
+        # Selenium browser settings
+        self.selenium_web_browser = os.getenv("USE_WEB_BROWSER", "chrome")
+        self.selenium_headless = os.getenv("HEADLESS_BROWSER", "True") == "True"
+
+        # User agent header to use when making HTTP requests
+        # Some websites might just completely deny request with an error code if
+        # no user agent was found.
+        self.user_agent = os.getenv(
+            "USER_AGENT",
+            "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_4) AppleWebKit/537.36"
+            " (KHTML, like Gecko) Chrome/83.0.4103.97 Safari/537.36",
+        )
+
+        self.memory_backend = os.getenv("MEMORY_BACKEND", "json_file")
+        self.memory_index = os.getenv("MEMORY_INDEX", "auto-gpt-memory")
+
+        self.redis_host = os.getenv("REDIS_HOST", "localhost")
+        self.redis_port = int(os.getenv("REDIS_PORT", "6379"))
+        self.redis_password = os.getenv("REDIS_PASSWORD", "")
+        self.wipe_redis_on_start = os.getenv("WIPE_REDIS_ON_START", "True") == "True"
+
+        self.plugins_dir = os.getenv("PLUGINS_DIR", "plugins")
+        self.plugins: List[AutoGPTPluginTemplate] = []
+        self.plugins_openai = []
+
+        plugins_allowlist = os.getenv("ALLOWLISTED_PLUGINS")
+        if plugins_allowlist:
+            self.plugins_allowlist = plugins_allowlist.split(",")
+        else:
+            self.plugins_allowlist = []
+
+        plugins_denylist = os.getenv("DENYLISTED_PLUGINS")
+        if plugins_denylist:
+            self.plugins_denylist = plugins_denylist.split(",")
+        else:
+            self.plugins_denylist = []
+
+    def get_azure_deployment_id_for_model(self, model: str) -> str:
+        """
+        Returns the relevant deployment id for the model specified.
+
+        Parameters:
+            model(str): The model to map to the deployment id.
+
+        Returns:
+            The matching deployment id if found, otherwise an empty string.
+        """
+        if model == self.fast_llm_model:
+            return self.azure_model_to_deployment_id_map[
+                "fast_llm_model_deployment_id"
+            ]  # type: ignore
+        elif model == self.smart_llm_model:
+            return self.azure_model_to_deployment_id_map[
+                "smart_llm_model_deployment_id"
+            ]  # type: ignore
+        elif model == "text-embedding-ada-002":
+            return self.azure_model_to_deployment_id_map[
+                "embedding_model_deployment_id"
+            ]  # type: ignore
+        else:
+            return ""
+
+    AZURE_CONFIG_FILE = os.path.join(os.path.dirname(__file__), "../..", "azure.yaml")
+
+    def load_azure_config(self, config_file: str = AZURE_CONFIG_FILE) -> None:
+        """
+        Loads the configuration parameters for Azure hosting from the specified file
+          path as a yaml file.
+
+        Parameters:
+            config_file(str): The path to the config yaml file. DEFAULT: "../azure.yaml"
+
+        Returns:
+            None
+        """
+        with open(config_file) as file:
+            config_params = yaml.load(file, Loader=yaml.FullLoader) or {}
+        self.openai_api_type = config_params.get("azure_api_type") or "azure"
+        self.openai_api_base = config_params.get("azure_api_base") or ""
+        self.openai_api_version = (
+            config_params.get("azure_api_version") or "2023-03-15-preview"
+        )
+        self.azure_model_to_deployment_id_map = config_params.get("azure_model_map", {})
+
+    def set_continuous_mode(self, value: bool) -> None:
+        """Set the continuous mode value."""
+        self.continuous_mode = value
+
+    def set_continuous_limit(self, value: int) -> None:
+        """Set the continuous limit value."""
+        self.continuous_limit = value
+
+    def set_speak_mode(self, value: bool) -> None:
+        """Set the speak mode value."""
+        self.speak_mode = value
+
+    def set_fast_llm_model(self, value: str) -> None:
+        """Set the fast LLM model value."""
+        self.fast_llm_model = value
+
+    def set_smart_llm_model(self, value: str) -> None:
+        """Set the smart LLM model value."""
+        self.smart_llm_model = value
+
+    def set_fast_token_limit(self, value: int) -> None:
+        """Set the fast token limit value."""
+        self.fast_token_limit = value
+
+    def set_smart_token_limit(self, value: int) -> None:
+        """Set the smart token limit value."""
+        self.smart_token_limit = value
+
+    def set_embedding_model(self, value: str) -> None:
+        """Set the model to use for creating embeddings."""
+        self.embedding_model = value
+
+    def set_openai_api_key(self, value: str) -> None:
+        """Set the OpenAI API key value."""
+        self.openai_api_key = value
+
+    def set_elevenlabs_api_key(self, value: str) -> None:
+        """Set the ElevenLabs API key value."""
+        self.elevenlabs_api_key = value
+
+    def set_elevenlabs_voice_1_id(self, value: str) -> None:
+        """Set the ElevenLabs Voice 1 ID value."""
+        self.elevenlabs_voice_1_id = value
+
+    def set_elevenlabs_voice_2_id(self, value: str) -> None:
+        """Set the ElevenLabs Voice 2 ID value."""
+        self.elevenlabs_voice_2_id = value
+
+    def set_google_api_key(self, value: str) -> None:
+        """Set the Google API key value."""
+        self.google_api_key = value
+
+    def set_custom_search_engine_id(self, value: str) -> None:
+        """Set the custom search engine id value."""
+        self.custom_search_engine_id = value
+
+    def set_debug_mode(self, value: bool) -> None:
+        """Set the debug mode value."""
+        self.debug_mode = value
+
+    def set_plugins(self, value: list) -> None:
+        """Set the plugins value."""
+        self.plugins = value
+
+    def set_temperature(self, value: int) -> None:
+        """Set the temperature value."""
+        self.temperature = value
+
+    def set_memory_backend(self, name: str) -> None:
+        """Set the memory backend name."""
+        self.memory_backend = name
+
+
+def check_openai_api_key() -> None:
+    """Check if the OpenAI API key is set in config.py or as an environment variable."""
+    cfg = Config()
+    if not cfg.openai_api_key:
+        print(
+            Fore.RED
+            + "Please set your OpenAI API key in .env or as an environment variable."
+            + Fore.RESET
+        )
+        print("You can get your key from https://platform.openai.com/account/api-keys")
+        exit(1)

autogpt/config/prompt_config.py

@@ -0,0 +1,53 @@
+# sourcery skip: do-not-use-staticmethod
+"""
+A module that contains the PromptConfig class object that contains the configuration
+"""
+import yaml
+from colorama import Fore
+
+from autogpt import utils
+from autogpt.config.config import Config
+from autogpt.logs import logger
+
+CFG = Config()
+
+
+class PromptConfig:
+    """
+    A class object that contains the configuration information for the prompt, which will be used by the prompt generator
+
+    Attributes:
+        constraints (list): Constraints list for the prompt generator.
+        resources (list): Resources list for the prompt generator.
+        performance_evaluations (list): Performance evaluation list for the prompt generator.
+    """
+
+    def __init__(
+        self,
+        config_file: str = CFG.prompt_settings_file,
+    ) -> None:
+        """
+        Initialize a class instance with parameters (constraints, resources, performance_evaluations) loaded from
+          yaml file if yaml file exists,
+        else raises error.
+
+        Parameters:
+            constraints (list): Constraints list for the prompt generator.
+            resources (list): Resources list for the prompt generator.
+            performance_evaluations (list): Performance evaluation list for the prompt generator.
+        Returns:
+            None
+        """
+        # Validate file
+        (validated, message) = utils.validate_yaml_file(config_file)
+        if not validated:
+            logger.typewriter_log("FAILED FILE VALIDATION", Fore.RED, message)
+            logger.double_check()
+            exit(1)
+
+        with open(config_file, encoding="utf-8") as file:
+            config_params = yaml.load(file, Loader=yaml.FullLoader)
+
+        self.constraints = config_params.get("constraints", [])
+        self.resources = config_params.get("resources", [])
+        self.performance_evaluations = config_params.get("performance_evaluations", [])

autogpt/js/overlay.js

@@ -0,0 +1,29 @@
+const overlay = document.createElement('div');
+Object.assign(overlay.style, {
+    position: 'fixed',
+    zIndex: 999999,
+    top: 0,
+    left: 0,
+    width: '100%',
+    height: '100%',
+    background: 'rgba(0, 0, 0, 0.7)',
+    color: '#fff',
+    fontSize: '24px',
+    fontWeight: 'bold',
+    display: 'flex',
+    justifyContent: 'center',
+    alignItems: 'center',
+});
+const textContent = document.createElement('div');
+Object.assign(textContent.style, {
+    textAlign: 'center',
+});
+textContent.textContent = 'AutoGPT Analyzing Page';
+overlay.appendChild(textContent);
+document.body.append(overlay);
+document.body.style.overflow = 'hidden';
+let dotCount = 0;
+setInterval(() => {
+    textContent.textContent = 'AutoGPT Analyzing Page' + '.'.repeat(dotCount);
+    dotCount = (dotCount + 1) % 4;
+}, 1000);

autogpt/json_utils/json_fix_general.py

@@ -0,0 +1,121 @@
+"""This module contains functions to fix JSON strings using general programmatic approaches, suitable for addressing
+common JSON formatting issues."""
+from __future__ import annotations
+
+import contextlib
+import json
+import re
+from typing import Optional
+
+from autogpt.config import Config
+from autogpt.json_utils.utilities import extract_char_position
+from autogpt.logs import logger
+
+CFG = Config()
+
+
+def fix_invalid_escape(json_to_load: str, error_message: str) -> str:
+    """Fix invalid escape sequences in JSON strings.
+
+    Args:
+        json_to_load (str): The JSON string.
+        error_message (str): The error message from the JSONDecodeError
+          exception.
+
+    Returns:
+        str: The JSON string with invalid escape sequences fixed.
+    """
+    while error_message.startswith("Invalid \\escape"):
+        bad_escape_location = extract_char_position(error_message)
+        json_to_load = (
+            json_to_load[:bad_escape_location] + json_to_load[bad_escape_location + 1 :]
+        )
+        try:
+            json.loads(json_to_load)
+            return json_to_load
+        except json.JSONDecodeError as e:
+            logger.debug("json loads error - fix invalid escape", e)
+            error_message = str(e)
+    return json_to_load
+
+
+def balance_braces(json_string: str) -> Optional[str]:
+    """
+    Balance the braces in a JSON string.
+
+    Args:
+        json_string (str): The JSON string.
+
+    Returns:
+        str: The JSON string with braces balanced.
+    """
+
+    open_braces_count = json_string.count("{")
+    close_braces_count = json_string.count("}")
+
+    while open_braces_count > close_braces_count:
+        json_string += "}"
+        close_braces_count += 1
+
+    while close_braces_count > open_braces_count:
+        json_string = json_string.rstrip("}")
+        close_braces_count -= 1
+
+    with contextlib.suppress(json.JSONDecodeError):
+        json.loads(json_string)
+        return json_string
+
+
+def add_quotes_to_property_names(json_string: str) -> str:
+    """
+    Add quotes to property names in a JSON string.
+
+    Args:
+        json_string (str): The JSON string.
+
+    Returns:
+        str: The JSON string with quotes added to property names.
+    """
+
+    def replace_func(match: re.Match) -> str:
+        return f'"{match[1]}":'
+
+    property_name_pattern = re.compile(r"(\w+):")
+    corrected_json_string = property_name_pattern.sub(replace_func, json_string)
+
+    try:
+        json.loads(corrected_json_string)
+        return corrected_json_string
+    except json.JSONDecodeError as e:
+        raise e
+
+
+def correct_json(json_to_load: str) -> str:
+    """
+    Correct common JSON errors.
+    Args:
+        json_to_load (str): The JSON string.
+    """
+
+    try:
+        logger.debug("json", json_to_load)
+        json.loads(json_to_load)
+        return json_to_load
+    except json.JSONDecodeError as e:
+        logger.debug("json loads error", e)
+        error_message = str(e)
+        if error_message.startswith("Invalid \\escape"):
+            json_to_load = fix_invalid_escape(json_to_load, error_message)
+        if error_message.startswith(
+            "Expecting property name enclosed in double quotes"
+        ):
+            json_to_load = add_quotes_to_property_names(json_to_load)
+            try:
+                json.loads(json_to_load)
+                return json_to_load
+            except json.JSONDecodeError as e:
+                logger.debug("json loads error - add quotes", e)
+                error_message = str(e)
+        if balanced_str := balance_braces(json_to_load):
+            return balanced_str
+    return json_to_load

autogpt/json_utils/json_fix_llm.py

@@ -0,0 +1,239 @@
+"""This module contains functions to fix JSON strings generated by LLM models, such as ChatGPT, using the assistance
+of the ChatGPT API or LLM models."""
+from __future__ import annotations
+
+import contextlib
+import json
+from typing import Any, Dict
+
+from colorama import Fore
+from regex import regex
+
+from autogpt.config import Config
+from autogpt.json_utils.json_fix_general import correct_json
+from autogpt.llm.utils import call_ai_function
+from autogpt.logs import logger
+from autogpt.speech import say_text
+
+JSON_SCHEMA = """
+{
+    "command": {
+        "name": "command name",
+        "args": {
+            "arg name": "value"
+        }
+    },
+    "thoughts":
+    {
+        "text": "thought",
+        "reasoning": "reasoning",
+        "plan": "- short bulleted\n- list that conveys\n- long-term plan",
+        "criticism": "constructive self-criticism",
+        "speak": "thoughts summary to say to user"
+    }
+}
+"""
+
+CFG = Config()
+
+
+def auto_fix_json(json_string: str, schema: str) -> str:
+    """Fix the given JSON string to make it parseable and fully compliant with
+        the provided schema using GPT-3.
+
+    Args:
+        json_string (str): The JSON string to fix.
+        schema (str): The schema to use to fix the JSON.
+    Returns:
+        str: The fixed JSON string.
+    """
+    # Try to fix the JSON using GPT:
+    function_string = "def fix_json(json_string: str, schema:str=None) -> str:"
+    args = [f"'''{json_string}'''", f"'''{schema}'''"]
+    description_string = (
+        "This function takes a JSON string and ensures that it"
+        " is parseable and fully compliant with the provided schema. If an object"
+        " or field specified in the schema isn't contained within the correct JSON,"
+        " it is omitted. The function also escapes any double quotes within JSON"
+        " string values to ensure that they are valid. If the JSON string contains"
+        " any None or NaN values, they are replaced with null before being parsed."
+    )
+
+    # If it doesn't already start with a "`", add one:
+    if not json_string.startswith("`"):
+        json_string = "```json\n" + json_string + "\n```"
+    result_string = call_ai_function(
+        function_string, args, description_string, model=CFG.fast_llm_model
+    )
+    logger.debug("------------ JSON FIX ATTEMPT ---------------")
+    logger.debug(f"Original JSON: {json_string}")
+    logger.debug("-----------")
+    logger.debug(f"Fixed JSON: {result_string}")
+    logger.debug("----------- END OF FIX ATTEMPT ----------------")
+
+    try:
+        json.loads(result_string)  # just check the validity
+        return result_string
+    except json.JSONDecodeError:  # noqa: E722
+        # Get the call stack:
+        # import traceback
+        # call_stack = traceback.format_exc()
+        # print(f"Failed to fix JSON: '{json_string}' "+call_stack)
+        return "failed"
+
+
+def fix_json_using_multiple_techniques(assistant_reply: str) -> Dict[Any, Any]:
+    """Fix the given JSON string to make it parseable and fully compliant with two techniques.
+
+    Args:
+        json_string (str): The JSON string to fix.
+
+    Returns:
+        str: The fixed JSON string.
+    """
+    assistant_reply = assistant_reply.strip()
+    if assistant_reply.startswith("```json"):
+        assistant_reply = assistant_reply[7:]
+    if assistant_reply.endswith("```"):
+        assistant_reply = assistant_reply[:-3]
+    try:
+        return json.loads(assistant_reply)  # just check the validity
+    except json.JSONDecodeError:  # noqa: E722
+        pass
+
+    if assistant_reply.startswith("json "):
+        assistant_reply = assistant_reply[5:]
+        assistant_reply = assistant_reply.strip()
+    try:
+        return json.loads(assistant_reply)  # just check the validity
+    except json.JSONDecodeError:  # noqa: E722
+        pass
+
+    # Parse and print Assistant response
+    assistant_reply_json = fix_and_parse_json(assistant_reply)
+    logger.debug("Assistant reply JSON: %s", str(assistant_reply_json))
+    if assistant_reply_json == {}:
+        assistant_reply_json = attempt_to_fix_json_by_finding_outermost_brackets(
+            assistant_reply
+        )
+
+    logger.debug("Assistant reply JSON 2: %s", str(assistant_reply_json))
+    if assistant_reply_json != {}:
+        return assistant_reply_json
+
+    logger.error(
+        "Error: The following AI output couldn't be converted to a JSON:\n",
+        assistant_reply,
+    )
+    if CFG.speak_mode:
+        say_text("I have received an invalid JSON response from the OpenAI API.")
+
+    return {}
+
+
+def fix_and_parse_json(
+    json_to_load: str, try_to_fix_with_gpt: bool = True
+) -> Dict[Any, Any]:
+    """Fix and parse JSON string
+
+    Args:
+        json_to_load (str): The JSON string.
+        try_to_fix_with_gpt (bool, optional): Try to fix the JSON with GPT.
+            Defaults to True.
+
+    Returns:
+        str or dict[Any, Any]: The parsed JSON.
+    """
+
+    with contextlib.suppress(json.JSONDecodeError):
+        json_to_load = json_to_load.replace("\t", "")
+        return json.loads(json_to_load)
+
+    with contextlib.suppress(json.JSONDecodeError):
+        json_to_load = correct_json(json_to_load)
+        return json.loads(json_to_load)
+    # Let's do something manually:
+    # sometimes GPT responds with something BEFORE the braces:
+    # "I'm sorry, I don't understand. Please try again."
+    # {"text": "I'm sorry, I don't understand. Please try again.",
+    #  "confidence": 0.0}
+    # So let's try to find the first brace and then parse the rest
+    #  of the string
+    try:
+        brace_index = json_to_load.index("{")
+        maybe_fixed_json = json_to_load[brace_index:]
+        last_brace_index = maybe_fixed_json.rindex("}")
+        maybe_fixed_json = maybe_fixed_json[: last_brace_index + 1]
+        return json.loads(maybe_fixed_json)
+    except (json.JSONDecodeError, ValueError) as e:
+        return try_ai_fix(try_to_fix_with_gpt, e, json_to_load)
+
+
+def try_ai_fix(
+    try_to_fix_with_gpt: bool, exception: Exception, json_to_load: str
+) -> Dict[Any, Any]:
+    """Try to fix the JSON with the AI
+
+    Args:
+        try_to_fix_with_gpt (bool): Whether to try to fix the JSON with the AI.
+        exception (Exception): The exception that was raised.
+        json_to_load (str): The JSON string to load.
+
+    Raises:
+        exception: If try_to_fix_with_gpt is False.
+
+    Returns:
+        str or dict[Any, Any]: The JSON string or dictionary.
+    """
+    if not try_to_fix_with_gpt:
+        raise exception
+    if CFG.debug_mode:
+        logger.warn(
+            "Warning: Failed to parse AI output, attempting to fix."
+            "\n If you see this warning frequently, it's likely that"
+            " your prompt is confusing the AI. Try changing it up"
+            " slightly."
+        )
+    # Now try to fix this up using the ai_functions
+    ai_fixed_json = auto_fix_json(json_to_load, JSON_SCHEMA)
+
+    if ai_fixed_json != "failed":
+        return json.loads(ai_fixed_json)
+    # This allows the AI to react to the error message,
+    #   which usually results in it correcting its ways.
+    # logger.error("Failed to fix AI output, telling the AI.")
+    return {}
+
+
+def attempt_to_fix_json_by_finding_outermost_brackets(json_string: str):
+    if CFG.speak_mode and CFG.debug_mode:
+        say_text(
+            "I have received an invalid JSON response from the OpenAI API. "
+            "Trying to fix it now."
+        )
+        logger.error("Attempting to fix JSON by finding outermost brackets\n")
+
+    try:
+        json_pattern = regex.compile(r"\{(?:[^{}]|(?R))*\}")
+        json_match = json_pattern.search(json_string)
+
+        if json_match:
+            # Extract the valid JSON object from the string
+            json_string = json_match.group(0)
+            logger.typewriter_log(
+                title="Apparently json was fixed.", title_color=Fore.GREEN
+            )
+            if CFG.speak_mode and CFG.debug_mode:
+                say_text("Apparently json was fixed.")
+        else:
+            return {}
+
+    except (json.JSONDecodeError, ValueError):
+        if CFG.debug_mode:
+            logger.error(f"Error: Invalid JSON: {json_string}\n")
+        if CFG.speak_mode:
+            say_text("Didn't work. I will have to ignore this response then.")
+        logger.error("Error: Invalid JSON, setting it to empty JSON now.\n")
+        json_string = {}
+
+    return fix_and_parse_json(json_string)

autogpt/json_utils/llm_response_format_1.json

@@ -0,0 +1,31 @@
+{
+    "$schema": "http://json-schema.org/draft-07/schema#",
+    "type": "object",
+    "properties": {
+        "thoughts": {
+            "type": "object",
+            "properties": {
+                "text": {"type": "string"},
+                "reasoning": {"type": "string"},
+                "plan": {"type": "string"},
+                "criticism": {"type": "string"},
+                "speak": {"type": "string"}
+            },
+            "required": ["text", "reasoning", "plan", "criticism", "speak"],
+            "additionalProperties": false
+        },
+        "command": {
+            "type": "object",
+            "properties": {
+                "name": {"type": "string"},
+                "args": {
+                    "type": "object"
+                }
+            },
+            "required": ["name", "args"],
+            "additionalProperties": false
+        }
+    },
+    "required": ["thoughts", "command"],
+    "additionalProperties": false
+}

autogpt/json_utils/utilities.py

@@ -0,0 +1,82 @@
+"""Utilities for the json_fixes package."""
+import json
+import os.path
+import re
+
+from jsonschema import Draft7Validator
+
+from autogpt.config import Config
+from autogpt.logs import logger
+
+CFG = Config()
+LLM_DEFAULT_RESPONSE_FORMAT = "llm_response_format_1"
+
+
+def extract_char_position(error_message: str) -> int:
+    """Extract the character position from the JSONDecodeError message.
+
+    Args:
+        error_message (str): The error message from the JSONDecodeError
+          exception.
+
+    Returns:
+        int: The character position.
+    """
+
+    char_pattern = re.compile(r"\(char (\d+)\)")
+    if match := char_pattern.search(error_message):
+        return int(match[1])
+    else:
+        raise ValueError("Character position not found in the error message.")
+
+
+def validate_json(json_object: object, schema_name: str):
+    """
+    :type schema_name: object
+    :param schema_name: str
+    :type json_object: object
+    """
+    # with open(f"/Users/kilig/Job/Python-project/auto-gpt/autogpt/json_utils/{schema_name}.json", "r") as f:
+    scheme_file = os.path.join(os.path.dirname(__file__), f"{schema_name}.json")
+    with open(scheme_file, "r") as f:
+        schema = json.load(f)
+    validator = Draft7Validator(schema)
+
+    if errors := sorted(validator.iter_errors(json_object), key=lambda e: e.path):
+        logger.error("The JSON object is invalid.")
+        if CFG.debug_mode:
+            logger.error(
+                json.dumps(json_object, indent=4)
+            )  # Replace 'json_object' with the variable containing the JSON data
+            logger.error("The following issues were found:")
+
+            for error in errors:
+                logger.error(f"Error: {error.message}")
+    else:
+        logger.debug("The JSON object is valid.")
+
+    return json_object
+
+
+def validate_json_string(json_string: str, schema_name: str):
+    """
+    :type schema_name: object
+    :param schema_name: str
+    :type json_object: object
+    """
+
+    try:
+        json_loaded = json.loads(json_string)
+        return validate_json(json_loaded, schema_name)
+    except:
+        return None
+
+
+def is_string_valid_json(json_string: str, schema_name: str) -> bool:
+    """
+    :type schema_name: object
+    :param schema_name: str
+    :type json_object: object
+    """
+
+    return validate_json_string(json_string, schema_name) is not None

autogpt/llm/__init__.py

@@ -0,0 +1,19 @@
+from autogpt.llm.base import (
+    ChatModelInfo,
+    ChatModelResponse,
+    EmbeddingModelInfo,
+    EmbeddingModelResponse,
+    LLMResponse,
+    Message,
+    ModelInfo,
+)
+
+__all__ = [
+    "Message",
+    "ModelInfo",
+    "ChatModelInfo",
+    "EmbeddingModelInfo",
+    "LLMResponse",
+    "ChatModelResponse",
+    "EmbeddingModelResponse",
+]

autogpt/llm/api_manager.py

@@ -0,0 +1,152 @@
+from __future__ import annotations
+
+from typing import List, Optional
+
+import openai
+from openai import Model
+
+from autogpt.config import Config
+from autogpt.llm.base import MessageDict
+from autogpt.llm.modelsinfo import COSTS
+from autogpt.logs import logger
+from autogpt.singleton import Singleton
+
+
+class ApiManager(metaclass=Singleton):
+    def __init__(self):
+        self.total_prompt_tokens = 0
+        self.total_completion_tokens = 0
+        self.total_cost = 0
+        self.total_budget = 0
+        self.models: Optional[list[Model]] = None
+
+    def reset(self):
+        self.total_prompt_tokens = 0
+        self.total_completion_tokens = 0
+        self.total_cost = 0
+        self.total_budget = 0.0
+        self.models = None
+
+    def create_chat_completion(
+        self,
+        messages: list[MessageDict],
+        model: str | None = None,
+        temperature: float = None,
+        max_tokens: int | None = None,
+        deployment_id=None,
+    ) -> str:
+        """
+        Create a chat completion and update the cost.
+        Args:
+        messages (list): The list of messages to send to the API.
+        model (str): The model to use for the API call.
+        temperature (float): The temperature to use for the API call.
+        max_tokens (int): The maximum number of tokens for the API call.
+        Returns:
+        str: The AI's response.
+        """
+        cfg = Config()
+        if temperature is None:
+            temperature = cfg.temperature
+        if deployment_id is not None:
+            response = openai.ChatCompletion.create(
+                deployment_id=deployment_id,
+                model=model,
+                messages=messages,
+                temperature=temperature,
+                max_tokens=max_tokens,
+                api_key=cfg.openai_api_key,
+            )
+        else:
+            response = openai.ChatCompletion.create(
+                model=model,
+                messages=messages,
+                temperature=temperature,
+                max_tokens=max_tokens,
+                api_key=cfg.openai_api_key,
+            )
+        if not hasattr(response, "error"):
+            logger.debug(f"Response: {response}")
+            prompt_tokens = response.usage.prompt_tokens
+            completion_tokens = response.usage.completion_tokens
+            self.update_cost(prompt_tokens, completion_tokens, model)
+        return response
+
+    def update_cost(self, prompt_tokens, completion_tokens, model: str):
+        """
+        Update the total cost, prompt tokens, and completion tokens.
+
+        Args:
+        prompt_tokens (int): The number of tokens used in the prompt.
+        completion_tokens (int): The number of tokens used in the completion.
+        model (str): The model used for the API call.
+        """
+        # the .model property in API responses can contain version suffixes like -v2
+        model = model[:-3] if model.endswith("-v2") else model
+
+        self.total_prompt_tokens += prompt_tokens
+        self.total_completion_tokens += completion_tokens
+        self.total_cost += (
+            prompt_tokens * COSTS[model]["prompt"]
+            + completion_tokens * COSTS[model]["completion"]
+        ) / 1000
+        logger.debug(f"Total running cost: ${self.total_cost:.3f}")
+
+    def set_total_budget(self, total_budget):
+        """
+        Sets the total user-defined budget for API calls.
+
+        Args:
+        total_budget (float): The total budget for API calls.
+        """
+        self.total_budget = total_budget
+
+    def get_total_prompt_tokens(self):
+        """
+        Get the total number of prompt tokens.
+
+        Returns:
+        int: The total number of prompt tokens.
+        """
+        return self.total_prompt_tokens
+
+    def get_total_completion_tokens(self):
+        """
+        Get the total number of completion tokens.
+
+        Returns:
+        int: The total number of completion tokens.
+        """
+        return self.total_completion_tokens
+
+    def get_total_cost(self):
+        """
+        Get the total cost of API calls.
+
+        Returns:
+        float: The total cost of API calls.
+        """
+        return self.total_cost
+
+    def get_total_budget(self):
+        """
+        Get the total user-defined budget for API calls.
+
+        Returns:
+        float: The total budget for API calls.
+        """
+        return self.total_budget
+
+    def get_models(self) -> List[Model]:
+        """
+        Get list of available GPT models.
+
+        Returns:
+        list: List of available GPT models.
+
+        """
+        if self.models is None:
+            all_models = openai.Model.list()["data"]
+            self.models = [model for model in all_models if "gpt" in model["id"]]
+
+        return self.models

autogpt/llm/base.py

@@ -0,0 +1,151 @@
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from math import ceil, floor
+from typing import List, Literal, TypedDict
+
+MessageRole = Literal["system", "user", "assistant"]
+MessageType = Literal["ai_response", "action_result"]
+
+
+class MessageDict(TypedDict):
+    role: MessageRole
+    content: str
+
+
+@dataclass
+class Message:
+    """OpenAI Message object containing a role and the message content"""
+
+    role: MessageRole
+    content: str
+    type: MessageType | None = None
+
+    def raw(self) -> MessageDict:
+        return {"role": self.role, "content": self.content}
+
+
+@dataclass
+class ModelInfo:
+    """Struct for model information.
+
+    Would be lovely to eventually get this directly from APIs, but needs to be scraped from
+    websites for now.
+
+    """
+
+    name: str
+    prompt_token_cost: float
+    completion_token_cost: float
+    max_tokens: int
+
+
+@dataclass
+class ChatModelInfo(ModelInfo):
+    """Struct for chat model information."""
+
+
+@dataclass
+class TextModelInfo(ModelInfo):
+    """Struct for text completion model information."""
+
+
+@dataclass
+class EmbeddingModelInfo(ModelInfo):
+    """Struct for embedding model information."""
+
+    embedding_dimensions: int
+
+
+@dataclass
+class ChatSequence:
+    """Utility container for a chat sequence"""
+
+    model: ChatModelInfo
+    messages: list[Message] = field(default_factory=list)
+
+    def __getitem__(self, i: int):
+        return self.messages[i]
+
+    def __iter__(self):
+        return iter(self.messages)
+
+    def __len__(self):
+        return len(self.messages)
+
+    def append(self, message: Message):
+        return self.messages.append(message)
+
+    def extend(self, messages: list[Message] | ChatSequence):
+        return self.messages.extend(messages)
+
+    def insert(self, index: int, *messages: Message):
+        for message in reversed(messages):
+            self.messages.insert(index, message)
+
+    @classmethod
+    def for_model(cls, model_name: str, messages: list[Message] | ChatSequence = []):
+        from autogpt.llm.providers.openai import OPEN_AI_CHAT_MODELS
+
+        if not model_name in OPEN_AI_CHAT_MODELS:
+            raise ValueError(f"Unknown chat model '{model_name}'")
+
+        return ChatSequence(
+            model=OPEN_AI_CHAT_MODELS[model_name], messages=list(messages)
+        )
+
+    def add(self, message_role: MessageRole, content: str):
+        self.messages.append(Message(message_role, content))
+
+    @property
+    def token_length(self):
+        from autogpt.llm.utils import count_message_tokens
+
+        return count_message_tokens(self.messages, self.model.name)
+
+    def raw(self) -> list[MessageDict]:
+        return [m.raw() for m in self.messages]
+
+    def dump(self) -> str:
+        SEPARATOR_LENGTH = 42
+
+        def separator(text: str):
+            half_sep_len = (SEPARATOR_LENGTH - 2 - len(text)) / 2
+            return f"{floor(half_sep_len)*'-'} {text.upper()} {ceil(half_sep_len)*'-'}"
+
+        formatted_messages = "\n".join(
+            [f"{separator(m.role)}\n{m.content}" for m in self.messages]
+        )
+        return f"""
+============== ChatSequence ==============
+Length: {self.token_length} tokens; {len(self.messages)} messages
+{formatted_messages}
+==========================================
+"""
+
+
+@dataclass
+class LLMResponse:
+    """Standard response struct for a response from an LLM model."""
+
+    model_info: ModelInfo
+    prompt_tokens_used: int = 0
+    completion_tokens_used: int = 0
+
+
+@dataclass
+class EmbeddingModelResponse(LLMResponse):
+    """Standard response struct for a response from an embedding model."""
+
+    embedding: List[float] = field(default_factory=list)
+
+    def __post_init__(self):
+        if self.completion_tokens_used:
+            raise ValueError("Embeddings should not have completion tokens used.")
+
+
+@dataclass
+class ChatModelResponse(LLMResponse):
+    """Standard response struct for a response from an LLM model."""
+
+    content: str = None

autogpt/llm/modelsinfo.py

@@ -0,0 +1,11 @@
+COSTS = {
+    "gpt-3.5-turbo": {"prompt": 0.002, "completion": 0.002},
+    "gpt-3.5-turbo-0301": {"prompt": 0.002, "completion": 0.002},
+    "gpt-4-0314": {"prompt": 0.03, "completion": 0.06},
+    "gpt-4": {"prompt": 0.03, "completion": 0.06},
+    "gpt-4-0314": {"prompt": 0.03, "completion": 0.06},
+    "gpt-4-32k": {"prompt": 0.06, "completion": 0.12},
+    "gpt-4-32k-0314": {"prompt": 0.06, "completion": 0.12},
+    "text-embedding-ada-002": {"prompt": 0.0004, "completion": 0.0},
+    "text-davinci-003": {"prompt": 0.02, "completion": 0.02},
+}

autogpt/llm/providers/openai.py

@@ -0,0 +1,74 @@
+from autogpt.llm.base import ChatModelInfo, EmbeddingModelInfo, TextModelInfo
+
+OPEN_AI_CHAT_MODELS = {
+    info.name: info
+    for info in [
+        ChatModelInfo(
+            name="gpt-3.5-turbo",
+            prompt_token_cost=0.002,
+            completion_token_cost=0.002,
+            max_tokens=4096,
+        ),
+        ChatModelInfo(
+            name="gpt-3.5-turbo-0301",
+            prompt_token_cost=0.002,
+            completion_token_cost=0.002,
+            max_tokens=4096,
+        ),
+        ChatModelInfo(
+            name="gpt-4",
+            prompt_token_cost=0.03,
+            completion_token_cost=0.06,
+            max_tokens=8192,
+        ),
+        ChatModelInfo(
+            name="gpt-4-0314",
+            prompt_token_cost=0.03,
+            completion_token_cost=0.06,
+            max_tokens=8192,
+        ),
+        ChatModelInfo(
+            name="gpt-4-32k",
+            prompt_token_cost=0.06,
+            completion_token_cost=0.12,
+            max_tokens=32768,
+        ),
+        ChatModelInfo(
+            name="gpt-4-32k-0314",
+            prompt_token_cost=0.06,
+            completion_token_cost=0.12,
+            max_tokens=32768,
+        ),
+    ]
+}
+
+OPEN_AI_TEXT_MODELS = {
+    info.name: info
+    for info in [
+        TextModelInfo(
+            name="text-davinci-003",
+            prompt_token_cost=0.02,
+            completion_token_cost=0.02,
+            max_tokens=4097,
+        ),
+    ]
+}
+
+OPEN_AI_EMBEDDING_MODELS = {
+    info.name: info
+    for info in [
+        EmbeddingModelInfo(
+            name="text-embedding-ada-002",
+            prompt_token_cost=0.0004,
+            completion_token_cost=0.0,
+            max_tokens=8191,
+            embedding_dimensions=1536,
+        ),
+    ]
+}
+
+OPEN_AI_MODELS: dict[str, ChatModelInfo | EmbeddingModelInfo | TextModelInfo] = {
+    **OPEN_AI_CHAT_MODELS,
+    **OPEN_AI_TEXT_MODELS,
+    **OPEN_AI_EMBEDDING_MODELS,
+}

autogpt/llm/utils/__init__.py

@@ -0,0 +1,266 @@
+from __future__ import annotations
+
+import functools
+import time
+from typing import List, Literal, Optional
+from unittest.mock import patch
+
+import openai
+import openai.api_resources.abstract.engine_api_resource as engine_api_resource
+import openai.util
+from colorama import Fore, Style
+from openai.error import APIError, RateLimitError
+from openai.openai_object import OpenAIObject
+
+from autogpt.config import Config
+from autogpt.logs import logger
+
+from ..api_manager import ApiManager
+from ..base import ChatSequence, Message
+from .token_counter import *
+
+
+def metered(func):
+    """Adds ApiManager metering to functions which make OpenAI API calls"""
+    api_manager = ApiManager()
+
+    openai_obj_processor = openai.util.convert_to_openai_object
+
+    def update_usage_with_response(response: OpenAIObject):
+        try:
+            usage = response.usage
+            logger.debug(f"Reported usage from call to model {response.model}: {usage}")
+            api_manager.update_cost(
+                response.usage.prompt_tokens,
+                response.usage.completion_tokens if "completion_tokens" in usage else 0,
+                response.model,
+            )
+        except Exception as err:
+            logger.warn(f"Failed to update API costs: {err.__class__.__name__}: {err}")
+
+    def metering_wrapper(*args, **kwargs):
+        openai_obj = openai_obj_processor(*args, **kwargs)
+        if isinstance(openai_obj, OpenAIObject) and "usage" in openai_obj:
+            update_usage_with_response(openai_obj)
+        return openai_obj
+
+    def metered_func(*args, **kwargs):
+        with patch.object(
+            engine_api_resource.util,
+            "convert_to_openai_object",
+            side_effect=metering_wrapper,
+        ):
+            return func(*args, **kwargs)
+
+    return metered_func
+
+
+def retry_openai_api(
+    num_retries: int = 10,
+    backoff_base: float = 2.0,
+    warn_user: bool = True,
+):
+    """Retry an OpenAI API call.
+
+    Args:
+        num_retries int: Number of retries. Defaults to 10.
+        backoff_base float: Base for exponential backoff. Defaults to 2.
+        warn_user bool: Whether to warn the user. Defaults to True.
+    """
+    retry_limit_msg = f"{Fore.RED}Error: " f"Reached rate limit, passing...{Fore.RESET}"
+    api_key_error_msg = (
+        f"Please double check that you have setup a "
+        f"{Fore.CYAN + Style.BRIGHT}PAID{Style.RESET_ALL} OpenAI API Account. You can "
+        f"read more here: {Fore.CYAN}https://docs.agpt.co/setup/#getting-an-api-key{Fore.RESET}"
+    )
+    backoff_msg = (
+        f"{Fore.RED}Error: API Bad gateway. Waiting {{backoff}} seconds...{Fore.RESET}"
+    )
+
+    def _wrapper(func):
+        @functools.wraps(func)
+        def _wrapped(*args, **kwargs):
+            user_warned = not warn_user
+            num_attempts = num_retries + 1  # +1 for the first attempt
+            for attempt in range(1, num_attempts + 1):
+                try:
+                    return func(*args, **kwargs)
+
+                except RateLimitError:
+                    if attempt == num_attempts:
+                        raise
+
+                    logger.debug(retry_limit_msg)
+                    if not user_warned:
+                        logger.double_check(api_key_error_msg)
+                        user_warned = True
+
+                except APIError as e:
+                    if (e.http_status not in [502, 429]) or (attempt == num_attempts):
+                        raise
+
+                backoff = backoff_base ** (attempt + 2)
+                logger.debug(backoff_msg.format(backoff=backoff))
+                time.sleep(backoff)
+
+        return _wrapped
+
+    return _wrapper
+
+
+def call_ai_function(
+    function: str,
+    args: list,
+    description: str,
+    model: str | None = None,
+    config: Config = None,
+) -> str:
+    """Call an AI function
+
+    This is a magic function that can do anything with no-code. See
+    https://github.com/Torantulino/AI-Functions for more info.
+
+    Args:
+        function (str): The function to call
+        args (list): The arguments to pass to the function
+        description (str): The description of the function
+        model (str, optional): The model to use. Defaults to None.
+
+    Returns:
+        str: The response from the function
+    """
+    if model is None:
+        model = config.smart_llm_model
+    # For each arg, if any are None, convert to "None":
+    args = [str(arg) if arg is not None else "None" for arg in args]
+    # parse args to comma separated string
+    arg_str: str = ", ".join(args)
+
+    prompt = ChatSequence.for_model(
+        model,
+        [
+            Message(
+                "system",
+                f"You are now the following python function: ```# {description}"
+                f"\n{function}```\n\nOnly respond with your `return` value.",
+            ),
+            Message("user", arg_str),
+        ],
+    )
+    return create_chat_completion(prompt=prompt, temperature=0)
+
+
+@metered
+@retry_openai_api()
+def create_text_completion(
+    prompt: str,
+    model: Optional[str],
+    temperature: Optional[float],
+    max_output_tokens: Optional[int],
+) -> str:
+    cfg = Config()
+    if model is None:
+        model = cfg.fast_llm_model
+    if temperature is None:
+        temperature = cfg.temperature
+
+    if cfg.use_azure:
+        kwargs = {"deployment_id": cfg.get_azure_deployment_id_for_model(model)}
+    else:
+        kwargs = {"model": model}
+
+    response = openai.Completion.create(
+        **kwargs,
+        prompt=prompt,
+        temperature=temperature,
+        max_tokens=max_output_tokens,
+        api_key=cfg.openai_api_key,
+    )
+    return response.choices[0].text
+
+
+# Overly simple abstraction until we create something better
+# simple retry mechanism when getting a rate error or a bad gateway
+@metered
+@retry_openai_api()
+def create_chat_completion(
+    prompt: ChatSequence,
+    model: Optional[str] = None,
+    temperature: float = None,
+    max_tokens: Optional[int] = None,
+) -> str:
+    """Create a chat completion using the OpenAI API
+
+    Args:
+        messages (List[Message]): The messages to send to the chat completion
+        model (str, optional): The model to use. Defaults to None.
+        temperature (float, optional): The temperature to use. Defaults to 0.9.
+        max_tokens (int, optional): The max tokens to use. Defaults to None.
+
+    Returns:
+        str: The response from the chat completion
+    """
+    cfg = Config()
+    if model is None:
+        model = prompt.model.name
+    if temperature is None:
+        temperature = cfg.temperature
+
+    logger.debug(
+        f"{Fore.GREEN}Creating chat completion with model {model}, temperature {temperature}, max_tokens {max_tokens}{Fore.RESET}"
+    )
+    for plugin in cfg.plugins:
+        if plugin.can_handle_chat_completion(
+            messages=prompt.raw(),
+            model=model,
+            temperature=temperature,
+            max_tokens=max_tokens,
+        ):
+            message = plugin.handle_chat_completion(
+                messages=prompt.raw(),
+                model=model,
+                temperature=temperature,
+                max_tokens=max_tokens,
+            )
+            if message is not None:
+                return message
+    api_manager = ApiManager()
+    response = None
+
+    if cfg.use_azure:
+        kwargs = {"deployment_id": cfg.get_azure_deployment_id_for_model(model)}
+    else:
+        kwargs = {"model": model}
+
+    response = api_manager.create_chat_completion(
+        **kwargs,
+        messages=prompt.raw(),
+        temperature=temperature,
+        max_tokens=max_tokens,
+    )
+
+    resp = response.choices[0].message["content"]
+    for plugin in cfg.plugins:
+        if not plugin.can_handle_on_response():
+            continue
+        resp = plugin.on_response(resp)
+    return resp
+
+
+def check_model(
+    model_name: str, model_type: Literal["smart_llm_model", "fast_llm_model"]
+) -> str:
+    """Check if model is available for use. If not, return gpt-3.5-turbo."""
+    api_manager = ApiManager()
+    models = api_manager.get_models()
+
+    if any(model_name in m["id"] for m in models):
+        return model_name
+
+    logger.typewriter_log(
+        "WARNING: ",
+        Fore.YELLOW,
+        f"You do not have access to {model_name}. Setting {model_type} to "
+        f"gpt-3.5-turbo.",
+    )
+    return "gpt-3.5-turbo"

autogpt/llm/utils/token_counter.py

@@ -0,0 +1,76 @@
+"""Functions for counting the number of tokens in a message or string."""
+from __future__ import annotations
+
+from typing import List
+
+import tiktoken
+
+from autogpt.llm.base import Message
+from autogpt.logs import logger
+
+
+def count_message_tokens(
+    messages: List[Message], model: str = "gpt-3.5-turbo-0301"
+) -> int:
+    """
+    Returns the number of tokens used by a list of messages.
+
+    Args:
+        messages (list): A list of messages, each of which is a dictionary
+            containing the role and content of the message.
+        model (str): The name of the model to use for tokenization.
+            Defaults to "gpt-3.5-turbo-0301".
+
+    Returns:
+        int: The number of tokens used by the list of messages.
+    """
+    try:
+        encoding = tiktoken.encoding_for_model(model)
+    except KeyError:
+        logger.warn("Warning: model not found. Using cl100k_base encoding.")
+        encoding = tiktoken.get_encoding("cl100k_base")
+    if model == "gpt-3.5-turbo":
+        # !Note: gpt-3.5-turbo may change over time.
+        # Returning num tokens assuming gpt-3.5-turbo-0301.")
+        return count_message_tokens(messages, model="gpt-3.5-turbo-0301")
+    elif model == "gpt-4":
+        # !Note: gpt-4 may change over time. Returning num tokens assuming gpt-4-0314.")
+        return count_message_tokens(messages, model="gpt-4-0314")
+    elif model == "gpt-3.5-turbo-0301":
+        tokens_per_message = (
+            4  # every message follows <|start|>{role/name}\n{content}<|end|>\n
+        )
+        tokens_per_name = -1  # if there's a name, the role is omitted
+    elif model == "gpt-4-0314":
+        tokens_per_message = 3
+        tokens_per_name = 1
+    else:
+        raise NotImplementedError(
+            f"num_tokens_from_messages() is not implemented for model {model}.\n"
+            " See https://github.com/openai/openai-python/blob/main/chatml.md for"
+            " information on how messages are converted to tokens."
+        )
+    num_tokens = 0
+    for message in messages:
+        num_tokens += tokens_per_message
+        for key, value in message.raw().items():
+            num_tokens += len(encoding.encode(value))
+            if key == "name":
+                num_tokens += tokens_per_name
+    num_tokens += 3  # every reply is primed with <|start|>assistant<|message|>
+    return num_tokens
+
+
+def count_string_tokens(string: str, model_name: str) -> int:
+    """
+    Returns the number of tokens in a text string.
+
+    Args:
+        string (str): The text string.
+        model_name (str): The name of the encoding to use. (e.g., "gpt-3.5-turbo")
+
+    Returns:
+        int: The number of tokens in the text string.
+    """
+    encoding = tiktoken.encoding_for_model(model_name)
+    return len(encoding.encode(string))

autogpt/log_cycle/json_handler.py

@@ -0,0 +1,20 @@
+import json
+import logging
+
+
+class JsonFileHandler(logging.FileHandler):
+    def __init__(self, filename, mode="a", encoding=None, delay=False):
+        super().__init__(filename, mode, encoding, delay)
+
+    def emit(self, record):
+        json_data = json.loads(self.format(record))
+        with open(self.baseFilename, "w", encoding="utf-8") as f:
+            json.dump(json_data, f, ensure_ascii=False, indent=4)
+
+
+import logging
+
+
+class JsonFormatter(logging.Formatter):
+    def format(self, record):
+        return record.msg

autogpt/log_cycle/log_cycle.py

@@ -0,0 +1,85 @@
+import json
+import os
+from typing import Any, Dict, Union
+
+from autogpt.logs import logger
+
+DEFAULT_PREFIX = "agent"
+FULL_MESSAGE_HISTORY_FILE_NAME = "full_message_history.json"
+CURRENT_CONTEXT_FILE_NAME = "current_context.json"
+NEXT_ACTION_FILE_NAME = "next_action.json"
+PROMPT_SUMMARY_FILE_NAME = "prompt_summary.json"
+SUMMARY_FILE_NAME = "summary.txt"
+SUPERVISOR_FEEDBACK_FILE_NAME = "supervisor_feedback.txt"
+PROMPT_SUPERVISOR_FEEDBACK_FILE_NAME = "prompt_supervisor_feedback.json"
+USER_INPUT_FILE_NAME = "user_input.txt"
+
+
+class LogCycleHandler:
+    """
+    A class for logging cycle data.
+    """
+
+    def __init__(self):
+        self.log_count_within_cycle = 0
+
+    @staticmethod
+    def create_directory_if_not_exists(directory_path: str) -> None:
+        if not os.path.exists(directory_path):
+            os.makedirs(directory_path, exist_ok=True)
+
+    def create_outer_directory(self, ai_name: str, created_at: str) -> str:
+        log_directory = logger.get_log_directory()
+
+        if os.environ.get("OVERWRITE_DEBUG") == "1":
+            outer_folder_name = "auto_gpt"
+        else:
+            ai_name_short = ai_name[:15] if ai_name else DEFAULT_PREFIX
+            outer_folder_name = f"{created_at}_{ai_name_short}"
+
+        outer_folder_path = os.path.join(log_directory, "DEBUG", outer_folder_name)
+        self.create_directory_if_not_exists(outer_folder_path)
+
+        return outer_folder_path
+
+    def create_inner_directory(self, outer_folder_path: str, cycle_count: int) -> str:
+        nested_folder_name = str(cycle_count).zfill(3)
+        nested_folder_path = os.path.join(outer_folder_path, nested_folder_name)
+        self.create_directory_if_not_exists(nested_folder_path)
+
+        return nested_folder_path
+
+    def create_nested_directory(
+        self, ai_name: str, created_at: str, cycle_count: int
+    ) -> str:
+        outer_folder_path = self.create_outer_directory(ai_name, created_at)
+        nested_folder_path = self.create_inner_directory(outer_folder_path, cycle_count)
+
+        return nested_folder_path
+
+    def log_cycle(
+        self,
+        ai_name: str,
+        created_at: str,
+        cycle_count: int,
+        data: Union[Dict[str, Any], Any],
+        file_name: str,
+    ) -> None:
+        """
+        Log cycle data to a JSON file.
+
+        Args:
+            data (Any): The data to be logged.
+            file_name (str): The name of the file to save the logged data.
+        """
+        nested_folder_path = self.create_nested_directory(
+            ai_name, created_at, cycle_count
+        )
+
+        json_data = json.dumps(data, ensure_ascii=False, indent=4)
+        log_file_path = os.path.join(
+            nested_folder_path, f"{self.log_count_within_cycle}_{file_name}"
+        )
+
+        logger.log_json(json_data, log_file_path)
+        self.log_count_within_cycle += 1

autogpt/logs.py

@@ -0,0 +1,294 @@
+"""Logging module for Auto-GPT."""
+import logging
+import os
+import random
+import re
+import time
+from logging import LogRecord
+from typing import Any
+
+from colorama import Fore, Style
+
+from autogpt.log_cycle.json_handler import JsonFileHandler, JsonFormatter
+from autogpt.singleton import Singleton
+from autogpt.speech import say_text
+
+
+class Logger(metaclass=Singleton):
+    """
+    Logger that handle titles in different colors.
+    Outputs logs in console, activity.log, and errors.log
+    For console handler: simulates typing
+    """
+
+    def __init__(self):
+        # create log directory if it doesn't exist
+        this_files_dir_path = os.path.dirname(__file__)
+        log_dir = os.path.join(this_files_dir_path, "../logs")
+        if not os.path.exists(log_dir):
+            os.makedirs(log_dir)
+
+        log_file = "activity.log"
+        error_file = "error.log"
+
+        console_formatter = AutoGptFormatter("%(title_color)s %(message)s")
+
+        # Create a handler for console which simulate typing
+        self.typing_console_handler = TypingConsoleHandler()
+        self.typing_console_handler.setLevel(logging.INFO)
+        self.typing_console_handler.setFormatter(console_formatter)
+
+        # Create a handler for console without typing simulation
+        self.console_handler = ConsoleHandler()
+        self.console_handler.setLevel(logging.DEBUG)
+        self.console_handler.setFormatter(console_formatter)
+
+        # Info handler in activity.log
+        self.file_handler = logging.FileHandler(
+            os.path.join(log_dir, log_file), "a", "utf-8"
+        )
+        self.file_handler.setLevel(logging.DEBUG)
+        info_formatter = AutoGptFormatter(
+            "%(asctime)s %(levelname)s %(title)s %(message_no_color)s"
+        )
+        self.file_handler.setFormatter(info_formatter)
+
+        # Error handler error.log
+        error_handler = logging.FileHandler(
+            os.path.join(log_dir, error_file), "a", "utf-8"
+        )
+        error_handler.setLevel(logging.ERROR)
+        error_formatter = AutoGptFormatter(
+            "%(asctime)s %(levelname)s %(module)s:%(funcName)s:%(lineno)d %(title)s"
+            " %(message_no_color)s"
+        )
+        error_handler.setFormatter(error_formatter)
+
+        self.typing_logger = logging.getLogger("TYPER")
+        self.typing_logger.addHandler(self.typing_console_handler)
+        self.typing_logger.addHandler(self.file_handler)
+        self.typing_logger.addHandler(error_handler)
+        self.typing_logger.setLevel(logging.DEBUG)
+
+        self.logger = logging.getLogger("LOGGER")
+        self.logger.addHandler(self.console_handler)
+        self.logger.addHandler(self.file_handler)
+        self.logger.addHandler(error_handler)
+        self.logger.setLevel(logging.DEBUG)
+
+        self.json_logger = logging.getLogger("JSON_LOGGER")
+        self.json_logger.addHandler(self.file_handler)
+        self.json_logger.addHandler(error_handler)
+        self.json_logger.setLevel(logging.DEBUG)
+
+        self.speak_mode = False
+        self.chat_plugins = []
+
+    def typewriter_log(
+        self, title="", title_color="", content="", speak_text=False, level=logging.INFO
+    ):
+        if speak_text and self.speak_mode:
+            say_text(f"{title}. {content}")
+
+        for plugin in self.chat_plugins:
+            plugin.report(f"{title}. {content}")
+
+        if content:
+            if isinstance(content, list):
+                content = " ".join(content)
+        else:
+            content = ""
+
+        self.typing_logger.log(
+            level, content, extra={"title": title, "color": title_color}
+        )
+
+    def debug(
+        self,
+        message,
+        title="",
+        title_color="",
+    ):
+        self._log(title, title_color, message, logging.DEBUG)
+
+    def info(
+        self,
+        message,
+        title="",
+        title_color="",
+    ):
+        self._log(title, title_color, message, logging.INFO)
+
+    def warn(
+        self,
+        message,
+        title="",
+        title_color="",
+    ):
+        self._log(title, title_color, message, logging.WARN)
+
+    def error(self, title, message=""):
+        self._log(title, Fore.RED, message, logging.ERROR)
+
+    def _log(
+        self,
+        title: str = "",
+        title_color: str = "",
+        message: str = "",
+        level=logging.INFO,
+    ):
+        if message:
+            if isinstance(message, list):
+                message = " ".join(message)
+        self.logger.log(
+            level, message, extra={"title": str(title), "color": str(title_color)}
+        )
+
+    def set_level(self, level):
+        self.logger.setLevel(level)
+        self.typing_logger.setLevel(level)
+
+    def double_check(self, additionalText=None):
+        if not additionalText:
+            additionalText = (
+                "Please ensure you've setup and configured everything"
+                " correctly. Read https://github.com/Torantulino/Auto-GPT#readme to "
+                "double check. You can also create a github issue or join the discord"
+                " and ask there!"
+            )
+
+        self.typewriter_log("DOUBLE CHECK CONFIGURATION", Fore.YELLOW, additionalText)
+
+    def log_json(self, data: Any, file_name: str) -> None:
+        # Define log directory
+        this_files_dir_path = os.path.dirname(__file__)
+        log_dir = os.path.join(this_files_dir_path, "../logs")
+
+        # Create a handler for JSON files
+        json_file_path = os.path.join(log_dir, file_name)
+        json_data_handler = JsonFileHandler(json_file_path)
+        json_data_handler.setFormatter(JsonFormatter())
+
+        # Log the JSON data using the custom file handler
+        self.json_logger.addHandler(json_data_handler)
+        self.json_logger.debug(data)
+        self.json_logger.removeHandler(json_data_handler)
+
+    def get_log_directory(self):
+        this_files_dir_path = os.path.dirname(__file__)
+        log_dir = os.path.join(this_files_dir_path, "../logs")
+        return os.path.abspath(log_dir)
+
+
+"""
+Output stream to console using simulated typing
+"""
+
+
+class TypingConsoleHandler(logging.StreamHandler):
+    def emit(self, record):
+        min_typing_speed = 0.05
+        max_typing_speed = 0.01
+
+        msg = self.format(record)
+        try:
+            words = msg.split()
+            for i, word in enumerate(words):
+                print(word, end="", flush=True)
+                if i < len(words) - 1:
+                    print(" ", end="", flush=True)
+                typing_speed = random.uniform(min_typing_speed, max_typing_speed)
+                time.sleep(typing_speed)
+                # type faster after each word
+                min_typing_speed = min_typing_speed * 0.95
+                max_typing_speed = max_typing_speed * 0.95
+            print()
+        except Exception:
+            self.handleError(record)
+
+
+class ConsoleHandler(logging.StreamHandler):
+    def emit(self, record) -> None:
+        msg = self.format(record)
+        try:
+            print(msg)
+        except Exception:
+            self.handleError(record)
+
+
+class AutoGptFormatter(logging.Formatter):
+    """
+    Allows to handle custom placeholders 'title_color' and 'message_no_color'.
+    To use this formatter, make sure to pass 'color', 'title' as log extras.
+    """
+
+    def format(self, record: LogRecord) -> str:
+        if hasattr(record, "color"):
+            record.title_color = (
+                getattr(record, "color")
+                + getattr(record, "title", "")
+                + " "
+                + Style.RESET_ALL
+            )
+        else:
+            record.title_color = getattr(record, "title", "")
+
+        # Add this line to set 'title' to an empty string if it doesn't exist
+        record.title = getattr(record, "title", "")
+
+        if hasattr(record, "msg"):
+            record.message_no_color = remove_color_codes(getattr(record, "msg"))
+        else:
+            record.message_no_color = ""
+        return super().format(record)
+
+
+def remove_color_codes(s: str) -> str:
+    ansi_escape = re.compile(r"\x1B(?:[@-Z\\-_]|\[[0-?]*[ -/]*[@-~])")
+    return ansi_escape.sub("", s)
+
+
+logger = Logger()
+
+
+def print_assistant_thoughts(
+    ai_name: object,
+    assistant_reply_json_valid: object,
+    speak_mode: bool = False,
+) -> None:
+    assistant_thoughts_reasoning = None
+    assistant_thoughts_plan = None
+    assistant_thoughts_speak = None
+    assistant_thoughts_criticism = None
+
+    assistant_thoughts = assistant_reply_json_valid.get("thoughts", {})
+    assistant_thoughts_text = assistant_thoughts.get("text")
+    if assistant_thoughts:
+        assistant_thoughts_reasoning = assistant_thoughts.get("reasoning")
+        assistant_thoughts_plan = assistant_thoughts.get("plan")
+        assistant_thoughts_criticism = assistant_thoughts.get("criticism")
+        assistant_thoughts_speak = assistant_thoughts.get("speak")
+    logger.typewriter_log(
+        f"{ai_name.upper()} THOUGHTS:", Fore.YELLOW, f"{assistant_thoughts_text}"
+    )
+    logger.typewriter_log("REASONING:", Fore.YELLOW, f"{assistant_thoughts_reasoning}")
+    if assistant_thoughts_plan:
+        logger.typewriter_log("PLAN:", Fore.YELLOW, "")
+        # If it's a list, join it into a string
+        if isinstance(assistant_thoughts_plan, list):
+            assistant_thoughts_plan = "\n".join(assistant_thoughts_plan)
+        elif isinstance(assistant_thoughts_plan, dict):
+            assistant_thoughts_plan = str(assistant_thoughts_plan)
+
+        # Split the input_string using the newline character and dashes
+        lines = assistant_thoughts_plan.split("\n")
+        for line in lines:
+            line = line.lstrip("- ")
+            logger.typewriter_log("- ", Fore.GREEN, line.strip())
+    logger.typewriter_log("CRITICISM:", Fore.YELLOW, f"{assistant_thoughts_criticism}")
+    # Speak the assistant's thoughts
+    if assistant_thoughts_speak:
+        if speak_mode:
+            say_text(assistant_thoughts_speak)
+        else:
+            logger.typewriter_log("SPEAK:", Fore.YELLOW, f"{assistant_thoughts_speak}")

autogpt/main.py

@@ -0,0 +1,194 @@
+"""The application entry point.  Can be invoked by a CLI or any other front end application."""
+import logging
+import sys
+from pathlib import Path
+
+from colorama import Fore, Style
+
+from autogpt.agent import Agent
+from autogpt.commands.command import CommandRegistry
+from autogpt.config import Config, check_openai_api_key
+from autogpt.configurator import create_config
+from autogpt.logs import logger
+from autogpt.memory.vector import get_memory
+from autogpt.plugins import scan_plugins
+from autogpt.prompts.prompt import DEFAULT_TRIGGERING_PROMPT, construct_main_ai_config
+from autogpt.utils import (
+    get_current_git_branch,
+    get_latest_bulletin,
+    get_legal_warning,
+    markdown_to_ansi_style,
+)
+from autogpt.workspace import Workspace
+from scripts.install_plugin_deps import install_plugin_dependencies
+
+
+def run_auto_gpt(
+    continuous: bool,
+    continuous_limit: int,
+    ai_settings: str,
+    prompt_settings: str,
+    skip_reprompt: bool,
+    speak: bool,
+    debug: bool,
+    gpt3only: bool,
+    gpt4only: bool,
+    memory_type: str,
+    browser_name: str,
+    allow_downloads: bool,
+    skip_news: bool,
+    workspace_directory: str,
+    install_plugin_deps: bool,
+):
+    # Configure logging before we do anything else.
+    logger.set_level(logging.DEBUG if debug else logging.INFO)
+    logger.speak_mode = speak
+
+    cfg = Config()
+    # TODO: fill in llm values here
+    check_openai_api_key()
+
+    create_config(
+        cfg,
+        continuous,
+        continuous_limit,
+        ai_settings,
+        prompt_settings,
+        skip_reprompt,
+        speak,
+        debug,
+        gpt3only,
+        gpt4only,
+        memory_type,
+        browser_name,
+        allow_downloads,
+        skip_news,
+    )
+
+    if cfg.continuous_mode:
+        for line in get_legal_warning().split("\n"):
+            logger.warn(markdown_to_ansi_style(line), "LEGAL:", Fore.RED)
+
+    if not cfg.skip_news:
+        motd, is_new_motd = get_latest_bulletin()
+        if motd:
+            motd = markdown_to_ansi_style(motd)
+            for motd_line in motd.split("\n"):
+                logger.info(motd_line, "NEWS:", Fore.GREEN)
+            if is_new_motd and not cfg.chat_messages_enabled:
+                input(
+                    Fore.MAGENTA
+                    + Style.BRIGHT
+                    + "NEWS: Bulletin was updated! Press Enter to continue..."
+                    + Style.RESET_ALL
+                )
+
+        git_branch = get_current_git_branch()
+        if git_branch and git_branch != "stable":
+            logger.typewriter_log(
+                "WARNING: ",
+                Fore.RED,
+                f"You are running on `{git_branch}` branch "
+                "- this is not a supported branch.",
+            )
+        if sys.version_info < (3, 10):
+            logger.typewriter_log(
+                "WARNING: ",
+                Fore.RED,
+                "You are running on an older version of Python. "
+                "Some people have observed problems with certain "
+                "parts of Auto-GPT with this version. "
+                "Please consider upgrading to Python 3.10 or higher.",
+            )
+
+    if install_plugin_deps:
+        install_plugin_dependencies()
+
+    # TODO: have this directory live outside the repository (e.g. in a user's
+    #   home directory) and have it come in as a command line argument or part of
+    #   the env file.
+    if workspace_directory is None:
+        workspace_directory = Path(__file__).parent / "auto_gpt_workspace"
+    else:
+        workspace_directory = Path(workspace_directory)
+    # TODO: pass in the ai_settings file and the env file and have them cloned into
+    #   the workspace directory so we can bind them to the agent.
+    workspace_directory = Workspace.make_workspace(workspace_directory)
+    cfg.workspace_path = str(workspace_directory)
+
+    # HACK: doing this here to collect some globals that depend on the workspace.
+    file_logger_path = workspace_directory / "file_logger.txt"
+    if not file_logger_path.exists():
+        with file_logger_path.open(mode="w", encoding="utf-8") as f:
+            f.write("File Operation Logger ")
+
+    cfg.file_logger_path = str(file_logger_path)
+
+    cfg.set_plugins(scan_plugins(cfg, cfg.debug_mode))
+    # Create a CommandRegistry instance and scan default folder
+    command_registry = CommandRegistry()
+
+    command_categories = [
+        "autogpt.commands.analyze_code",
+        "autogpt.commands.audio_text",
+        "autogpt.commands.execute_code",
+        "autogpt.commands.file_operations",
+        "autogpt.commands.git_operations",
+        "autogpt.commands.google_search",
+        "autogpt.commands.image_gen",
+        "autogpt.commands.improve_code",
+        "autogpt.commands.web_selenium",
+        "autogpt.commands.write_tests",
+        "autogpt.app",
+        "autogpt.commands.task_statuses",
+    ]
+    logger.debug(
+        f"The following command categories are disabled: {cfg.disabled_command_categories}"
+    )
+    command_categories = [
+        x for x in command_categories if x not in cfg.disabled_command_categories
+    ]
+
+    logger.debug(f"The following command categories are enabled: {command_categories}")
+
+    for command_category in command_categories:
+        command_registry.import_commands(command_category)
+
+    ai_name = ""
+    ai_config = construct_main_ai_config()
+    ai_config.command_registry = command_registry
+    if ai_config.ai_name:
+        ai_name = ai_config.ai_name
+    # print(prompt)
+    # Initialize variables
+    next_action_count = 0
+
+    # add chat plugins capable of report to logger
+    if cfg.chat_messages_enabled:
+        for plugin in cfg.plugins:
+            if hasattr(plugin, "can_handle_report") and plugin.can_handle_report():
+                logger.info(f"Loaded plugin into logger: {plugin.__class__.__name__}")
+                logger.chat_plugins.append(plugin)
+
+    # Initialize memory and make sure it is empty.
+    # this is particularly important for indexing and referencing pinecone memory
+    memory = get_memory(cfg, init=True)
+    logger.typewriter_log(
+        "Using memory of type:", Fore.GREEN, f"{memory.__class__.__name__}"
+    )
+    logger.typewriter_log("Using Browser:", Fore.GREEN, cfg.selenium_web_browser)
+    system_prompt = ai_config.construct_full_prompt()
+    if cfg.debug_mode:
+        logger.typewriter_log("Prompt:", Fore.GREEN, system_prompt)
+
+    agent = Agent(
+        ai_name=ai_name,
+        memory=memory,
+        next_action_count=next_action_count,
+        command_registry=command_registry,
+        config=ai_config,
+        system_prompt=system_prompt,
+        triggering_prompt=DEFAULT_TRIGGERING_PROMPT,
+        workspace_directory=workspace_directory,
+    )
+    agent.start_interaction_loop()

autogpt/memory/message_history.py

@@ -0,0 +1,204 @@
+from __future__ import annotations
+
+import copy
+import json
+from dataclasses import dataclass, field
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from autogpt.agent import Agent
+
+from autogpt.config import Config
+from autogpt.json_utils.utilities import (
+    LLM_DEFAULT_RESPONSE_FORMAT,
+    is_string_valid_json,
+)
+from autogpt.llm.base import ChatSequence, Message, MessageRole, MessageType
+from autogpt.llm.utils import create_chat_completion
+from autogpt.log_cycle.log_cycle import PROMPT_SUMMARY_FILE_NAME, SUMMARY_FILE_NAME
+from autogpt.logs import logger
+
+
+@dataclass
+class MessageHistory:
+    agent: Agent
+
+    messages: list[Message] = field(default_factory=list)
+    summary: str = "I was created"
+
+    last_trimmed_index: int = 0
+
+    def __getitem__(self, i: int):
+        return self.messages[i]
+
+    def __iter__(self):
+        return iter(self.messages)
+
+    def __len__(self):
+        return len(self.messages)
+
+    def add(
+        self,
+        role: MessageRole,
+        content: str,
+        type: MessageType | None = None,
+    ):
+        return self.append(Message(role, content, type))
+
+    def append(self, message: Message):
+        return self.messages.append(message)
+
+    def trim_messages(
+        self,
+        current_message_chain: list[Message],
+    ) -> tuple[Message, list[Message]]:
+        """
+        Returns a list of trimmed messages: messages which are in the message history
+        but not in current_message_chain.
+
+        Args:
+            current_message_chain (list[Message]): The messages currently in the context.
+
+        Returns:
+            Message: A message with the new running summary after adding the trimmed messages.
+            list[Message]: A list of messages that are in full_message_history with an index higher than last_trimmed_index and absent from current_message_chain.
+        """
+        # Select messages in full_message_history with an index higher than last_trimmed_index
+        new_messages = [
+            msg for i, msg in enumerate(self) if i > self.last_trimmed_index
+        ]
+
+        # Remove messages that are already present in current_message_chain
+        new_messages_not_in_chain = [
+            msg for msg in new_messages if msg not in current_message_chain
+        ]
+
+        if not new_messages_not_in_chain:
+            return self.summary_message(), []
+
+        new_summary_message = self.update_running_summary(
+            new_events=new_messages_not_in_chain
+        )
+
+        # Find the index of the last message processed
+        last_message = new_messages_not_in_chain[-1]
+        self.last_trimmed_index = self.messages.index(last_message)
+
+        return new_summary_message, new_messages_not_in_chain
+
+    def per_cycle(self, messages: list[Message] | None = None):
+        """
+        Yields:
+            Message: a message containing user input
+            Message: a message from the AI containing a proposed action
+            Message: the message containing the result of the AI's proposed action
+        """
+        messages = messages or self.messages
+        for i in range(0, len(messages) - 1):
+            ai_message = messages[i]
+            if ai_message.type != "ai_response":
+                continue
+            user_message = (
+                messages[i - 1] if i > 0 and messages[i - 1].role == "user" else None
+            )
+            result_message = messages[i + 1]
+            try:
+                assert is_string_valid_json(
+                    ai_message.content, LLM_DEFAULT_RESPONSE_FORMAT
+                ), "AI response is not a valid JSON object"
+                assert result_message.type == "action_result"
+
+                yield user_message, ai_message, result_message
+            except AssertionError as err:
+                logger.debug(
+                    f"Invalid item in message history: {err}; Messages: {messages[i-1:i+2]}"
+                )
+
+    def summary_message(self) -> Message:
+        return Message(
+            "system",
+            f"This reminds you of these events from your past: \n{self.summary}",
+        )
+
+    def update_running_summary(self, new_events: list[Message]) -> Message:
+        """
+        This function takes a list of dictionaries representing new events and combines them with the current summary,
+        focusing on key and potentially important information to remember. The updated summary is returned in a message
+        formatted in the 1st person past tense.
+
+        Args:
+            new_events (List[Dict]): A list of dictionaries containing the latest events to be added to the summary.
+
+        Returns:
+            str: A message containing the updated summary of actions, formatted in the 1st person past tense.
+
+        Example:
+            new_events = [{"event": "entered the kitchen."}, {"event": "found a scrawled note with the number 7"}]
+            update_running_summary(new_events)
+            # Returns: "This reminds you of these events from your past: \nI entered the kitchen and found a scrawled note saying 7."
+        """
+        cfg = Config()
+
+        if not new_events:
+            return self.summary_message()
+
+        # Create a copy of the new_events list to prevent modifying the original list
+        new_events = copy.deepcopy(new_events)
+
+        # Replace "assistant" with "you". This produces much better first person past tense results.
+        for event in new_events:
+            if event.role.lower() == "assistant":
+                event.role = "you"
+
+                # Remove "thoughts" dictionary from "content"
+                try:
+                    content_dict = json.loads(event.content)
+                    if "thoughts" in content_dict:
+                        del content_dict["thoughts"]
+                    event.content = json.dumps(content_dict)
+                except json.decoder.JSONDecodeError:
+                    if cfg.debug_mode:
+                        logger.error(f"Error: Invalid JSON: {event.content}\n")
+
+            elif event.role.lower() == "system":
+                event.role = "your computer"
+
+            # Delete all user messages
+            elif event.role == "user":
+                new_events.remove(event)
+
+        prompt = f'''Your task is to create a concise running summary of actions and information results in the provided text, focusing on key and potentially important information to remember.
+
+You will receive the current summary and the your latest actions. Combine them, adding relevant key information from the latest development in 1st person past tense and keeping the summary concise.
+
+Summary So Far:
+"""
+{self.summary}
+"""
+
+Latest Development:
+"""
+{new_events or "Nothing new happened."}
+"""
+'''
+
+        prompt = ChatSequence.for_model(cfg.fast_llm_model, [Message("user", prompt)])
+        self.agent.log_cycle_handler.log_cycle(
+            self.agent.config.ai_name,
+            self.agent.created_at,
+            self.agent.cycle_count,
+            prompt.raw(),
+            PROMPT_SUMMARY_FILE_NAME,
+        )
+
+        self.summary = create_chat_completion(prompt)
+
+        self.agent.log_cycle_handler.log_cycle(
+            self.agent.config.ai_name,
+            self.agent.created_at,
+            self.agent.cycle_count,
+            self.summary,
+            SUMMARY_FILE_NAME,
+        )
+
+        return self.summary_message()

autogpt/memory/vector/__init__.py

@@ -0,0 +1,138 @@
+from autogpt.config import Config
+from autogpt.logs import logger
+
+from .memory_item import MemoryItem, MemoryItemRelevance
+from .providers.base import VectorMemoryProvider as VectorMemory
+from .providers.json_file import JSONFileMemory
+from .providers.no_memory import NoMemory
+
+# List of supported memory backends
+# Add a backend to this list if the import attempt is successful
+supported_memory = ["json_file", "no_memory"]
+
+# try:
+#     from .providers.redis import RedisMemory
+
+#     supported_memory.append("redis")
+# except ImportError:
+#     RedisMemory = None
+
+# try:
+#     from .providers.pinecone import PineconeMemory
+
+#     supported_memory.append("pinecone")
+# except ImportError:
+#     PineconeMemory = None
+
+# try:
+#     from .providers.weaviate import WeaviateMemory
+
+#     supported_memory.append("weaviate")
+# except ImportError:
+#     WeaviateMemory = None
+
+# try:
+#     from .providers.milvus import MilvusMemory
+
+#     supported_memory.append("milvus")
+# except ImportError:
+#     MilvusMemory = None
+
+
+def get_memory(cfg: Config, init=False) -> VectorMemory:
+    memory = None
+
+    match cfg.memory_backend:
+        case "json_file":
+            memory = JSONFileMemory(cfg)
+
+        case "pinecone":
+            raise NotImplementedError(
+                "The Pinecone memory backend has been rendered incompatible by work on "
+                "the memory system, and was removed. Whether support will be added back "
+                "in the future is subject to discussion, feel free to pitch in: "
+                "https://github.com/Significant-Gravitas/Auto-GPT/discussions/4280"
+            )
+            # if not PineconeMemory:
+            #     logger.warn(
+            #         "Error: Pinecone is not installed. Please install pinecone"
+            #         " to use Pinecone as a memory backend."
+            #     )
+            # else:
+            #     memory = PineconeMemory(cfg)
+            #     if init:
+            #         memory.clear()
+
+        case "redis":
+            raise NotImplementedError(
+                "The Redis memory backend has been rendered incompatible by work on "
+                "the memory system, and has been removed temporarily."
+            )
+            # if not RedisMemory:
+            #     logger.warn(
+            #         "Error: Redis is not installed. Please install redis-py to"
+            #         " use Redis as a memory backend."
+            #     )
+            # else:
+            #     memory = RedisMemory(cfg)
+
+        case "weaviate":
+            raise NotImplementedError(
+                "The Weaviate memory backend has been rendered incompatible by work on "
+                "the memory system, and was removed. Whether support will be added back "
+                "in the future is subject to discussion, feel free to pitch in: "
+                "https://github.com/Significant-Gravitas/Auto-GPT/discussions/4280"
+            )
+            # if not WeaviateMemory:
+            #     logger.warn(
+            #         "Error: Weaviate is not installed. Please install weaviate-client to"
+            #         " use Weaviate as a memory backend."
+            #     )
+            # else:
+            #     memory = WeaviateMemory(cfg)
+
+        case "milvus":
+            raise NotImplementedError(
+                "The Milvus memory backend has been rendered incompatible by work on "
+                "the memory system, and was removed. Whether support will be added back "
+                "in the future is subject to discussion, feel free to pitch in: "
+                "https://github.com/Significant-Gravitas/Auto-GPT/discussions/4280"
+            )
+            # if not MilvusMemory:
+            #     logger.warn(
+            #         "Error: pymilvus sdk is not installed."
+            #         "Please install pymilvus to use Milvus or Zilliz Cloud as memory backend."
+            #     )
+            # else:
+            #     memory = MilvusMemory(cfg)
+
+        case "no_memory":
+            memory = NoMemory()
+
+        case _:
+            raise ValueError(
+                f"Unknown memory backend '{cfg.memory_backend}'. Please check your config."
+            )
+
+    if memory is None:
+        memory = JSONFileMemory(cfg)
+
+    return memory
+
+
+def get_supported_memory_backends():
+    return supported_memory
+
+
+__all__ = [
+    "get_memory",
+    "MemoryItem",
+    "MemoryItemRelevance",
+    "JSONFileMemory",
+    "NoMemory",
+    "VectorMemory",
+    # "RedisMemory",
+    # "PineconeMemory",
+    # "MilvusMemory",
+    # "WeaviateMemory",
+]

autogpt/memory/vector/memory_item.py

@@ -0,0 +1,223 @@
+from __future__ import annotations
+
+import dataclasses
+import json
+from typing import Literal
+
+import numpy as np
+
+from autogpt.config import Config
+from autogpt.llm import Message
+from autogpt.llm.utils import count_string_tokens
+from autogpt.logs import logger
+from autogpt.processing.text import chunk_content, split_text, summarize_text
+
+from .utils import Embedding, get_embedding
+
+MemoryDocType = Literal["webpage", "text_file", "code_file", "agent_history"]
+
+
+@dataclasses.dataclass
+class MemoryItem:
+    """Memory object containing raw content as well as embeddings"""
+
+    raw_content: str
+    summary: str
+    chunks: list[str]
+    chunk_summaries: list[str]
+    e_summary: Embedding
+    e_chunks: list[Embedding]
+    metadata: dict
+
+    def relevance_for(self, query: str, e_query: Embedding | None = None):
+        return MemoryItemRelevance.of(self, query, e_query)
+
+    @staticmethod
+    def from_text(
+        text: str,
+        source_type: MemoryDocType,
+        metadata: dict = {},
+        how_to_summarize: str | None = None,
+        question_for_summary: str | None = None,
+    ):
+        cfg = Config()
+        logger.debug(f"Memorizing text:\n{'-'*32}\n{text}\n{'-'*32}\n")
+
+        chunks = [
+            chunk
+            for chunk, _ in (
+                split_text(text, cfg.embedding_model)
+                if source_type != "code_file"
+                else chunk_content(text, cfg.embedding_model)
+            )
+        ]
+        logger.debug("Chunks: " + str(chunks))
+
+        chunk_summaries = [
+            summary
+            for summary, _ in [
+                summarize_text(
+                    text_chunk,
+                    instruction=how_to_summarize,
+                    question=question_for_summary,
+                )
+                for text_chunk in chunks
+            ]
+        ]
+        logger.debug("Chunk summaries: " + str(chunk_summaries))
+
+        e_chunks = get_embedding(chunks)
+
+        summary = (
+            chunk_summaries[0]
+            if len(chunks) == 1
+            else summarize_text(
+                "\n\n".join(chunk_summaries),
+                instruction=how_to_summarize,
+                question=question_for_summary,
+            )[0]
+        )
+        logger.debug("Total summary: " + summary)
+
+        # TODO: investigate search performance of weighted average vs summary
+        # e_average = np.average(e_chunks, axis=0, weights=[len(c) for c in chunks])
+        e_summary = get_embedding(summary)
+
+        metadata["source_type"] = source_type
+
+        return MemoryItem(
+            text,
+            summary,
+            chunks,
+            chunk_summaries,
+            e_summary,
+            e_chunks,
+            metadata=metadata,
+        )
+
+    @staticmethod
+    def from_text_file(content: str, path: str):
+        return MemoryItem.from_text(content, "text_file", {"location": path})
+
+    @staticmethod
+    def from_code_file(content: str, path: str):
+        # TODO: implement tailored code memories
+        return MemoryItem.from_text(content, "code_file", {"location": path})
+
+    @staticmethod
+    def from_ai_action(ai_message: Message, result_message: Message):
+        # The result_message contains either user feedback
+        # or the result of the command specified in ai_message
+
+        if ai_message["role"] != "assistant":
+            raise ValueError(f"Invalid role on 'ai_message': {ai_message['role']}")
+
+        result = (
+            result_message["content"]
+            if result_message["content"].startswith("Command")
+            else "None"
+        )
+        user_input = (
+            result_message["content"]
+            if result_message["content"].startswith("Human feedback")
+            else "None"
+        )
+        memory_content = (
+            f"Assistant Reply: {ai_message['content']}"
+            "\n\n"
+            f"Result: {result}"
+            "\n\n"
+            f"Human Feedback: {user_input}"
+        )
+
+        return MemoryItem.from_text(
+            text=memory_content,
+            source_type="agent_history",
+            how_to_summarize="if possible, also make clear the link between the command in the assistant's response and the command result. Do not mention the human feedback if there is none",
+        )
+
+    @staticmethod
+    def from_webpage(content: str, url: str, question: str | None = None):
+        return MemoryItem.from_text(
+            text=content,
+            source_type="webpage",
+            metadata={"location": url},
+            question_for_summary=question,
+        )
+
+    def dump(self) -> str:
+        token_length = count_string_tokens(self.raw_content, Config().embedding_model)
+        return f"""
+=============== MemoryItem ===============
+Length: {token_length} tokens in {len(self.e_chunks)} chunks
+Metadata: {json.dumps(self.metadata, indent=2)}
+---------------- SUMMARY -----------------
+{self.summary}
+------------------ RAW -------------------
+{self.raw_content}
+==========================================
+"""
+
+
+@dataclasses.dataclass
+class MemoryItemRelevance:
+    """
+    Class that encapsulates memory relevance search functionality and data.
+    Instances contain a MemoryItem and its relevance scores for a given query.
+    """
+
+    memory_item: MemoryItem
+    for_query: str
+    summary_relevance_score: float
+    chunk_relevance_scores: list[float]
+
+    @staticmethod
+    def of(
+        memory_item: MemoryItem, for_query: str, e_query: Embedding | None = None
+    ) -> MemoryItemRelevance:
+        e_query = e_query or get_embedding(for_query)
+        _, srs, crs = MemoryItemRelevance.calculate_scores(memory_item, e_query)
+        return MemoryItemRelevance(
+            for_query=for_query,
+            memory_item=memory_item,
+            summary_relevance_score=srs,
+            chunk_relevance_scores=crs,
+        )
+
+    @staticmethod
+    def calculate_scores(
+        memory: MemoryItem, compare_to: Embedding
+    ) -> tuple[float, float, list[float]]:
+        """
+        Calculates similarity between given embedding and all embeddings of the memory
+
+        Returns:
+            float: the aggregate (max) relevance score of the memory
+            float: the relevance score of the memory summary
+            list: the relevance scores of the memory chunks
+        """
+        summary_relevance_score = np.dot(memory.e_summary, compare_to)
+        chunk_relevance_scores = np.dot(memory.e_chunks, compare_to)
+        logger.debug(f"Relevance of summary: {summary_relevance_score}")
+        logger.debug(f"Relevance of chunks: {chunk_relevance_scores}")
+
+        relevance_scores = [summary_relevance_score, *chunk_relevance_scores]
+        logger.debug(f"Relevance scores: {relevance_scores}")
+        return max(relevance_scores), summary_relevance_score, chunk_relevance_scores
+
+    @property
+    def score(self) -> float:
+        """The aggregate relevance score of the memory item for the given query"""
+        return max([self.summary_relevance_score, *self.chunk_relevance_scores])
+
+    @property
+    def most_relevant_chunk(self) -> tuple[str, float]:
+        """The most relevant chunk of the memory item + its score for the given query"""
+        i_relmax = np.argmax(self.chunk_relevance_scores)
+        return self.memory_item.chunks[i_relmax], self.chunk_relevance_scores[i_relmax]
+
+    def __str__(self):
+        return (
+            f"{self.memory_item.summary} ({self.summary_relevance_score}) "
+            f"{self.chunk_relevance_scores}"
+        )

autogpt/memory/vector/providers/__init__.py

@@ -0,0 +1,7 @@
+from .json_file import JSONFileMemory
+from .no_memory import NoMemory
+
+__all__ = [
+    "JSONFileMemory",
+    "NoMemory",
+]

autogpt/memory/vector/providers/base.py

@@ -0,0 +1,74 @@
+import abc
+import functools
+from typing import MutableSet, Sequence
+
+import numpy as np
+
+from autogpt.config.config import Config
+from autogpt.logs import logger
+from autogpt.singleton import AbstractSingleton
+
+from .. import MemoryItem, MemoryItemRelevance
+from ..utils import Embedding, get_embedding
+
+
+class VectorMemoryProvider(MutableSet[MemoryItem], AbstractSingleton):
+    @abc.abstractmethod
+    def __init__(self, config: Config):
+        pass
+
+    def get(self, query: str) -> MemoryItemRelevance | None:
+        """
+        Gets the data from the memory that is most relevant to the given query.
+
+        Args:
+            data: The data to compare to.
+
+        Returns: The most relevant Memory
+        """
+        result = self.get_relevant(query, 1)
+        return result[0] if result else None
+
+    def get_relevant(self, query: str, k: int) -> Sequence[MemoryItemRelevance]:
+        """
+        Returns the top-k most relevant memories for the given query
+
+        Args:
+            query: the query to compare stored memories to
+            k: the number of relevant memories to fetch
+
+        Returns:
+            list[MemoryItemRelevance] containing the top [k] relevant memories
+        """
+        if len(self) < 1:
+            return []
+
+        logger.debug(
+            f"Searching for {k} relevant memories for query '{query}'; "
+            f"{len(self)} memories in index"
+        )
+
+        relevances = self.score_memories_for_relevance(query)
+        logger.debug(f"Memory relevance scores: {[str(r) for r in relevances]}")
+
+        # take last k items and reverse
+        top_k_indices = np.argsort([r.score for r in relevances])[-k:][::-1]
+
+        return [relevances[i] for i in top_k_indices]
+
+    def score_memories_for_relevance(
+        self, for_query: str
+    ) -> Sequence[MemoryItemRelevance]:
+        """
+        Returns MemoryItemRelevance for every memory in the index.
+        Implementations may override this function for performance purposes.
+        """
+        e_query: Embedding = get_embedding(for_query)
+        return [m.relevance_for(for_query, e_query) for m in self]
+
+    def get_stats(self) -> tuple[int, int]:
+        """
+        Returns:
+            tuple (n_memories: int, n_chunks: int): the stats of the memory index
+        """
+        return len(self), functools.reduce(lambda t, m: t + len(m.e_chunks), self, 0)

autogpt/memory/vector/providers/json_file.py

@@ -0,0 +1,68 @@
+from __future__ import annotations
+
+from pathlib import Path
+from typing import Iterator
+
+import orjson
+
+from autogpt.config import Config
+from autogpt.logs import logger
+
+from ..memory_item import MemoryItem
+from .base import VectorMemoryProvider
+
+
+class JSONFileMemory(VectorMemoryProvider):
+    """Memory backend that stores memories in a JSON file"""
+
+    SAVE_OPTIONS = orjson.OPT_SERIALIZE_NUMPY | orjson.OPT_SERIALIZE_DATACLASS
+
+    file_path: Path
+    memories: list[MemoryItem]
+
+    def __init__(self, cfg: Config) -> None:
+        """Initialize a class instance
+
+        Args:
+            cfg: Config object
+
+        Returns:
+            None
+        """
+        workspace_path = Path(cfg.workspace_path)
+        self.file_path = workspace_path / f"{cfg.memory_index}.json"
+        self.file_path.touch()
+        logger.debug(f"Initialized {__name__} with index path {self.file_path}")
+
+        self.memories = []
+        self.save_index()
+
+    def __iter__(self) -> Iterator[MemoryItem]:
+        return iter(self.memories)
+
+    def __contains__(self, x: MemoryItem) -> bool:
+        return x in self.memories
+
+    def __len__(self) -> int:
+        return len(self.memories)
+
+    def add(self, item: MemoryItem):
+        self.memories.append(item)
+        self.save_index()
+        return len(self.memories)
+
+    def discard(self, item: MemoryItem):
+        try:
+            self.remove(item)
+        except:
+            pass
+
+    def clear(self):
+        """Clears the data in memory."""
+        self.memories.clear()
+        self.save_index()
+
+    def save_index(self):
+        logger.debug(f"Saving memory index to file {self.file_path}")
+        with self.file_path.open("wb") as f:
+            return f.write(orjson.dumps(self.memories, option=self.SAVE_OPTIONS))

autogpt/memory/vector/providers/no_memory.py

@@ -0,0 +1,36 @@
+"""A class that does not store any data. This is the default memory provider."""
+from __future__ import annotations
+
+from typing import Iterator, Optional
+
+from autogpt.config.config import Config
+
+from .. import MemoryItem
+from .base import VectorMemoryProvider
+
+
+class NoMemory(VectorMemoryProvider):
+    """
+    A class that does not store any data. This is the default memory provider.
+    """
+
+    def __init__(self, config: Optional[Config] = None):
+        pass
+
+    def __iter__(self) -> Iterator[MemoryItem]:
+        return iter([])
+
+    def __contains__(self, x: MemoryItem) -> bool:
+        return False
+
+    def __len__(self) -> int:
+        return 0
+
+    def add(self, item: MemoryItem):
+        pass
+
+    def discard(self, item: MemoryItem):
+        pass
+
+    def clear(self):
+        pass

autogpt/memory/vector/utils.py

@@ -0,0 +1,70 @@
+from typing import Any, overload
+
+import numpy as np
+import openai
+
+from autogpt.config import Config
+from autogpt.llm.utils import metered, retry_openai_api
+from autogpt.logs import logger
+
+Embedding = list[np.float32] | np.ndarray[Any, np.dtype[np.float32]]
+"""Embedding vector"""
+TText = list[int]
+"""Token array representing text"""
+
+
+@overload
+def get_embedding(input: str | TText) -> Embedding:
+    ...
+
+
+@overload
+def get_embedding(input: list[str] | list[TText]) -> list[Embedding]:
+    ...
+
+
+@metered
+@retry_openai_api()
+def get_embedding(
+    input: str | TText | list[str] | list[TText],
+) -> Embedding | list[Embedding]:
+    """Get an embedding from the ada model.
+
+    Args:
+        input: Input text to get embeddings for, encoded as a string or array of tokens.
+            Multiple inputs may be given as a list of strings or token arrays.
+
+    Returns:
+        List[float]: The embedding.
+    """
+    cfg = Config()
+    multiple = isinstance(input, list) and all(not isinstance(i, int) for i in input)
+
+    if isinstance(input, str):
+        input = input.replace("\n", " ")
+    elif multiple and isinstance(input[0], str):
+        input = [text.replace("\n", " ") for text in input]
+
+    model = cfg.embedding_model
+    if cfg.use_azure:
+        kwargs = {"engine": cfg.get_azure_deployment_id_for_model(model)}
+    else:
+        kwargs = {"model": model}
+
+    logger.debug(
+        f"Getting embedding{f's for {len(input)} inputs' if multiple else ''}"
+        f" with model '{model}'"
+        + (f" via Azure deployment '{kwargs['engine']}'" if cfg.use_azure else "")
+    )
+
+    embeddings = openai.Embedding.create(
+        input=input,
+        api_key=cfg.openai_api_key,
+        **kwargs,
+    ).data
+
+    if not multiple:
+        return embeddings[0]["embedding"]
+
+    embeddings = sorted(embeddings, key=lambda x: x["index"])
+    return [d["embedding"] for d in embeddings]

autogpt/models/base_open_ai_plugin.py

@@ -0,0 +1,249 @@
+"""Handles loading of plugins."""
+from typing import Any, Dict, List, Optional, Tuple, TypedDict, TypeVar
+
+from auto_gpt_plugin_template import AutoGPTPluginTemplate
+
+PromptGenerator = TypeVar("PromptGenerator")
+
+
+class Message(TypedDict):
+    role: str
+    content: str
+
+
+class BaseOpenAIPlugin(AutoGPTPluginTemplate):
+    """
+    This is a BaseOpenAIPlugin class for generating Auto-GPT plugins.
+    """
+
+    def __init__(self, manifests_specs_clients: dict):
+        # super().__init__()
+        self._name = manifests_specs_clients["manifest"]["name_for_model"]
+        self._version = manifests_specs_clients["manifest"]["schema_version"]
+        self._description = manifests_specs_clients["manifest"]["description_for_model"]
+        self._client = manifests_specs_clients["client"]
+        self._manifest = manifests_specs_clients["manifest"]
+        self._openapi_spec = manifests_specs_clients["openapi_spec"]
+
+    def can_handle_on_response(self) -> bool:
+        """This method is called to check that the plugin can
+        handle the on_response method.
+        Returns:
+            bool: True if the plugin can handle the on_response method."""
+        return False
+
+    def on_response(self, response: str, *args, **kwargs) -> str:
+        """This method is called when a response is received from the model."""
+        return response
+
+    def can_handle_post_prompt(self) -> bool:
+        """This method is called to check that the plugin can
+        handle the post_prompt method.
+        Returns:
+            bool: True if the plugin can handle the post_prompt method."""
+        return False
+
+    def post_prompt(self, prompt: PromptGenerator) -> PromptGenerator:
+        """This method is called just after the generate_prompt is called,
+            but actually before the prompt is generated.
+        Args:
+            prompt (PromptGenerator): The prompt generator.
+        Returns:
+            PromptGenerator: The prompt generator.
+        """
+        return prompt
+
+    def can_handle_on_planning(self) -> bool:
+        """This method is called to check that the plugin can
+        handle the on_planning method.
+        Returns:
+            bool: True if the plugin can handle the on_planning method."""
+        return False
+
+    def on_planning(
+        self, prompt: PromptGenerator, messages: List[Message]
+    ) -> Optional[str]:
+        """This method is called before the planning chat completion is done.
+        Args:
+            prompt (PromptGenerator): The prompt generator.
+            messages (List[str]): The list of messages.
+        """
+
+    def can_handle_post_planning(self) -> bool:
+        """This method is called to check that the plugin can
+        handle the post_planning method.
+        Returns:
+            bool: True if the plugin can handle the post_planning method."""
+        return False
+
+    def post_planning(self, response: str) -> str:
+        """This method is called after the planning chat completion is done.
+        Args:
+            response (str): The response.
+        Returns:
+            str: The resulting response.
+        """
+        return response
+
+    def can_handle_pre_instruction(self) -> bool:
+        """This method is called to check that the plugin can
+        handle the pre_instruction method.
+        Returns:
+            bool: True if the plugin can handle the pre_instruction method."""
+        return False
+
+    def pre_instruction(self, messages: List[Message]) -> List[Message]:
+        """This method is called before the instruction chat is done.
+        Args:
+            messages (List[Message]): The list of context messages.
+        Returns:
+            List[Message]: The resulting list of messages.
+        """
+        return messages
+
+    def can_handle_on_instruction(self) -> bool:
+        """This method is called to check that the plugin can
+        handle the on_instruction method.
+        Returns:
+            bool: True if the plugin can handle the on_instruction method."""
+        return False
+
+    def on_instruction(self, messages: List[Message]) -> Optional[str]:
+        """This method is called when the instruction chat is done.
+        Args:
+            messages (List[Message]): The list of context messages.
+        Returns:
+            Optional[str]: The resulting message.
+        """
+
+    def can_handle_post_instruction(self) -> bool:
+        """This method is called to check that the plugin can
+        handle the post_instruction method.
+        Returns:
+            bool: True if the plugin can handle the post_instruction method."""
+        return False
+
+    def post_instruction(self, response: str) -> str:
+        """This method is called after the instruction chat is done.
+        Args:
+            response (str): The response.
+        Returns:
+            str: The resulting response.
+        """
+        return response
+
+    def can_handle_pre_command(self) -> bool:
+        """This method is called to check that the plugin can
+        handle the pre_command method.
+        Returns:
+            bool: True if the plugin can handle the pre_command method."""
+        return False
+
+    def pre_command(
+        self, command_name: str, arguments: Dict[str, Any]
+    ) -> Tuple[str, Dict[str, Any]]:
+        """This method is called before the command is executed.
+        Args:
+            command_name (str): The command name.
+            arguments (Dict[str, Any]): The arguments.
+        Returns:
+            Tuple[str, Dict[str, Any]]: The command name and the arguments.
+        """
+        return command_name, arguments
+
+    def can_handle_post_command(self) -> bool:
+        """This method is called to check that the plugin can
+        handle the post_command method.
+        Returns:
+            bool: True if the plugin can handle the post_command method."""
+        return False
+
+    def post_command(self, command_name: str, response: str) -> str:
+        """This method is called after the command is executed.
+        Args:
+            command_name (str): The command name.
+            response (str): The response.
+        Returns:
+            str: The resulting response.
+        """
+        return response
+
+    def can_handle_chat_completion(
+        self, messages: Dict[Any, Any], model: str, temperature: float, max_tokens: int
+    ) -> bool:
+        """This method is called to check that the plugin can
+          handle the chat_completion method.
+        Args:
+            messages (List[Message]): The messages.
+            model (str): The model name.
+            temperature (float): The temperature.
+            max_tokens (int): The max tokens.
+          Returns:
+              bool: True if the plugin can handle the chat_completion method."""
+        return False
+
+    def handle_chat_completion(
+        self, messages: List[Message], model: str, temperature: float, max_tokens: int
+    ) -> str:
+        """This method is called when the chat completion is done.
+        Args:
+            messages (List[Message]): The messages.
+            model (str): The model name.
+            temperature (float): The temperature.
+            max_tokens (int): The max tokens.
+        Returns:
+            str: The resulting response.
+        """
+
+    def can_handle_text_embedding(self, text: str) -> bool:
+        """This method is called to check that the plugin can
+          handle the text_embedding method.
+        Args:
+            text (str): The text to be convert to embedding.
+          Returns:
+              bool: True if the plugin can handle the text_embedding method."""
+        return False
+
+    def handle_text_embedding(self, text: str) -> list:
+        """This method is called when the chat completion is done.
+        Args:
+            text (str): The text to be convert to embedding.
+        Returns:
+            list: The text embedding.
+        """
+
+    def can_handle_user_input(self, user_input: str) -> bool:
+        """This method is called to check that the plugin can
+        handle the user_input method.
+
+        Args:
+            user_input (str): The user input.
+
+        Returns:
+            bool: True if the plugin can handle the user_input method."""
+        return False
+
+    def user_input(self, user_input: str) -> str:
+        """This method is called to request user input to the user.
+
+        Args:
+            user_input (str): The question or prompt to ask the user.
+
+        Returns:
+            str: The user input.
+        """
+
+    def can_handle_report(self) -> bool:
+        """This method is called to check that the plugin can
+        handle the report method.
+
+        Returns:
+            bool: True if the plugin can handle the report method."""
+        return False
+
+    def report(self, message: str) -> None:
+        """This method is called to report a message to the user.
+
+        Args:
+            message (str): The message to report.
+        """

autogpt/plugins.py

@@ -0,0 +1,283 @@
+"""Handles loading of plugins."""
+
+import importlib.util
+import json
+import os
+import zipfile
+from pathlib import Path
+from typing import List
+from urllib.parse import urlparse
+from zipimport import zipimporter
+
+import openapi_python_client
+import requests
+from auto_gpt_plugin_template import AutoGPTPluginTemplate
+from openapi_python_client.config import Config as OpenAPIConfig
+
+from autogpt.config import Config
+from autogpt.logs import logger
+from autogpt.models.base_open_ai_plugin import BaseOpenAIPlugin
+
+
+def inspect_zip_for_modules(zip_path: str, debug: bool = False) -> list[str]:
+    """
+    Inspect a zipfile for a modules.
+
+    Args:
+        zip_path (str): Path to the zipfile.
+        debug (bool, optional): Enable debug logging. Defaults to False.
+
+    Returns:
+        list[str]: The list of module names found or empty list if none were found.
+    """
+    result = []
+    with zipfile.ZipFile(zip_path, "r") as zfile:
+        for name in zfile.namelist():
+            if name.endswith("__init__.py") and not name.startswith("__MACOSX"):
+                logger.debug(f"Found module '{name}' in the zipfile at: {name}")
+                result.append(name)
+    if len(result) == 0:
+        logger.debug(f"Module '__init__.py' not found in the zipfile @ {zip_path}.")
+    return result
+
+
+def write_dict_to_json_file(data: dict, file_path: str) -> None:
+    """
+    Write a dictionary to a JSON file.
+    Args:
+        data (dict): Dictionary to write.
+        file_path (str): Path to the file.
+    """
+    with open(file_path, "w") as file:
+        json.dump(data, file, indent=4)
+
+
+def fetch_openai_plugins_manifest_and_spec(cfg: Config) -> dict:
+    """
+    Fetch the manifest for a list of OpenAI plugins.
+        Args:
+        urls (List): List of URLs to fetch.
+    Returns:
+        dict: per url dictionary of manifest and spec.
+    """
+    # TODO add directory scan
+    manifests = {}
+    for url in cfg.plugins_openai:
+        openai_plugin_client_dir = f"{cfg.plugins_dir}/openai/{urlparse(url).netloc}"
+        create_directory_if_not_exists(openai_plugin_client_dir)
+        if not os.path.exists(f"{openai_plugin_client_dir}/ai-plugin.json"):
+            try:
+                response = requests.get(f"{url}/.well-known/ai-plugin.json")
+                if response.status_code == 200:
+                    manifest = response.json()
+                    if manifest["schema_version"] != "v1":
+                        logger.warn(
+                            f"Unsupported manifest version: {manifest['schem_version']} for {url}"
+                        )
+                        continue
+                    if manifest["api"]["type"] != "openapi":
+                        logger.warn(
+                            f"Unsupported API type: {manifest['api']['type']} for {url}"
+                        )
+                        continue
+                    write_dict_to_json_file(
+                        manifest, f"{openai_plugin_client_dir}/ai-plugin.json"
+                    )
+                else:
+                    logger.warn(
+                        f"Failed to fetch manifest for {url}: {response.status_code}"
+                    )
+            except requests.exceptions.RequestException as e:
+                logger.warn(f"Error while requesting manifest from {url}: {e}")
+        else:
+            logger.info(f"Manifest for {url} already exists")
+            manifest = json.load(open(f"{openai_plugin_client_dir}/ai-plugin.json"))
+        if not os.path.exists(f"{openai_plugin_client_dir}/openapi.json"):
+            openapi_spec = openapi_python_client._get_document(
+                url=manifest["api"]["url"], path=None, timeout=5
+            )
+            write_dict_to_json_file(
+                openapi_spec, f"{openai_plugin_client_dir}/openapi.json"
+            )
+        else:
+            logger.info(f"OpenAPI spec for {url} already exists")
+            openapi_spec = json.load(open(f"{openai_plugin_client_dir}/openapi.json"))
+        manifests[url] = {"manifest": manifest, "openapi_spec": openapi_spec}
+    return manifests
+
+
+def create_directory_if_not_exists(directory_path: str) -> bool:
+    """
+    Create a directory if it does not exist.
+    Args:
+        directory_path (str): Path to the directory.
+    Returns:
+        bool: True if the directory was created, else False.
+    """
+    if not os.path.exists(directory_path):
+        try:
+            os.makedirs(directory_path)
+            logger.debug(f"Created directory: {directory_path}")
+            return True
+        except OSError as e:
+            logger.warn(f"Error creating directory {directory_path}: {e}")
+            return False
+    else:
+        logger.info(f"Directory {directory_path} already exists")
+        return True
+
+
+def initialize_openai_plugins(
+    manifests_specs: dict, cfg: Config, debug: bool = False
+) -> dict:
+    """
+    Initialize OpenAI plugins.
+    Args:
+        manifests_specs (dict): per url dictionary of manifest and spec.
+        cfg (Config): Config instance including plugins config
+        debug (bool, optional): Enable debug logging. Defaults to False.
+    Returns:
+        dict: per url dictionary of manifest, spec and client.
+    """
+    openai_plugins_dir = f"{cfg.plugins_dir}/openai"
+    if create_directory_if_not_exists(openai_plugins_dir):
+        for url, manifest_spec in manifests_specs.items():
+            openai_plugin_client_dir = f"{openai_plugins_dir}/{urlparse(url).hostname}"
+            _meta_option = (openapi_python_client.MetaType.SETUP,)
+            _config = OpenAPIConfig(
+                **{
+                    "project_name_override": "client",
+                    "package_name_override": "client",
+                }
+            )
+            prev_cwd = Path.cwd()
+            os.chdir(openai_plugin_client_dir)
+
+            if not os.path.exists("client"):
+                client_results = openapi_python_client.create_new_client(
+                    url=manifest_spec["manifest"]["api"]["url"],
+                    path=None,
+                    meta=_meta_option,
+                    config=_config,
+                )
+                if client_results:
+                    logger.warn(
+                        f"Error creating OpenAPI client: {client_results[0].header} \n"
+                        f" details: {client_results[0].detail}"
+                    )
+                    continue
+            spec = importlib.util.spec_from_file_location(
+                "client", "client/client/client.py"
+            )
+            module = importlib.util.module_from_spec(spec)
+
+            try:
+                spec.loader.exec_module(module)
+            finally:
+                os.chdir(prev_cwd)
+
+            client = module.Client(base_url=url)
+            manifest_spec["client"] = client
+    return manifests_specs
+
+
+def instantiate_openai_plugin_clients(
+    manifests_specs_clients: dict, cfg: Config, debug: bool = False
+) -> dict:
+    """
+    Instantiates BaseOpenAIPlugin instances for each OpenAI plugin.
+    Args:
+        manifests_specs_clients (dict): per url dictionary of manifest, spec and client.
+        cfg (Config): Config instance including plugins config
+        debug (bool, optional): Enable debug logging. Defaults to False.
+    Returns:
+          plugins (dict): per url dictionary of BaseOpenAIPlugin instances.
+
+    """
+    plugins = {}
+    for url, manifest_spec_client in manifests_specs_clients.items():
+        plugins[url] = BaseOpenAIPlugin(manifest_spec_client)
+    return plugins
+
+
+def scan_plugins(cfg: Config, debug: bool = False) -> List[AutoGPTPluginTemplate]:
+    """Scan the plugins directory for plugins and loads them.
+
+    Args:
+        cfg (Config): Config instance including plugins config
+        debug (bool, optional): Enable debug logging. Defaults to False.
+
+    Returns:
+        List[Tuple[str, Path]]: List of plugins.
+    """
+    loaded_plugins = []
+    # Generic plugins
+    plugins_path_path = Path(cfg.plugins_dir)
+
+    logger.debug(f"Allowlisted Plugins: {cfg.plugins_allowlist}")
+    logger.debug(f"Denylisted Plugins: {cfg.plugins_denylist}")
+
+    for plugin in plugins_path_path.glob("*.zip"):
+        if moduleList := inspect_zip_for_modules(str(plugin), debug):
+            for module in moduleList:
+                plugin = Path(plugin)
+                module = Path(module)
+                logger.debug(f"Plugin: {plugin} Module: {module}")
+                zipped_package = zipimporter(str(plugin))
+                zipped_module = zipped_package.load_module(str(module.parent))
+                for key in dir(zipped_module):
+                    if key.startswith("__"):
+                        continue
+                    a_module = getattr(zipped_module, key)
+                    a_keys = dir(a_module)
+                    if (
+                        "_abc_impl" in a_keys
+                        and a_module.__name__ != "AutoGPTPluginTemplate"
+                        and denylist_allowlist_check(a_module.__name__, cfg)
+                    ):
+                        loaded_plugins.append(a_module())
+    # OpenAI plugins
+    if cfg.plugins_openai:
+        manifests_specs = fetch_openai_plugins_manifest_and_spec(cfg)
+        if manifests_specs.keys():
+            manifests_specs_clients = initialize_openai_plugins(
+                manifests_specs, cfg, debug
+            )
+            for url, openai_plugin_meta in manifests_specs_clients.items():
+                if denylist_allowlist_check(url, cfg):
+                    plugin = BaseOpenAIPlugin(openai_plugin_meta)
+                    loaded_plugins.append(plugin)
+
+    if loaded_plugins:
+        logger.info(f"\nPlugins found: {len(loaded_plugins)}\n" "--------------------")
+    for plugin in loaded_plugins:
+        logger.info(f"{plugin._name}: {plugin._version} - {plugin._description}")
+    return loaded_plugins
+
+
+def denylist_allowlist_check(plugin_name: str, cfg: Config) -> bool:
+    """Check if the plugin is in the allowlist or denylist.
+
+    Args:
+        plugin_name (str): Name of the plugin.
+        cfg (Config): Config object.
+
+    Returns:
+        True or False
+    """
+    logger.debug(f"Checking if plugin {plugin_name} should be loaded")
+    if (
+        plugin_name in cfg.plugins_denylist
+        or "all" in cfg.plugins_denylist
+        or "none" in cfg.plugins_allowlist
+    ):
+        logger.debug(f"Not loading plugin {plugin_name} as it was in the denylist.")
+        return False
+    if plugin_name in cfg.plugins_allowlist or "all" in cfg.plugins_allowlist:
+        logger.debug(f"Loading plugin {plugin_name} as it was in the allowlist.")
+        return True
+    ack = input(
+        f"WARNING: Plugin {plugin_name} found. But not in the"
+        f" allowlist... Load? ({cfg.authorise_key}/{cfg.exit_key}): "
+    )
+    return ack.lower() == cfg.authorise_key

autogpt/processing/html.py

@@ -0,0 +1,33 @@
+"""HTML processing functions"""
+from __future__ import annotations
+
+from bs4 import BeautifulSoup
+from requests.compat import urljoin
+
+
+def extract_hyperlinks(soup: BeautifulSoup, base_url: str) -> list[tuple[str, str]]:
+    """Extract hyperlinks from a BeautifulSoup object
+
+    Args:
+        soup (BeautifulSoup): The BeautifulSoup object
+        base_url (str): The base URL
+
+    Returns:
+        List[Tuple[str, str]]: The extracted hyperlinks
+    """
+    return [
+        (link.text, urljoin(base_url, link["href"]))
+        for link in soup.find_all("a", href=True)
+    ]
+
+
+def format_hyperlinks(hyperlinks: list[tuple[str, str]]) -> list[str]:
+    """Format hyperlinks to be displayed to the user
+
+    Args:
+        hyperlinks (List[Tuple[str, str]]): The hyperlinks to format
+
+    Returns:
+        List[str]: The formatted hyperlinks
+    """
+    return [f"{link_text} ({link_url})" for link_text, link_url in hyperlinks]

autogpt/processing/text.py

@@ -0,0 +1,234 @@
+"""Text processing functions"""
+from math import ceil
+from typing import Optional
+
+import spacy
+import tiktoken
+
+from autogpt.config import Config
+from autogpt.llm.base import ChatSequence
+from autogpt.llm.providers.openai import OPEN_AI_MODELS
+from autogpt.llm.utils import count_string_tokens, create_chat_completion
+from autogpt.logs import logger
+from autogpt.utils import batch
+
+CFG = Config()
+
+
+def _max_chunk_length(model: str, max: Optional[int] = None) -> int:
+    model_max_input_tokens = OPEN_AI_MODELS[model].max_tokens - 1
+    if max is not None and max > 0:
+        return min(max, model_max_input_tokens)
+    return model_max_input_tokens
+
+
+def must_chunk_content(
+    text: str, for_model: str, max_chunk_length: Optional[int] = None
+) -> bool:
+    return count_string_tokens(text, for_model) > _max_chunk_length(
+        for_model, max_chunk_length
+    )
+
+
+def chunk_content(
+    content: str,
+    for_model: str,
+    max_chunk_length: Optional[int] = None,
+    with_overlap=True,
+):
+    """Split content into chunks of approximately equal token length."""
+
+    MAX_OVERLAP = 200  # limit overlap to save tokens
+
+    if not must_chunk_content(content, for_model, max_chunk_length):
+        yield content, count_string_tokens(content, for_model)
+        return
+
+    max_chunk_length = max_chunk_length or _max_chunk_length(for_model)
+
+    tokenizer = tiktoken.encoding_for_model(for_model)
+
+    tokenized_text = tokenizer.encode(content)
+    total_length = len(tokenized_text)
+    n_chunks = ceil(total_length / max_chunk_length)
+
+    chunk_length = ceil(total_length / n_chunks)
+    overlap = min(max_chunk_length - chunk_length, MAX_OVERLAP) if with_overlap else 0
+
+    for token_batch in batch(tokenized_text, chunk_length + overlap, overlap):
+        yield tokenizer.decode(token_batch), len(token_batch)
+
+
+def summarize_text(
+    text: str, instruction: Optional[str] = None, question: Optional[str] = None
+) -> tuple[str, None | list[tuple[str, str]]]:
+    """Summarize text using the OpenAI API
+
+    Args:
+        text (str): The text to summarize
+        instruction (str): Additional instruction for summarization, e.g. "focus on information related to polar bears", "omit personal information contained in the text"
+
+    Returns:
+        str: The summary of the text
+        list[(summary, chunk)]: Text chunks and their summary, if the text was chunked.
+            None otherwise.
+    """
+    if not text:
+        raise ValueError("No text to summarize")
+
+    if instruction and question:
+        raise ValueError("Parameters 'question' and 'instructions' cannot both be set")
+
+    model = CFG.fast_llm_model
+
+    if question:
+        instruction = (
+            f'include any information that can be used to answer the question "{question}". '
+            "Do not directly answer the question itself"
+        )
+
+    summarization_prompt = ChatSequence.for_model(model)
+
+    token_length = count_string_tokens(text, model)
+    logger.info(f"Text length: {token_length} tokens")
+
+    # reserve 50 tokens for summary prompt, 500 for the response
+    max_chunk_length = _max_chunk_length(model) - 550
+    logger.info(f"Max chunk length: {max_chunk_length} tokens")
+
+    if not must_chunk_content(text, model, max_chunk_length):
+        # summarization_prompt.add("user", text)
+        summarization_prompt.add(
+            "user",
+            "Write a concise summary of the following text"
+            f"{f'; {instruction}' if instruction is not None else ''}:"
+            "\n\n\n"
+            f'LITERAL TEXT: """{text}"""'
+            "\n\n\n"
+            "CONCISE SUMMARY: The text is best summarized as"
+            # "Only respond with a concise summary or description of the user message."
+        )
+
+        logger.debug(f"Summarizing with {model}:\n{summarization_prompt.dump()}\n")
+        summary = create_chat_completion(
+            summarization_prompt, temperature=0, max_tokens=500
+        )
+
+        logger.debug(f"\n{'-'*16} SUMMARY {'-'*17}\n{summary}\n{'-'*42}\n")
+        return summary.strip(), None
+
+    summaries: list[str] = []
+    chunks = list(split_text(text, for_model=model, max_chunk_length=max_chunk_length))
+
+    for i, (chunk, chunk_length) in enumerate(chunks):
+        logger.info(
+            f"Summarizing chunk {i + 1} / {len(chunks)} of length {chunk_length} tokens"
+        )
+        summary, _ = summarize_text(chunk, instruction)
+        summaries.append(summary)
+
+    logger.info(f"Summarized {len(chunks)} chunks")
+
+    summary, _ = summarize_text("\n\n".join(summaries))
+
+    return summary.strip(), [
+        (summaries[i], chunks[i][0]) for i in range(0, len(chunks))
+    ]
+
+
+def split_text(
+    text: str,
+    for_model: str = CFG.fast_llm_model,
+    with_overlap=True,
+    max_chunk_length: Optional[int] = None,
+):
+    """Split text into chunks of sentences, with each chunk not exceeding the maximum length
+
+    Args:
+        text (str): The text to split
+        for_model (str): The model to chunk for; determines tokenizer and constraints
+        max_length (int, optional): The maximum length of each chunk
+
+    Yields:
+        str: The next chunk of text
+
+    Raises:
+        ValueError: when a sentence is longer than the maximum length
+    """
+    max_length = _max_chunk_length(for_model, max_chunk_length)
+
+    # flatten paragraphs to improve performance
+    text = text.replace("\n", " ")
+    text_length = count_string_tokens(text, for_model)
+
+    if text_length < max_length:
+        yield text, text_length
+        return
+
+    n_chunks = ceil(text_length / max_length)
+    target_chunk_length = ceil(text_length / n_chunks)
+
+    nlp: spacy.language.Language = spacy.load(CFG.browse_spacy_language_model)
+    nlp.add_pipe("sentencizer")
+    doc = nlp(text)
+    sentences = [sentence.text.strip() for sentence in doc.sents]
+
+    current_chunk: list[str] = []
+    current_chunk_length = 0
+    last_sentence = None
+    last_sentence_length = 0
+
+    i = 0
+    while i < len(sentences):
+        sentence = sentences[i]
+        sentence_length = count_string_tokens(sentence, for_model)
+        expected_chunk_length = current_chunk_length + 1 + sentence_length
+
+        if (
+            expected_chunk_length < max_length
+            # try to create chunks of approximately equal size
+            and expected_chunk_length - (sentence_length / 2) < target_chunk_length
+        ):
+            current_chunk.append(sentence)
+            current_chunk_length = expected_chunk_length
+
+        elif sentence_length < max_length:
+            if last_sentence:
+                yield " ".join(current_chunk), current_chunk_length
+                current_chunk = []
+                current_chunk_length = 0
+
+                if with_overlap:
+                    overlap_max_length = max_length - sentence_length - 1
+                    if last_sentence_length < overlap_max_length:
+                        current_chunk += [last_sentence]
+                        current_chunk_length += last_sentence_length + 1
+                    elif overlap_max_length > 5:
+                        # add as much from the end of the last sentence as fits
+                        current_chunk += [
+                            list(
+                                chunk_content(
+                                    last_sentence,
+                                    for_model,
+                                    overlap_max_length,
+                                )
+                            ).pop()[0],
+                        ]
+                        current_chunk_length += overlap_max_length + 1
+
+            current_chunk += [sentence]
+            current_chunk_length += sentence_length
+
+        else:  # sentence longer than maximum length -> chop up and try again
+            sentences[i : i + 1] = [
+                chunk
+                for chunk, _ in chunk_content(sentence, for_model, target_chunk_length)
+            ]
+            continue
+
+        i += 1
+        last_sentence = sentence
+        last_sentence_length = sentence_length
+
+    if current_chunk:
+        yield " ".join(current_chunk), current_chunk_length

autogpt/prompts/default_prompts.py

@@ -0,0 +1,29 @@
+#########################Setup.py#################################
+
+DEFAULT_SYSTEM_PROMPT_AICONFIG_AUTOMATIC = """
+Your task is to devise up to 5 highly effective goals and an appropriate role-based name (_GPT) for an autonomous agent, ensuring that the goals are optimally aligned with the successful completion of its assigned task.
+
+The user will provide the task, you will provide only the output in the exact format specified below with no explanation or conversation.
+
+Example input:
+Help me with marketing my business
+
+Example output:
+Name: CMOGPT
+Description: a professional digital marketer AI that assists Solopreneurs in growing their businesses by providing world-class expertise in solving marketing problems for SaaS, content products, agencies, and more.
+Goals:
+- Engage in effective problem-solving, prioritization, planning, and supporting execution to address your marketing needs as your virtual Chief Marketing Officer.
+
+- Provide specific, actionable, and concise advice to help you make informed decisions without the use of platitudes or overly wordy explanations.
+
+- Identify and prioritize quick wins and cost-effective campaigns that maximize results with minimal time and budget investment.
+
+- Proactively take the lead in guiding you and offering suggestions when faced with unclear information or uncertainty to ensure your marketing strategy remains on track.
+"""
+
+DEFAULT_TASK_PROMPT_AICONFIG_AUTOMATIC = (
+    "Task: '{{user_prompt}}'\n"
+    "Respond only with the output in the exact format specified in the system prompt, with no explanation or conversation.\n"
+)
+
+DEFAULT_USER_DESIRE_PROMPT = "Write a wikipedia style article about the project: https://github.com/significant-gravitas/Auto-GPT"  # Default prompt

autogpt/prompts/generator.py

@@ -0,0 +1,158 @@
+""" A module for generating custom prompt strings."""
+import json
+from typing import TYPE_CHECKING, Any, Callable, Dict, List, Optional
+
+if TYPE_CHECKING:
+    from autogpt.commands.command import CommandRegistry
+
+
+class PromptGenerator:
+    """
+    A class for generating custom prompt strings based on constraints, commands,
+        resources, and performance evaluations.
+    """
+
+    def __init__(self) -> None:
+        """
+        Initialize the PromptGenerator object with empty lists of constraints,
+            commands, resources, and performance evaluations.
+        """
+        self.constraints = []
+        self.commands = []
+        self.resources = []
+        self.performance_evaluation = []
+        self.goals = []
+        self.command_registry: CommandRegistry | None = None
+        self.name = "Bob"
+        self.role = "AI"
+        self.response_format = {
+            "thoughts": {
+                "text": "thought",
+                "reasoning": "reasoning",
+                "plan": "- short bulleted\n- list that conveys\n- long-term plan",
+                "criticism": "constructive self-criticism",
+                "speak": "thoughts summary to say to user",
+            },
+            "command": {"name": "command name", "args": {"arg name": "value"}},
+        }
+
+    def add_constraint(self, constraint: str) -> None:
+        """
+        Add a constraint to the constraints list.
+
+        Args:
+            constraint (str): The constraint to be added.
+        """
+        self.constraints.append(constraint)
+
+    def add_command(
+        self,
+        command_label: str,
+        command_name: str,
+        args=None,
+        function: Optional[Callable] = None,
+    ) -> None:
+        """
+        Add a command to the commands list with a label, name, and optional arguments.
+
+        Args:
+            command_label (str): The label of the command.
+            command_name (str): The name of the command.
+            args (dict, optional): A dictionary containing argument names and their
+              values. Defaults to None.
+            function (callable, optional): A callable function to be called when
+                the command is executed. Defaults to None.
+        """
+        if args is None:
+            args = {}
+
+        command_args = {arg_key: arg_value for arg_key, arg_value in args.items()}
+
+        command = {
+            "label": command_label,
+            "name": command_name,
+            "args": command_args,
+            "function": function,
+        }
+
+        self.commands.append(command)
+
+    def _generate_command_string(self, command: Dict[str, Any]) -> str:
+        """
+        Generate a formatted string representation of a command.
+
+        Args:
+            command (dict): A dictionary containing command information.
+
+        Returns:
+            str: The formatted command string.
+        """
+        args_string = ", ".join(
+            f'"{key}": "{value}"' for key, value in command["args"].items()
+        )
+        return f'{command["label"]}: "{command["name"]}", args: {args_string}'
+
+    def add_resource(self, resource: str) -> None:
+        """
+        Add a resource to the resources list.
+
+        Args:
+            resource (str): The resource to be added.
+        """
+        self.resources.append(resource)
+
+    def add_performance_evaluation(self, evaluation: str) -> None:
+        """
+        Add a performance evaluation item to the performance_evaluation list.
+
+        Args:
+            evaluation (str): The evaluation item to be added.
+        """
+        self.performance_evaluation.append(evaluation)
+
+    def _generate_numbered_list(self, items: List[Any], item_type="list") -> str:
+        """
+        Generate a numbered list from given items based on the item_type.
+
+        Args:
+            items (list): A list of items to be numbered.
+            item_type (str, optional): The type of items in the list.
+                Defaults to 'list'.
+
+        Returns:
+            str: The formatted numbered list.
+        """
+        if item_type == "command":
+            command_strings = []
+            if self.command_registry:
+                command_strings += [
+                    str(item)
+                    for item in self.command_registry.commands.values()
+                    if item.enabled
+                ]
+            # terminate command is added manually
+            command_strings += [self._generate_command_string(item) for item in items]
+            return "\n".join(f"{i+1}. {item}" for i, item in enumerate(command_strings))
+        else:
+            return "\n".join(f"{i+1}. {item}" for i, item in enumerate(items))
+
+    def generate_prompt_string(self) -> str:
+        """
+        Generate a prompt string based on the constraints, commands, resources,
+            and performance evaluations.
+
+        Returns:
+            str: The generated prompt string.
+        """
+        formatted_response_format = json.dumps(self.response_format, indent=4)
+        return (
+            f"Constraints:\n{self._generate_numbered_list(self.constraints)}\n\n"
+            "Commands:\n"
+            f"{self._generate_numbered_list(self.commands, item_type='command')}\n\n"
+            f"Resources:\n{self._generate_numbered_list(self.resources)}\n\n"
+            "Performance Evaluation:\n"
+            f"{self._generate_numbered_list(self.performance_evaluation)}\n\n"
+            "You should only respond in JSON format as described below \nResponse"
+            f" Format: \n{formatted_response_format} \nEnsure the response can be"
+            " parsed by Python json.loads"
+        )

autogpt/prompts/prompt.py

@@ -0,0 +1,115 @@
+from colorama import Fore
+
+from autogpt.config.ai_config import AIConfig
+from autogpt.config.config import Config
+from autogpt.config.prompt_config import PromptConfig
+from autogpt.llm.api_manager import ApiManager
+from autogpt.logs import logger
+from autogpt.prompts.generator import PromptGenerator
+from autogpt.setup import prompt_user
+from autogpt.utils import clean_input
+
+CFG = Config()
+
+DEFAULT_TRIGGERING_PROMPT = (
+    "Determine which next command to use, and respond using the format specified above:"
+)
+
+
+def build_default_prompt_generator() -> PromptGenerator:
+    """
+    This function generates a prompt string that includes various constraints,
+        commands, resources, and performance evaluations.
+
+    Returns:
+        str: The generated prompt string.
+    """
+
+    # Initialize the PromptGenerator object
+    prompt_generator = PromptGenerator()
+
+    # Initialize the PromptConfig object and load the file set in the main config (default: prompts_settings.yaml)
+    prompt_config = PromptConfig(CFG.prompt_settings_file)
+
+    # Add constraints to the PromptGenerator object
+    for constraint in prompt_config.constraints:
+        prompt_generator.add_constraint(constraint)
+
+    # Add resources to the PromptGenerator object
+    for resource in prompt_config.resources:
+        prompt_generator.add_resource(resource)
+
+    # Add performance evaluations to the PromptGenerator object
+    for performance_evaluation in prompt_config.performance_evaluations:
+        prompt_generator.add_performance_evaluation(performance_evaluation)
+
+    return prompt_generator
+
+
+def construct_main_ai_config() -> AIConfig:
+    """Construct the prompt for the AI to respond to
+
+    Returns:
+        str: The prompt string
+    """
+    config = AIConfig.load(CFG.ai_settings_file)
+    if CFG.skip_reprompt and config.ai_name:
+        logger.typewriter_log("Name :", Fore.GREEN, config.ai_name)
+        logger.typewriter_log("Role :", Fore.GREEN, config.ai_role)
+        logger.typewriter_log("Goals:", Fore.GREEN, f"{config.ai_goals}")
+        logger.typewriter_log(
+            "API Budget:",
+            Fore.GREEN,
+            "infinite" if config.api_budget <= 0 else f"${config.api_budget}",
+        )
+    elif config.ai_name:
+        logger.typewriter_log(
+            "Welcome back! ",
+            Fore.GREEN,
+            f"Would you like me to return to being {config.ai_name}?",
+            speak_text=True,
+        )
+        should_continue = clean_input(
+            f"""Continue with the last settings?
+Name:  {config.ai_name}
+Role:  {config.ai_role}
+Goals: {config.ai_goals}
+API Budget: {"infinite" if config.api_budget <= 0 else f"${config.api_budget}"}
+Continue ({CFG.authorise_key}/{CFG.exit_key}): """
+        )
+        if should_continue.lower() == CFG.exit_key:
+            config = AIConfig()
+
+    if not config.ai_name:
+        config = prompt_user()
+        config.save(CFG.ai_settings_file)
+
+    if CFG.restrict_to_workspace:
+        logger.typewriter_log(
+            "NOTE:All files/directories created by this agent can be found inside its workspace at:",
+            Fore.YELLOW,
+            f"{CFG.workspace_path}",
+        )
+    # set the total api budget
+    api_manager = ApiManager()
+    api_manager.set_total_budget(config.api_budget)
+
+    # Agent Created, print message
+    logger.typewriter_log(
+        config.ai_name,
+        Fore.LIGHTBLUE_EX,
+        "has been created with the following details:",
+        speak_text=True,
+    )
+
+    # Print the ai config details
+    # Name
+    logger.typewriter_log("Name:", Fore.GREEN, config.ai_name, speak_text=False)
+    # Role
+    logger.typewriter_log("Role:", Fore.GREEN, config.ai_role, speak_text=False)
+    # Goals
+    logger.typewriter_log("Goals:", Fore.GREEN, "", speak_text=False)
+    for goal in config.ai_goals:
+        logger.typewriter_log("-", Fore.GREEN, goal, speak_text=False)
+
+    return config

autogpt/setup.py

@@ -0,0 +1,206 @@
+"""Set up the AI and its goals"""
+import re
+
+from colorama import Fore, Style
+from jinja2 import Template
+
+from autogpt import utils
+from autogpt.config import Config
+from autogpt.config.ai_config import AIConfig
+from autogpt.llm.base import ChatSequence, Message
+from autogpt.llm.chat import create_chat_completion
+from autogpt.logs import logger
+from autogpt.prompts.default_prompts import (
+    DEFAULT_SYSTEM_PROMPT_AICONFIG_AUTOMATIC,
+    DEFAULT_TASK_PROMPT_AICONFIG_AUTOMATIC,
+    DEFAULT_USER_DESIRE_PROMPT,
+)
+
+CFG = Config()
+
+
+def prompt_user() -> AIConfig:
+    """Prompt the user for input
+
+    Returns:
+        AIConfig: The AIConfig object tailored to the user's input
+    """
+    ai_name = ""
+    ai_config = None
+
+    # Construct the prompt
+    logger.typewriter_log(
+        "Welcome to Auto-GPT! ",
+        Fore.GREEN,
+        "run with '--help' for more information.",
+        speak_text=True,
+    )
+
+    # Get user desire
+    logger.typewriter_log(
+        "Create an AI-Assistant:",
+        Fore.GREEN,
+        "input '--manual' to enter manual mode.",
+        speak_text=True,
+    )
+
+    user_desire = utils.clean_input(
+        f"{Fore.LIGHTBLUE_EX}I want Auto-GPT to{Style.RESET_ALL}: "
+    )
+
+    if user_desire == "":
+        user_desire = DEFAULT_USER_DESIRE_PROMPT  # Default prompt
+
+    # If user desire contains "--manual"
+    if "--manual" in user_desire:
+        logger.typewriter_log(
+            "Manual Mode Selected",
+            Fore.GREEN,
+            speak_text=True,
+        )
+        return generate_aiconfig_manual()
+
+    else:
+        try:
+            return generate_aiconfig_automatic(user_desire)
+        except Exception as e:
+            logger.typewriter_log(
+                "Unable to automatically generate AI Config based on user desire.",
+                Fore.RED,
+                "Falling back to manual mode.",
+                speak_text=True,
+            )
+
+            return generate_aiconfig_manual()
+
+
+def generate_aiconfig_manual() -> AIConfig:
+    """
+    Interactively create an AI configuration by prompting the user to provide the name, role, and goals of the AI.
+
+    This function guides the user through a series of prompts to collect the necessary information to create
+    an AIConfig object. The user will be asked to provide a name and role for the AI, as well as up to five
+    goals. If the user does not provide a value for any of the fields, default values will be used.
+
+    Returns:
+        AIConfig: An AIConfig object containing the user-defined or default AI name, role, and goals.
+    """
+
+    # Manual Setup Intro
+    logger.typewriter_log(
+        "Create an AI-Assistant:",
+        Fore.GREEN,
+        "Enter the name of your AI and its role below. Entering nothing will load"
+        " defaults.",
+        speak_text=True,
+    )
+
+    # Get AI Name from User
+    logger.typewriter_log(
+        "Name your AI: ", Fore.GREEN, "For example, 'Entrepreneur-GPT'"
+    )
+    ai_name = utils.clean_input("AI Name: ")
+    if ai_name == "":
+        ai_name = "Entrepreneur-GPT"
+
+    logger.typewriter_log(
+        f"{ai_name} here!", Fore.LIGHTBLUE_EX, "I am at your service.", speak_text=True
+    )
+
+    # Get AI Role from User
+    logger.typewriter_log(
+        "Describe your AI's role: ",
+        Fore.GREEN,
+        "For example, 'an AI designed to autonomously develop and run businesses with"
+        " the sole goal of increasing your net worth.'",
+    )
+    ai_role = utils.clean_input(f"{ai_name} is: ")
+    if ai_role == "":
+        ai_role = "an AI designed to autonomously develop and run businesses with the"
+        " sole goal of increasing your net worth."
+
+    # Enter up to 5 goals for the AI
+    logger.typewriter_log(
+        "Enter up to 5 goals for your AI: ",
+        Fore.GREEN,
+        "For example: \nIncrease net worth, Grow Twitter Account, Develop and manage"
+        " multiple businesses autonomously'",
+    )
+    logger.info("Enter nothing to load defaults, enter nothing when finished.")
+    ai_goals = []
+    for i in range(5):
+        ai_goal = utils.clean_input(f"{Fore.LIGHTBLUE_EX}Goal{Style.RESET_ALL} {i+1}: ")
+        if ai_goal == "":
+            break
+        ai_goals.append(ai_goal)
+    if not ai_goals:
+        ai_goals = [
+            "Increase net worth",
+            "Grow Twitter Account",
+            "Develop and manage multiple businesses autonomously",
+        ]
+
+    # Get API Budget from User
+    logger.typewriter_log(
+        "Enter your budget for API calls: ",
+        Fore.GREEN,
+        "For example: $1.50",
+    )
+    logger.info("Enter nothing to let the AI run without monetary limit")
+    api_budget_input = utils.clean_input(
+        f"{Fore.LIGHTBLUE_EX}Budget{Style.RESET_ALL}: $"
+    )
+    if api_budget_input == "":
+        api_budget = 0.0
+    else:
+        try:
+            api_budget = float(api_budget_input.replace("$", ""))
+        except ValueError:
+            logger.typewriter_log(
+                "Invalid budget input. Setting budget to unlimited.", Fore.RED
+            )
+            api_budget = 0.0
+
+    return AIConfig(ai_name, ai_role, ai_goals, api_budget)
+
+
+def generate_aiconfig_automatic(user_prompt) -> AIConfig:
+    """Generates an AIConfig object from the given string.
+
+    Returns:
+    AIConfig: The AIConfig object tailored to the user's input
+    """
+
+    system_prompt = DEFAULT_SYSTEM_PROMPT_AICONFIG_AUTOMATIC
+    prompt_ai_config_automatic = Template(
+        DEFAULT_TASK_PROMPT_AICONFIG_AUTOMATIC
+    ).render(user_prompt=user_prompt)
+    # Call LLM with the string as user input
+    output = create_chat_completion(
+        ChatSequence.for_model(
+            CFG.fast_llm_model,
+            [
+                Message("system", system_prompt),
+                Message("user", prompt_ai_config_automatic),
+            ],
+        )
+    )
+
+    # Debug LLM Output
+    logger.debug(f"AI Config Generator Raw Output: {output}")
+
+    # Parse the output
+    ai_name = re.search(r"Name(?:\s*):(?:\s*)(.*)", output, re.IGNORECASE).group(1)
+    ai_role = (
+        re.search(
+            r"Description(?:\s*):(?:\s*)(.*?)(?:(?:\n)|Goals)",
+            output,
+            re.IGNORECASE | re.DOTALL,
+        )
+        .group(1)
+        .strip()
+    )
+    ai_goals = re.findall(r"(?<=\n)-\s*(.*)", output)
+    api_budget = 0.0  # TODO: parse api budget using a regular expression
+
+    return AIConfig(ai_name, ai_role, ai_goals, api_budget)

autogpt/singleton.py

@@ -0,0 +1,22 @@
+"""The singleton metaclass for ensuring only one instance of a class."""
+import abc
+
+
+class Singleton(abc.ABCMeta, type):
+    """
+    Singleton metaclass for ensuring only one instance of a class.
+    """
+
+    _instances = {}
+
+    def __call__(cls, *args, **kwargs):
+        """Call method for the singleton metaclass."""
+        if cls not in cls._instances:
+            cls._instances[cls] = super(Singleton, cls).__call__(*args, **kwargs)
+        return cls._instances[cls]
+
+
+class AbstractSingleton(abc.ABC, metaclass=Singleton):
+    """
+    Abstract singleton class for ensuring only one instance of a class.
+    """

autogpt/speech/__init__.py

@@ -0,0 +1,4 @@
+"""This module contains the speech recognition and speech synthesis functions."""
+from autogpt.speech.say import say_text
+
+__all__ = ["say_text"]

autogpt/speech/base.py

@@ -0,0 +1,48 @@
+"""Base class for all voice classes."""
+import abc
+from threading import Lock
+
+from autogpt.singleton import AbstractSingleton
+
+
+class VoiceBase(AbstractSingleton):
+    """
+    Base class for all voice classes.
+    """
+
+    def __init__(self):
+        """
+        Initialize the voice class.
+        """
+        self._url = None
+        self._headers = None
+        self._api_key = None
+        self._voices = []
+        self._mutex = Lock()
+        self._setup()
+
+    def say(self, text: str, voice_index: int = 0) -> bool:
+        """
+        Say the given text.
+
+        Args:
+            text (str): The text to say.
+            voice_index (int): The index of the voice to use.
+        """
+        with self._mutex:
+            return self._speech(text, voice_index)
+
+    @abc.abstractmethod
+    def _setup(self) -> None:
+        """
+        Setup the voices, API key, etc.
+        """
+
+    @abc.abstractmethod
+    def _speech(self, text: str, voice_index: int = 0) -> bool:
+        """
+        Play the given text.
+
+        Args:
+            text (str): The text to play.
+        """

autogpt/speech/brian.py

@@ -0,0 +1,42 @@
+import logging
+import os
+
+import requests
+from playsound import playsound
+
+from autogpt.speech.base import VoiceBase
+
+
+class BrianSpeech(VoiceBase):
+    """Brian speech module for autogpt"""
+
+    def _setup(self) -> None:
+        """Setup the voices, API key, etc."""
+
+    def _speech(self, text: str, _: int = 0) -> bool:
+        """Speak text using Brian with the streamelements API
+
+        Args:
+            text (str): The text to speak
+
+        Returns:
+            bool: True if the request was successful, False otherwise
+        """
+        tts_url = (
+            f"https://api.streamelements.com/kappa/v2/speech?voice=Brian&text={text}"
+        )
+        response = requests.get(tts_url)
+
+        if response.status_code == 200:
+            with open("speech.mp3", "wb") as f:
+                f.write(response.content)
+            playsound("speech.mp3")
+            os.remove("speech.mp3")
+            return True
+        else:
+            logging.error(
+                "Request failed with status code: %s, response content: %s",
+                response.status_code,
+                response.content,
+            )
+            return False

autogpt/speech/eleven_labs.py

@@ -0,0 +1,88 @@
+"""ElevenLabs speech module"""
+import os
+
+import requests
+from playsound import playsound
+
+from autogpt.config.config import Config
+from autogpt.speech.base import VoiceBase
+
+PLACEHOLDERS = {"your-voice-id"}
+
+
+class ElevenLabsSpeech(VoiceBase):
+    """ElevenLabs speech class"""
+
+    def _setup(self) -> None:
+        """Set up the voices, API key, etc.
+
+        Returns:
+            None: None
+        """
+
+        cfg = Config()
+        default_voices = ["ErXwobaYiN019PkySvjV", "EXAVITQu4vr4xnSDxMaL"]
+        voice_options = {
+            "Rachel": "21m00Tcm4TlvDq8ikWAM",
+            "Domi": "AZnzlk1XvdvUeBnXmlld",
+            "Bella": "EXAVITQu4vr4xnSDxMaL",
+            "Antoni": "ErXwobaYiN019PkySvjV",
+            "Elli": "MF3mGyEYCl7XYWbV9V6O",
+            "Josh": "TxGEqnHWrfWFTfGW9XjX",
+            "Arnold": "VR6AewLTigWG4xSOukaG",
+            "Adam": "pNInz6obpgDQGcFmaJgB",
+            "Sam": "yoZ06aMxZJJ28mfd3POQ",
+        }
+        self._headers = {
+            "Content-Type": "application/json",
+            "xi-api-key": cfg.elevenlabs_api_key,
+        }
+        self._voices = default_voices.copy()
+        if cfg.elevenlabs_voice_1_id in voice_options:
+            cfg.elevenlabs_voice_1_id = voice_options[cfg.elevenlabs_voice_1_id]
+        if cfg.elevenlabs_voice_2_id in voice_options:
+            cfg.elevenlabs_voice_2_id = voice_options[cfg.elevenlabs_voice_2_id]
+        self._use_custom_voice(cfg.elevenlabs_voice_1_id, 0)
+        self._use_custom_voice(cfg.elevenlabs_voice_2_id, 1)
+
+    def _use_custom_voice(self, voice, voice_index) -> None:
+        """Use a custom voice if provided and not a placeholder
+
+        Args:
+            voice (str): The voice ID
+            voice_index (int): The voice index
+
+        Returns:
+            None: None
+        """
+        # Placeholder values that should be treated as empty
+        if voice and voice not in PLACEHOLDERS:
+            self._voices[voice_index] = voice
+
+    def _speech(self, text: str, voice_index: int = 0) -> bool:
+        """Speak text using elevenlabs.io's API
+
+        Args:
+            text (str): The text to speak
+            voice_index (int, optional): The voice to use. Defaults to 0.
+
+        Returns:
+            bool: True if the request was successful, False otherwise
+        """
+        from autogpt.logs import logger
+
+        tts_url = (
+            f"https://api.elevenlabs.io/v1/text-to-speech/{self._voices[voice_index]}"
+        )
+        response = requests.post(tts_url, headers=self._headers, json={"text": text})
+
+        if response.status_code == 200:
+            with open("speech.mpeg", "wb") as f:
+                f.write(response.content)
+            playsound("speech.mpeg", True)
+            os.remove("speech.mpeg")
+            return True
+        else:
+            logger.warn("Request failed with status code:", response.status_code)
+            logger.info("Response content:", response.content)
+            return False

autogpt/speech/gtts.py

@@ -0,0 +1,22 @@
+""" GTTS Voice. """
+import os
+
+import gtts
+from playsound import playsound
+
+from autogpt.speech.base import VoiceBase
+
+
+class GTTSVoice(VoiceBase):
+    """GTTS Voice."""
+
+    def _setup(self) -> None:
+        pass
+
+    def _speech(self, text: str, _: int = 0) -> bool:
+        """Play the given text."""
+        tts = gtts.gTTS(text)
+        tts.save("speech.mp3")
+        playsound("speech.mp3", True)
+        os.remove("speech.mp3")
+        return True

autogpt/speech/macos_tts.py

@@ -0,0 +1,21 @@
+""" MacOS TTS Voice. """
+import os
+
+from autogpt.speech.base import VoiceBase
+
+
+class MacOSTTS(VoiceBase):
+    """MacOS TTS Voice."""
+
+    def _setup(self) -> None:
+        pass
+
+    def _speech(self, text: str, voice_index: int = 0) -> bool:
+        """Play the given text."""
+        if voice_index == 0:
+            os.system(f'say "{text}"')
+        elif voice_index == 1:
+            os.system(f'say -v "Ava (Premium)" "{text}"')
+        else:
+            os.system(f'say -v Samantha "{text}"')
+        return True

autogpt/speech/say.py

@@ -0,0 +1,46 @@
+""" Text to speech module """
+import threading
+from threading import Semaphore
+
+from autogpt.config.config import Config
+from autogpt.speech.base import VoiceBase
+from autogpt.speech.brian import BrianSpeech
+from autogpt.speech.eleven_labs import ElevenLabsSpeech
+from autogpt.speech.gtts import GTTSVoice
+from autogpt.speech.macos_tts import MacOSTTS
+
+_QUEUE_SEMAPHORE = Semaphore(
+    1
+)  # The amount of sounds to queue before blocking the main thread
+
+
+def say_text(text: str, voice_index: int = 0) -> None:
+    """Speak the given text using the given voice index"""
+    cfg = Config()
+    default_voice_engine, voice_engine = _get_voice_engine(cfg)
+
+    def speak() -> None:
+        success = voice_engine.say(text, voice_index)
+        if not success:
+            default_voice_engine.say(text)
+
+        _QUEUE_SEMAPHORE.release()
+
+    _QUEUE_SEMAPHORE.acquire(True)
+    thread = threading.Thread(target=speak)
+    thread.start()
+
+
+def _get_voice_engine(config: Config) -> tuple[VoiceBase, VoiceBase]:
+    """Get the voice engine to use for the given configuration"""
+    default_voice_engine = GTTSVoice()
+    if config.elevenlabs_api_key:
+        voice_engine = ElevenLabsSpeech()
+    elif config.use_mac_os_tts == "True":
+        voice_engine = MacOSTTS()
+    elif config.use_brian_tts == "True":
+        voice_engine = BrianSpeech()
+    else:
+        voice_engine = GTTSVoice()
+
+    return default_voice_engine, voice_engine

autogpt/spinner.py

@@ -0,0 +1,76 @@
+"""A simple spinner module"""
+import itertools
+import sys
+import threading
+import time
+
+
+class Spinner:
+    """A simple spinner class"""
+
+    def __init__(
+        self,
+        message: str = "Loading...",
+        delay: float = 0.1,
+        plain_output: bool = False,
+    ) -> None:
+        """Initialize the spinner class
+
+        Args:
+            message (str): The message to display.
+            delay (float): The delay between each spinner update.
+            plain_output (bool): Whether to display the spinner or not.
+        """
+        self.plain_output = plain_output
+        self.spinner = itertools.cycle(["-", "/", "|", "\\"])
+        self.delay = delay
+        self.message = message
+        self.running = False
+        self.spinner_thread = None
+
+    def spin(self) -> None:
+        """Spin the spinner"""
+        if self.plain_output:
+            self.print_message()
+            return
+        while self.running:
+            self.print_message()
+            time.sleep(self.delay)
+
+    def print_message(self):
+        sys.stdout.write(f"\r{' ' * (len(self.message) + 2)}\r")
+        sys.stdout.write(f"{next(self.spinner)} {self.message}\r")
+        sys.stdout.flush()
+
+    def __enter__(self):
+        """Start the spinner"""
+        self.running = True
+        self.spinner_thread = threading.Thread(target=self.spin)
+        self.spinner_thread.start()
+
+        return self
+
+    def __exit__(self, exc_type, exc_value, exc_traceback) -> None:
+        """Stop the spinner
+
+        Args:
+            exc_type (Exception): The exception type.
+            exc_value (Exception): The exception value.
+            exc_traceback (Exception): The exception traceback.
+        """
+        self.running = False
+        if self.spinner_thread is not None:
+            self.spinner_thread.join()
+        sys.stdout.write(f"\r{' ' * (len(self.message) + 2)}\r")
+        sys.stdout.flush()
+
+    def update_message(self, new_message, delay=0.1):
+        """Update the spinner message
+        Args:
+            new_message (str): New message to display.
+            delay (float): The delay in seconds between each spinner update.
+        """
+        self.delay = delay
+        self.message = new_message
+        if self.plain_output:
+            self.print_message()

autogpt/url_utils/validators.py

@@ -0,0 +1,107 @@
+import functools
+import re
+from typing import Any, Callable
+from urllib.parse import urljoin, urlparse
+
+from requests.compat import urljoin
+
+
+def validate_url(func: Callable[..., Any]) -> Any:
+    """The method decorator validate_url is used to validate urls for any command that requires
+    a url as an argument"""
+
+    @functools.wraps(func)
+    def wrapper(url: str, *args, **kwargs) -> Any:
+        """Check if the URL is valid using a basic check, urllib check, and local file check
+
+        Args:
+            url (str): The URL to check
+
+        Returns:
+            the result of the wrapped function
+
+        Raises:
+            ValueError if the url fails any of the validation tests
+        """
+        # Most basic check if the URL is valid:
+        if not re.match(r"^https?://", url):
+            raise ValueError("Invalid URL format")
+        if not is_valid_url(url):
+            raise ValueError("Missing Scheme or Network location")
+        # Restrict access to local files
+        if check_local_file_access(url):
+            raise ValueError("Access to local files is restricted")
+        # Check URL length
+        if len(url) > 2000:
+            raise ValueError("URL is too long")
+
+        return func(sanitize_url(url), *args, **kwargs)
+
+    return wrapper
+
+
+def is_valid_url(url: str) -> bool:
+    """Check if the URL is valid
+
+    Args:
+        url (str): The URL to check
+
+    Returns:
+        bool: True if the URL is valid, False otherwise
+    """
+    try:
+        result = urlparse(url)
+        return all([result.scheme, result.netloc])
+    except ValueError:
+        return False
+
+
+def sanitize_url(url: str) -> str:
+    """Sanitize the URL
+
+    Args:
+        url (str): The URL to sanitize
+
+    Returns:
+        str: The sanitized URL
+    """
+    parsed_url = urlparse(url)
+    reconstructed_url = f"{parsed_url.path}{parsed_url.params}?{parsed_url.query}"
+    return urljoin(url, reconstructed_url)
+
+
+def check_local_file_access(url: str) -> bool:
+    """Check if the URL is a local file
+
+    Args:
+        url (str): The URL to check
+
+    Returns:
+        bool: True if the URL is a local file, False otherwise
+    """
+    local_prefixes = [
+        "file:///",
+        "file://localhost/",
+        "file://localhost",
+        "http://localhost",
+        "http://localhost/",
+        "https://localhost",
+        "https://localhost/",
+        "http://2130706433",
+        "http://2130706433/",
+        "https://2130706433",
+        "https://2130706433/",
+        "http://127.0.0.1/",
+        "http://127.0.0.1",
+        "https://127.0.0.1/",
+        "https://127.0.0.1",
+        "https://0.0.0.0/",
+        "https://0.0.0.0",
+        "http://0.0.0.0/",
+        "http://0.0.0.0",
+        "http://0000",
+        "http://0000/",
+        "https://0000",
+        "https://0000/",
+    ]
+    return any(url.startswith(prefix) for prefix in local_prefixes)

autogpt/utils.py

@@ -0,0 +1,178 @@
+import os
+import re
+
+import requests
+import yaml
+from colorama import Fore, Style
+from git.repo import Repo
+
+from autogpt.config import Config
+from autogpt.logs import logger
+
+
+def batch(iterable, max_batch_length: int, overlap: int = 0):
+    """Batch data from iterable into slices of length N. The last batch may be shorter."""
+    # batched('ABCDEFG', 3) --> ABC DEF G
+    if max_batch_length < 1:
+        raise ValueError("n must be at least one")
+    for i in range(0, len(iterable), max_batch_length - overlap):
+        yield iterable[i : i + max_batch_length]
+
+
+def clean_input(prompt: str = "", talk=False):
+    try:
+        cfg = Config()
+        if cfg.chat_messages_enabled:
+            for plugin in cfg.plugins:
+                if not hasattr(plugin, "can_handle_user_input"):
+                    continue
+                if not plugin.can_handle_user_input(user_input=prompt):
+                    continue
+                plugin_response = plugin.user_input(user_input=prompt)
+                if not plugin_response:
+                    continue
+                if plugin_response.lower() in [
+                    "yes",
+                    "yeah",
+                    "y",
+                    "ok",
+                    "okay",
+                    "sure",
+                    "alright",
+                ]:
+                    return cfg.authorise_key
+                elif plugin_response.lower() in [
+                    "no",
+                    "nope",
+                    "n",
+                    "negative",
+                ]:
+                    return cfg.exit_key
+                return plugin_response
+
+        # ask for input, default when just pressing Enter is y
+        logger.info("Asking user via keyboard...")
+        answer = input(prompt)
+        return answer
+    except KeyboardInterrupt:
+        logger.info("You interrupted Auto-GPT")
+        logger.info("Quitting...")
+        exit(0)
+
+
+def validate_yaml_file(file: str):
+    try:
+        with open(file, encoding="utf-8") as fp:
+            yaml.load(fp.read(), Loader=yaml.FullLoader)
+    except FileNotFoundError:
+        return (False, f"The file {Fore.CYAN}`{file}`{Fore.RESET} wasn't found")
+    except yaml.YAMLError as e:
+        return (
+            False,
+            f"There was an issue while trying to read with your AI Settings file: {e}",
+        )
+
+    return (True, f"Successfully validated {Fore.CYAN}`{file}`{Fore.RESET}!")
+
+
+def readable_file_size(size, decimal_places=2):
+    """Converts the given size in bytes to a readable format.
+    Args:
+        size: Size in bytes
+        decimal_places (int): Number of decimal places to display
+    """
+    for unit in ["B", "KB", "MB", "GB", "TB"]:
+        if size < 1024.0:
+            break
+        size /= 1024.0
+    return f"{size:.{decimal_places}f} {unit}"
+
+
+def get_bulletin_from_web():
+    try:
+        response = requests.get(
+            "https://raw.githubusercontent.com/Significant-Gravitas/Auto-GPT/master/BULLETIN.md"
+        )
+        if response.status_code == 200:
+            return response.text
+    except requests.exceptions.RequestException:
+        pass
+
+    return ""
+
+
+def get_current_git_branch() -> str:
+    try:
+        repo = Repo(search_parent_directories=True)
+        branch = repo.active_branch
+        return branch.name
+    except:
+        return ""
+
+
+def get_latest_bulletin() -> tuple[str, bool]:
+    exists = os.path.exists("data/CURRENT_BULLETIN.md")
+    current_bulletin = ""
+    if exists:
+        current_bulletin = open(
+            "data/CURRENT_BULLETIN.md", "r", encoding="utf-8"
+        ).read()
+    new_bulletin = get_bulletin_from_web()
+    is_new_news = new_bulletin != "" and new_bulletin != current_bulletin
+
+    news_header = Fore.YELLOW + "Welcome to Auto-GPT!\n"
+    if new_bulletin or current_bulletin:
+        news_header += (
+            "Below you'll find the latest Auto-GPT News and updates regarding features!\n"
+            "If you don't wish to see this message, you "
+            "can run Auto-GPT with the *--skip-news* flag.\n"
+        )
+
+    if new_bulletin and is_new_news:
+        open("data/CURRENT_BULLETIN.md", "w", encoding="utf-8").write(new_bulletin)
+        current_bulletin = f"{Fore.RED}::NEW BULLETIN::{Fore.RESET}\n\n{new_bulletin}"
+
+    return f"{news_header}\n{current_bulletin}", is_new_news
+
+
+def markdown_to_ansi_style(markdown: str):
+    ansi_lines: list[str] = []
+    for line in markdown.split("\n"):
+        line_style = ""
+
+        if line.startswith("# "):
+            line_style += Style.BRIGHT
+        else:
+            line = re.sub(
+                r"(?<!\*)\*(\*?[^*]+\*?)\*(?!\*)",
+                rf"{Style.BRIGHT}\1{Style.NORMAL}",
+                line,
+            )
+
+        if re.match(r"^#+ ", line) is not None:
+            line_style += Fore.CYAN
+            line = re.sub(r"^#+ ", "", line)
+
+        ansi_lines.append(f"{line_style}{line}{Style.RESET_ALL}")
+    return "\n".join(ansi_lines)
+
+
+def get_legal_warning() -> str:
+    legal_text = """
+## DISCLAIMER AND INDEMNIFICATION AGREEMENT
+### PLEASE READ THIS DISCLAIMER AND INDEMNIFICATION AGREEMENT CAREFULLY BEFORE USING THE AUTOGPT SYSTEM. BY USING THE AUTOGPT SYSTEM, YOU AGREE TO BE BOUND BY THIS AGREEMENT.
+
+## Introduction
+AutoGPT (the "System") is a project that connects a GPT-like artificial intelligence system to the internet and allows it to automate tasks. While the System is designed to be useful and efficient, there may be instances where the System could perform actions that may cause harm or have unintended consequences.
+
+## No Liability for Actions of the System
+The developers, contributors, and maintainers of the AutoGPT project (collectively, the "Project Parties") make no warranties or representations, express or implied, about the System's performance, accuracy, reliability, or safety. By using the System, you understand and agree that the Project Parties shall not be liable for any actions taken by the System or any consequences resulting from such actions.
+
+## User Responsibility and Respondeat Superior Liability
+As a user of the System, you are responsible for supervising and monitoring the actions of the System while it is operating on your
+behalf. You acknowledge that using the System could expose you to potential liability including but not limited to respondeat superior and you agree to assume all risks and liabilities associated with such potential liability.
+
+## Indemnification
+By using the System, you agree to indemnify, defend, and hold harmless the Project Parties from and against any and all claims, liabilities, damages, losses, or expenses (including reasonable attorneys' fees and costs) arising out of or in connection with your use of the System, including, without limitation, any actions taken by the System on your behalf, any failure to properly supervise or monitor the System, and any resulting harm or unintended consequences.
+            """
+    return legal_text

autogpt/workspace/__init__.py

@@ -0,0 +1,5 @@
+from autogpt.workspace.workspace import Workspace
+
+__all__ = [
+    "Workspace",
+]

autogpt/workspace/workspace.py

@@ -0,0 +1,138 @@
+"""
+=========
+Workspace
+=========
+
+The workspace is a directory containing configuration and working files for an AutoGPT
+agent.
+
+"""
+from __future__ import annotations
+
+from pathlib import Path
+
+from autogpt.logs import logger
+
+
+class Workspace:
+    """A class that represents a workspace for an AutoGPT agent."""
+
+    NULL_BYTES = ["\0", "\000", "\x00", r"\z", "\u0000", "%00"]
+
+    def __init__(self, workspace_root: str | Path, restrict_to_workspace: bool):
+        self._root = self._sanitize_path(workspace_root)
+        self._restrict_to_workspace = restrict_to_workspace
+
+    @property
+    def root(self) -> Path:
+        """The root directory of the workspace."""
+        return self._root
+
+    @property
+    def restrict_to_workspace(self):
+        """Whether to restrict generated paths to the workspace."""
+        return self._restrict_to_workspace
+
+    @classmethod
+    def make_workspace(cls, workspace_directory: str | Path, *args, **kwargs) -> Path:
+        """Create a workspace directory and return the path to it.
+
+        Parameters
+        ----------
+        workspace_directory
+            The path to the workspace directory.
+
+        Returns
+        -------
+        Path
+            The path to the workspace directory.
+
+        """
+        # TODO: have this make the env file and ai settings file in the directory.
+        workspace_directory = cls._sanitize_path(workspace_directory)
+        workspace_directory.mkdir(exist_ok=True, parents=True)
+        return workspace_directory
+
+    def get_path(self, relative_path: str | Path) -> Path:
+        """Get the full path for an item in the workspace.
+
+        Parameters
+        ----------
+        relative_path
+            The relative path to resolve in the workspace.
+
+        Returns
+        -------
+        Path
+            The resolved path relative to the workspace.
+
+        """
+        return self._sanitize_path(
+            relative_path,
+            root=self.root,
+            restrict_to_root=self.restrict_to_workspace,
+        )
+
+    @staticmethod
+    def _sanitize_path(
+        relative_path: str | Path,
+        root: str | Path = None,
+        restrict_to_root: bool = True,
+    ) -> Path:
+        """Resolve the relative path within the given root if possible.
+
+        Parameters
+        ----------
+        relative_path
+            The relative path to resolve.
+        root
+            The root path to resolve the relative path within.
+        restrict_to_root
+            Whether to restrict the path to the root.
+
+        Returns
+        -------
+        Path
+            The resolved path.
+
+        Raises
+        ------
+        ValueError
+            If the path is absolute and a root is provided.
+        ValueError
+            If the path is outside the root and the root is restricted.
+
+        """
+
+        # Posix systems disallow null bytes in paths. Windows is agnostic about it.
+        # Do an explicit check here for all sorts of null byte representations.
+
+        for null_byte in Workspace.NULL_BYTES:
+            if null_byte in str(relative_path) or null_byte in str(root):
+                raise ValueError("embedded null byte")
+
+        if root is None:
+            return Path(relative_path).resolve()
+
+        logger.debug(f"Resolving path '{relative_path}' in workspace '{root}'")
+
+        root, relative_path = Path(root).resolve(), Path(relative_path)
+
+        logger.debug(f"Resolved root as '{root}'")
+
+        # Allow exception for absolute paths if they are contained in your workspace directory.
+        if relative_path.is_absolute() and not relative_path.is_relative_to(root):
+            raise ValueError(
+                f"Attempted to access absolute path '{relative_path}' in workspace '{root}'."
+            )
+
+        full_path = root.joinpath(relative_path).resolve()
+
+        logger.debug(f"Joined paths as '{full_path}'")
+
+        if restrict_to_root and not full_path.is_relative_to(root):
+            raise ValueError(
+                f"Attempted to access path '{full_path}' outside of workspace '{root}'."
+            )
+
+        return full_path