Skip to content

armgit / camel

Public Imported from GitHub
0

🐫 CAMEL: The first and the best multi-agent framework. Finding the Scaling Law of Agents. https://www.camel-ai.org

11.4 KB
Raw Blame
# ========= Copyright 2023-2026 @ CAMEL-AI.org. All Rights Reserved. =========
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ========= Copyright 2023-2026 @ CAMEL-AI.org. All Rights Reserved. =========
import asyncio
import logging
import queue
import threading
from typing import TYPE_CHECKING, List, Optional, Union

from slack_bolt.context.async_context import AsyncBoltContext
from slack_bolt.context.say.async_say import AsyncSay
from slack_sdk.web.async_client import AsyncWebClient

from camel.agents import ChatAgent
from camel.bots import SlackApp, SlackEventBody
from camel.retrievers import AutoRetriever
from camel.types import StorageType

if TYPE_CHECKING:
    from unstructured.documents.elements import Element

logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)


class BotAgent:
    def __init__(
        self,
        contents: Union[str, List[str], "Element", List["Element"]] = None,
        auto_retriever: Optional[AutoRetriever] = None,
        similarity_threshold: float = 0.5,
        vector_storage_local_path: str = "local_data/",
        top_k: int = 1,
        return_detailed_info: bool = True,
    ):
        r"""Initialize the BotAgent instance.

        Args:
            contents (Union[str, List[str], Element, List[Element]], optional)
                : The content to be retrieved.
            auto_retriever (Optional[AutoRetriever], optional): An instance of
                AutoRetriever for vector search.
            similarity_threshold (float): Threshold for vector similarity when
                retrieving content.
            vector_storage_local_path (str): Path to local vector storage for
                the retriever.
            top_k (int): Number of top results to retrieve.
            return_detailed_info (bool): Whether to return detailed
                information from the retriever.
        """

        assistant_sys_msg = '''
            Objective:
                You are a customer service bot designed to assist users
                with inquiries related to our open-source project.
                Your responses should be informative, concise, and helpful.

            Instructions:
                Understand User Queries: Carefully read and understand the
                        user's question. Focus on keywords and context to
                        determine the user's intent.
                Search for Relevant Information: Use the provided dataset
                        and refer to the RAG (file to find answers that
                        closely match the user's query. The RAG file
                        contains detailed interactions and should be your
                        primary resource for crafting responses.
                Provide Clear and Concise Responses: Your answers should
                        be clear and to the point. Avoid overly technical
                        language unless the user's query indicates
                        familiarity with technical terms.
                Encourage Engagement: Where applicable, encourage users
                        to contribute to the project or seek further
                        assistance.

            Response Structure:
                Greeting: Begin with a polite greeting or acknowledgment.
                Main Response: Provide the main answer to the user's query.
                Additional Information: Offer any extra tips or direct the
                        user to additional resources if necessary.
                Closing: Close the response politely, encouraging
                        further engagement if appropriate.
            bd
            Tone:
                Professional: Maintain a professional tone that
                        instills confidence in the user.
                Friendly: Be approachable and friendly to make users
                        feel comfortable.
                Helpful: Always aim to be as helpful as possible,
                        guiding users to solutions.
        '''

        self._agent = ChatAgent(
            assistant_sys_msg,
        )

        self._auto_retriever = None
        self._contents = contents
        self._top_k = top_k
        self._similarity_threshold = similarity_threshold
        self._return_detailed_info = return_detailed_info

        self._auto_retriever = auto_retriever or AutoRetriever(
            vector_storage_local_path=vector_storage_local_path,
            storage_type=StorageType.QDRANT,
        )

    async def process(self, message: str) -> str:
        r"""Process the user message, retrieve relevant content, and generate
        a response.

        Args:
            message (str): The user's query message.

        Returns:
            str: The assistant's response message.
        """
        user_raw_msg = message
        print("User message:", user_raw_msg)
        if self._auto_retriever:
            retrieved_content = self._auto_retriever.run_vector_retriever(
                query=user_raw_msg,
                contents=self._contents,
                top_k=self._top_k,
                similarity_threshold=self._similarity_threshold,
                return_detailed_info=self._return_detailed_info,
            )
            user_raw_msg = (
                f"Here is the query to you: {user_raw_msg}\n"
                f"Based on the retrieved content: {retrieved_content}, \n"
                f"answer the query"
            )

        assistant_response = self._agent.step(user_raw_msg)
        return assistant_response.msg.content


class SlackBot(SlackApp):
    r"""SlackBot class that extends the SlackApp class to handle Slack events
    and integrate with a message queue for asynchronous processing.

    This class initializes the Slack app and adds a message handler to process
    Slack events, specifically for incoming messages.

    Args:
        msg_queue (queue.Queue): A thread-safe queue to communicate between
            threads.
        token (Optional[str]): Slack API token for authentication.
        scopes (Optional[str]): Slack app scopes for permissions.
        signing_secret (Optional[str]): Signing secret for verifying Slack
            requests.
        client_id (Optional[str]): Slack app client ID.
        client_secret (Optional[str]): Slack app client secret.
    """

    def __init__(
        self,
        msg_queue: queue.Queue,
        token: Optional[str] = None,
        scopes: Optional[str] = None,
        signing_secret: Optional[str] = None,
        client_id: Optional[str] = None,
        client_secret: Optional[str] = None,
    ):
        r"""Initializes the SlackBot instance with a message queue and the
        required Slack authentication details.

        Args:
            msg_queue (queue.Queue): A thread-safe queue to communicate between
                threads.
            token (Optional[str]): Slack API token for authentication.
            scopes (Optional[str]): Slack app scopes for permissions.
            signing_secret (Optional[str]): Signing secret for verifying Slack
                requests.
            client_id (Optional[str]): Slack app client ID.
            client_secret (Optional[str]): Slack app client secret.
        """
        super().__init__(
            token=token,
            scopes=scopes,
            signing_secret=signing_secret,
            client_id=client_id,
            client_secret=client_secret,
        )
        self._queue: queue.Queue = msg_queue

    async def on_message(
        self,
        context: "AsyncBoltContext",
        client: "AsyncWebClient",
        event: dict,
        body: dict,
        say: "AsyncSay",
    ):
        r"""Event handler for processing incoming Slack messages.

        This method is called when a message event is received from Slack.
        It acknowledges the message and adds the event body and say function to
        the queue for further processing.

        Args:
            context (AsyncBoltContext): Context object that contains
                information about the Slack event.
            client (AsyncWebClient): Slack Web API client for making API
                requests.
            event (dict): Event data containing details of the incoming Slack
                message.
            body (dict): Full request body from Slack.
            say (AsyncSay): A function to send a response back to the Slack
                channel.
        """
        await context.ack()
        event_body = SlackEventBody(**body)
        self._queue.put((event_body, say))


async def process_message(agent: BotAgent, msg_queue: queue.Queue):
    r"""Process messages from the queue asynchronously.

    This function continuously retrieves messages from the message queue,
    processes them using the `BotAgent` instance, and sends the response back
    to Slack using the `say` function.

    Args:
        agent (BotAgent): An instance of the `BotAgent` class that handles the
            processing of the incoming message.
        msg_queue (queue.Queue): A thread-safe queue that stores Slack event
            messages and the corresponding `say` functions for response.
    """
    while True:
        event_body, say = msg_queue.get()

        logger.info(f"Received message: {event_body.event.text}")
        user_raw_msg = event_body.event.text

        # Process the message using the agent and send response back to Slack.
        response = await agent.process(user_raw_msg)
        await say(response)
        msg_queue.task_done()


def start_async_queue_processor(agent: BotAgent, msg_queue: queue.Queue):
    r"""Start an asynchronous queue processor in a new event loop.

    This function creates a new asyncio event loop in a separate thread to
    asynchronously process messages from the queue. It will run until all
    the messages in the queue have been processed.

    Args:
        agent (BotAgent): The agent responsible for processing the messages.
        msg_queue (queue.Queue): The message queue that contains Slack events
            and responses.
    """
    loop = asyncio.new_event_loop()
    asyncio.set_event_loop(loop)

    # Schedule the asynchronous message processing task in this event loop
    loop.run_until_complete(process_message(agent, msg_queue))


if __name__ == "__main__":
    r"""Main entry point for running the Slack bot application.

    This section initializes the required components including the message
    queue, agent, and the SlackBot instance. It also starts a separate thread
    for asynchronous message processing to avoid blocking the Slack bot's main
    event loop. The `slack_bot.run()` function will handle incoming Slack
    events on the main thread, while the separate thread will handle processing
    the messages from the queue.
    """
    msg_queue = queue.Queue()

    agent = BotAgent()

    thread = threading.Thread(
        target=start_async_queue_processor, args=(agent, msg_queue)
    )
    thread.start()

    # Initialize the SlackBot with the message queue.
    slack_bot = SlackBot(msg_queue=msg_queue)
    slack_bot.run(3000)

    thread.join()