Get the FREE Ultimate OpenClaw Setup Guide →
w

Pywayne Lark Bot Listener

Verified

@wangyendt

npx machina-cli add skill @wangyendt/lark-bot-listener --openclaw
Files (1)
SKILL.md
5.6 KB

Pywayne Lark Bot Listener

飞书消息监听器,通过 WebSocket 实时接收和处理飞书消息。

Quick Start

from pywayne.lark_bot_listener import LarkBotListener

# 初始化监听器
listener = LarkBotListener(
    app_id="your_app_id",
    app_secret="your_app_secret",
    message_expiry_time=60  # 消息去重过期时间(秒)
)

# 处理文本消息
@listener.text_handler()
async def handle_text(text: str):
    print(f"收到消息: {text}")

# 启动监听
listener.run()

Decorators - 消息处理装饰器

text_handler

文本消息处理装饰器,直接传递文本内容。

@listener.text_handler(group_only=False, user_only=False)
async def handle_text(text: str, chat_id: str, is_group: bool, group_name: str, user_name: str):
    print(f"收到来自 {user_name} 的消息: {text}")
    listener.send_message(chat_id, f"已收到:{text}")

参数说明(除 text 外均为可选):

参数类型说明
textstr文本内容(必需)
chat_idstr会话 ID
is_groupbool是否群组消息
group_namestr群组名称(私聊时为空)
user_namestr发送消息的用户姓名

装饰器参数

  • group_only=True: 只处理群组消息
  • user_only=True: 只处理私聊消息

image_handler

图片消息处理装饰器,自动下载图片到临时文件并清理。

from pathlib import Path
import cv2
import tempfile

@listener.image_handler()
async def handle_image(image_path: Path, user_name: str) -> Path:
    # 处理图片
    img = cv2.imread(str(image_path))
    # ...处理逻辑...
    # 返回新图片路径会自动发送回去,返回 None 则不发送
    return image_path

参数说明(除 image_path 外均为可选):

参数类型说明
image_pathPath临时图片文件路径(必需)
chat_idstr会话 ID
is_groupbool是否群组消息
group_namestr群组名称
user_namestr发送消息的用户姓名

返回值

  • 返回 Path: 自动上传并发送新图片
  • 返回 None: 不发送任何图片

file_handler

文件消息处理装饰器,自动下载文件到临时文件并清理。

@listener.file_handler()
async def handle_file(file_path: Path, user_name: str) -> Path:
    # 处理文件
    with open(file_path, 'r') as f:
        content = f.read()
    # ...处理逻辑...
    return file_path  # 返回文件路径会自动发送回去

参数说明(除 file_path 外均为可选):

参数类型说明
file_pathPath临时文件路径(必需)
chat_idstr会话 ID
is_groupbool是否群组消息
group_namestr群组名称
user_namestr发送消息的用户姓名

listen - 通用消息监听器

用于监听任意类型消息(包括富文本 post)。

@listener.listen(message_type="post")
async def handle_post(ctx: MessageContext):
    print(f"收到富文本消息: {ctx.content}")

MessageContext 属性

属性类型说明
chat_idstr会话 ID
user_idstr用户 ID
message_typestr消息类型
contentstr消息内容(文本消息为字符串,其他类型为 JSON 字符串)
is_groupbool是否群组消息
chat_typestr会话类型
message_idstr消息 ID

send_message - 发送消息

发送 Markdown 格式的消息到飞书。

listener.send_message(chat_id, "**这是加粗文本**")
listener.send_message(chat_id, "普通文本\n[链接](https://example.com)")

完整示例

from pywayne.lark_bot_listener import LarkBotListener
from pathlib import Path

listener = LarkBotListener(
    app_id="your_app_id",
    app_secret="your_app_secret"
)

# 文本消息 - AI 回复示例
@listener.text_handler()
async def handle_text(text: str, chat_id: str, user_name: str):
    # 处理逻辑...
    listener.send_message(chat_id, f"收到来自 {user_name} 的消息")

# 图片消息 - 自动下载和清理
@listener.image_handler()
async def handle_image(image_path: Path):
    # image_path 是临时文件,处理完会自动清理
    print(f"处理图片: {image_path}")

# 文件消息 - 自动下载和清理
@listener.file_handler(group_only=True)
async def handle_file(file_path: Path, group_name: str):
    print(f"收到文件: {file_path} 来自 {group_name}")

# 富文本消息
@listener.listen(message_type="post")
async def handle_post(ctx: MessageContext):
    import json
    post_content = json.loads(ctx.content)
    print(f"收到富文本: {post_content}")

# 启动监听
listener.run()

注意事项

  1. 异步处理: 所有处理函数使用 async/await 语法
  2. 消息去重: 每个处理函数独立去重,默认 60 秒过期
  3. 临时文件: 图片和文件下载到 系统临时目录/lark_bot_temp,处理完自动清理
  4. 错误隔离: 每个处理函数异常独立捕获,不影响其他函数
  5. 多注册: 同一消息可被多个处理函数处理

Source

git clone https://clawhub.ai/wangyendt/lark-bot-listenerView on GitHub

Overview

Pywayne Lark Bot Listener enables real-time Feishu/Lark message processing via WebSocket. It supports text, image, file, and post messages with automatic deduplication and optional auto-replies. Decorators simplify creating handlers and manage media download/cleanup for you.

How This Skill Works

The listener connects to Feishu via WebSocket and dispatches incoming messages to user-defined handlers. It includes message_expiry_time for deduplication, and provides decorators (text_handler, image_handler, file_handler) that download media to temporary files, clean up afterward, and can return a Path to auto-send updated media. A generic listen() handler supports rich post messages as well.

When to Use It

  • You need real-time processing of incoming Feishu messages (text, image, file, post) as they arrive.
  • Deduplication is important to avoid handling the same message more than once.
  • You want convenient decorators that automatically download media and clean up after processing.
  • You require optional automatic replies after handling a message.
  • You prefer a unified API that covers text, media, and rich post messages.

Quick Start

  1. Step 1: pip install pywayne-lark-bot-listener
  2. Step 2: Initialize LarkBotListener with app_id, app_secret, and optional message_expiry_time
  3. Step 3: Define handlers using @listener.text_handler(), @listener.image_handler(), @listener.file_handler(), and start with listener.run()

Best Practices

  • Set an appropriate message_expiry_time to balance dedup safety with resource usage.
  • Test handlers with all message types (text, image, file, post) to ensure correct processing.
  • Use group_only and user_only flags to scope handlers to the right conversations.
  • When handling media, return a Path to enable automatic sending of updated media; return None to skip sending.
  • Call listener.run() within your async event loop and handle errors gracefully.

Example Use Cases

  • Text messages: initialize the listener and handle_text to process incoming text and respond via listener.send_message.
  • Image messages: use image_handler to auto-download the image to a temp path, process it, and optionally return a new Path to resend.
  • File messages: use file_handler to download files to a temp path, read or analyze, and optionally resend the file.
  • Rich posts: listen for post-type messages with @listener.listen(message_type="post") to handle 富文本内容.
  • Comprehensive demo: combine text_handler, image_handler, file_handler, and listen for post, then call listener.run().

Frequently Asked Questions

Add this skill to your agents
Sponsor this space

Reach thousands of developers