feat: add wechat company service account

2026-01-19 01:21:06 +08:00 · 2023-02-16 01:55:20 +08:00
parent b13627ad46
commit 230ed15024
12 changed files with 249 additions and 51 deletions
--- a/README.md
+++ b/README.md
@@ -5,21 +5,18 @@
 **模型：**

 - [x] ChatGPT
- - ...
 
 **应用：**

 - [ ] 终端
 - [ ] Web
 - [x] 个人微信
- - [x] 公众号
+ - [x] 公众号 (个人/企业)
 - [ ] 企业微信
 - [ ] Telegram
 - [ ] QQ
 - [ ] 钉钉 
- - ... 
-
-
+ - [ ] 飞书

 # 快速开始

@@ -39,24 +36,159 @@ cd bot-on-anything/

 ### 2.配置说明

-核心配置文件为 `config.json`，
+核心配置文件为 `config.json`，项目中提供了模板文件 `config-template.json` ，可以从模板复制生成最终生效的 `config.json` 文件：

+```bash
+cp config-template.json config.json
+```
+
+配置文件结构如下：
+
+```bash
+{
+  "model": {
+    "type" : "openai",              # 选用的算法模型
+    "openai": {
+      # openAI配置
+    }
+  },
+  "channel": {
+    "type": "wechat_mp",            # 需要接入的应用 
+    "wechat": {
+        # 个人微信配置
+    },
+    "wechat_mp": {
+        # 公众号配置
+    }
+  }
+}
+```
+配置文件在最外层分成 `model` 和 `channel` 两部分，model 部分为模型配置，其中的 `type` 指定了选用哪个模型；`channel` 部分包含了应用渠道的配置，`type` 字段指定了接入哪个应用，同时下方对应的配置块也会生效。
+
+在使用时只需要更改 `model` 和 `channel` 配置块下的 `type` 字段，即可在任意模型和应用间完成切换，连接不同的通路。下面将依次介绍各个 模型 及 应用 的配置和运行过程。


 ## 二、选择模型

 ### 1.ChatGPT

+#### 1.1 注册 OpenAI 账号

-## 三、选择应用
+前往 [OpenAI注册页面](https://beta.openai.com/signup) 创建账号，参考这篇 [教程](https://www.cnblogs.com/damugua/p/16969508.html) 可以通过虚拟手机号来接收验证码。创建完账号则前往 [API管理页面](https://beta.openai.com/account/api-keys) 创建一个 API Key 并保存下来，后面需要在项目中配置这个key。

-### 1.微信
+> 项目中使用的对话模型是 davinci，计费方式是约每 750 字 (包含请求和回复) 消耗 $0.02，图片生成是每张消耗 $0.016，账号创建有免费的 $18 额度，使用完可以更换邮箱重新注册。
+
+#### 1.2 配置项说明
+
+```bash
+{
+  "model": {
+    "type" : "openai",
+   
+    "openai": {
+      "api_key": "YOUR API KEY",
+      "conversation_max_tokens": 1000,
+      "character_desc": "你是ChatGPT, 一个由OpenAI训练的大型语言模型, 你旨在回答并解决人们的任何问题，并且可以使用多种语言与人交流。"
+    }
+}
+```
+ `api_key`：填入上面注册账号时创建的 `OpenAI API KEY`
+ `conversation_max_tokens`：表示能够记忆的上下文最大字数（一问一答为一组对话，如果累积的对话字数超出限制，就会优先移除最早的一组对话）
+ `character_desc` 配置中保存着你对机器人说的一段话，他会记住这段话并作为他的设定，你可以为他定制任何人格      


-### 2.公众号
+## 三、运行应用

+### 1.个人微信

+与项目 [chatgpt-on-wechat](https://github.com/zhayujie/chatgpt-on-wechat) 的使用方式相同，目前接入个人微信可能导致账号被限制，暂时不建议使用。

+配置项说明：

-## 四、运行
+```bash
+"channel": {
+    "type": "wechat",
+    
+    "single_chat_prefix": ["bot", "@bot"],
+    "single_chat_reply_prefix": "[bot] ",
+    "group_chat_prefix": ["@bot"],
+    "group_name_white_list": ["ChatGPT测试群"],
+    "image_create_prefix": ["画", "看", "找一张"],
+    
+    "wechat": {
+    }
+}
+```
+个人微信的配置项放在和 `type` 同级的层次，表示这些为公共配置，会复用于其他应用。配置加载时会优先使用模块内的配置，如果未找到便使用公共配置。

+在项目根目录下执行 `python3 app.py` 即可启动程序，用手机扫码后完成登录，使用详情参考 [chatgpt-on-wechat](https://github.com/zhayujie/chatgpt-on-wechat)。
+
+### 2.个人订阅号
+
+#### 2.1 依赖安装
+
+安装 [werobot](https://github.com/offu/WeRoBot) 依赖：
+
+```bash
+pip3 install werobot
+```
+
+#### 2.2 配置
+
+```bash
+"channel": {
+    "type": "wechat_mp",
+        
+    "wechat_mp": {
+      "token": "YOUR TOKEN",           # token值
+      "port": "8088"                   # 程序启动监听的端口
+    }
+}
+```
+
+#### 2.1 运行程序
+
+在项目目录下运行 `python3 app.py`，终端显示如下则表示已成功运行：
+
+```
+[INFO][2023-02-16 01:39:53][app.py:12] - [INIT] load config: ...
+[INFO][2023-02-16 01:39:53][wechat_mp_channel.py:25] - [WX_Public] Wechat Public account service start!
+Bottle v0.12.23 server starting up (using AutoServer())...
+Listening on http://127.0.0.1:8088/
+Hit Ctrl-C to quit.
+```
+
+#### 2.2 设置公众号回调地址
+
+在 [微信公众平台](https://mp.weixin.qq.com/) 中进入个人订阅号，启用服务器配置：
+
+![wx_mp_config.png](docs/images/wx_mp_config.png)
+
+- 服务器地址 (URL)：在浏览器访问该URL需要能访问到服务器上运行的python程序 (默认为8088端口)
+- 令牌 (Token)：需和配置中的token一致
+
+#### 2.3 使用
+
+用户关注订阅号后，发送消息即可。
+
+> 注：用户发送消息后，微信后台会向配置的URL地址推送，但如果5s内未回复就会断开连接，同时重试3次，但往往请求openai接口不止5s。本项目中通过异步和缓存将5s超时限制优化至15s，但超出该时间仍无法正常回复。 同时每次5s连接断开时web框架会报错，待后续优化。
+ 
+
+### 3.企业服务号
+
+在企业服务号中，通过先异步访问openai接口，再通过客服接口主动推送用户的方式，解决了个人订阅号的15s超时问题。
+
+企业服务号配置只需修改type为`wechat_mp_service`，配置块仍复用 `wechat_mp`，在基础上增加了 `app_id` 和 `app_secret` 两个配置项。
+
+```bash
+"channel": {
+    "type": "wechat_mp_service",
+        
+    "wechat_mp": {
+      "token": "YOUR TOKEN",            # token值
+      "port": "8088",                   # 程序启动监听的端口
+      "app_id": "YOUR APP ID",          # appID
+      "app_secret": "YOUR APP SECRET"   # app secret
+    }
+}
+```
--- a/app.py
+++ b/app.py
@@ -12,7 +12,7 @@ if __name__ == '__main__':
        logger.info("[INIT] load config: {}".format(config.conf()))

        # create channel
-        channel = channel_factory.create_channel(config.conf().get("channel"))
+        channel = channel_factory.create_channel(config.conf().get("channel").get("type"))

        # startup channel
        channel.startup()
--- a/bridge/bridge.py
+++ b/bridge/bridge.py
@@ -6,4 +6,4 @@ class Bridge(object):
        pass

    def fetch_reply_content(self, query, context):
-        return model_factory.create_bot(config.conf().get("model")).reply(query, context)
+        return model_factory.create_bot(config.conf().get("model").get("type")).reply(query, context)
--- a/channel/channel_factory.py
+++ b/channel/channel_factory.py
@@ -14,8 +14,12 @@ def create_channel(channel_type):
        return WechatChannel()

    elif channel_type == const.WECHAT_MP:
-        from channel.wechat.wechat_mp_channel import WechatPublicAccount
-        return WechatPublicAccount()
+        from channel.wechat.wechat_mp_channel import WechatSubsribeAccount
+        return WechatSubsribeAccount()
+
+    elif channel_type == const.WECHAT_MP_SERVICE:
+        from channel.wechat.wechat_mp_service_channel import WechatServiceAccount
+        return WechatServiceAccount()

    else:
        raise RuntimeError
--- a/channel/wechat/wechat_channel.py
+++ b/channel/wechat/wechat_channel.py
@@ -9,7 +9,8 @@ from itchat.content import *
 from channel.channel import Channel
 from concurrent.futures import ThreadPoolExecutor
 from common.log import logger
-from config import conf
+from common import const
+from config import channel_conf_val, channel_conf
 import requests
 import io

@@ -45,7 +46,7 @@ class WechatChannel(Channel):
        to_user_id = msg['ToUserName']              # 接收人id
        other_user_id = msg['User']['UserName']     # 对手方id
        content = msg['Text']
-        match_prefix = self.check_prefix(content, conf().get('single_chat_prefix'))
+        match_prefix = self.check_prefix(content, channel_conf_val(const.WECHAT, 'single_chat_prefix'))
        if from_user_id == other_user_id and match_prefix is not None:
            # 好友向自己发送消息
            if match_prefix != '':
@@ -53,7 +54,7 @@ class WechatChannel(Channel):
                if len(str_list) == 2:
                    content = str_list[1].strip()

-            img_match_prefix = self.check_prefix(content, conf().get('image_create_prefix'))
+            img_match_prefix = self.check_prefix(content, channel_conf_val(const.WECHAT, 'image_create_prefix'))
            if img_match_prefix:
                content = content.split(img_match_prefix, 1)[1].strip()
                thread_pool.submit(self._do_send_img, content, from_user_id)
@@ -65,7 +66,7 @@ class WechatChannel(Channel):
            str_list = content.split(match_prefix, 1)
            if len(str_list) == 2:
                content = str_list[1].strip()
-            img_match_prefix = self.check_prefix(content, conf().get('image_create_prefix'))
+            img_match_prefix = self.check_prefix(content, channel_conf_val(const.WECHAT, 'image_create_prefix'))
            if img_match_prefix:
                content = content.split(img_match_prefix, 1)[1].strip()
                thread_pool.submit(self._do_send_img, content, to_user_id)
@@ -88,11 +89,11 @@ class WechatChannel(Channel):
        elif len(content_list) == 2:
            content = content_list[1]

-        config = conf()
-        match_prefix = (msg['IsAt'] and not config.get("group_at_off", False)) or self.check_prefix(origin_content, config.get('group_chat_prefix')) \
-                       or self.check_contain(origin_content, config.get('group_chat_keyword'))
-        if ('ALL_GROUP' in config.get('group_name_white_list') or group_name in config.get('group_name_white_list') or self.check_contain(group_name, config.get('group_name_keyword_white_list'))) and match_prefix:
-            img_match_prefix = self.check_prefix(content, conf().get('image_create_prefix'))
+        match_prefix = (msg['IsAt'] and not channel_conf_val(const.WECHAT, "group_at_off", False)) or self.check_prefix(origin_content, channel_conf_val(const.WECHAT, 'group_chat_prefix')) \
+                       or self.check_contain(origin_content, channel_conf_val(const.WECHAT, 'group_chat_keyword'))
+        group_white_list = channel_conf_val(const.WECHAT, 'group_name_white_list')
+        if ('ALL_GROUP' in group_white_list or group_name in group_white_list or self.check_contain(group_name, channel_conf_val(const.WECHAT, 'group_name_keyword_white_list'))) and match_prefix:
+            img_match_prefix = self.check_prefix(content, channel_conf_val(const.WECHAT, 'image_create_prefix'))
            if img_match_prefix:
                content = content.split(img_match_prefix, 1)[1].strip()
                thread_pool.submit(self._do_send_img, content, group_id)
@@ -111,7 +112,7 @@ class WechatChannel(Channel):
            context['from_user_id'] = reply_user_id
            reply_text = super().build_reply_content(query, context)
            if reply_text:
-                self.send(conf().get("single_chat_reply_prefix") + reply_text, reply_user_id)
+                self.send(channel_conf_val(const.WECHAT, "single_chat_reply_prefix") + reply_text, reply_user_id)
        except Exception as e:
            logger.exception(e)

@@ -146,7 +147,7 @@ class WechatChannel(Channel):
        reply_text = super().build_reply_content(query, context)
        if reply_text:
            reply_text = '@' + msg['ActualNickName'] + ' ' + reply_text.strip()
-            self.send(conf().get("group_chat_reply_prefix", "") + reply_text, msg['User']['UserName'])
+            self.send(channel_conf_val(const.WECHAT, "group_chat_reply_prefix", "") + reply_text, msg['User']['UserName'])


    def check_prefix(self, content, prefix_list):
--- a/channel/wechat/wechat_mp_channel.py
+++ b/channel/wechat/wechat_mp_channel.py
@@ -1,12 +1,12 @@
 import werobot
 import time
-import config
+from config import channel_conf
 from common import const
 from common.log import logger
 from channel.channel import Channel
 from concurrent.futures import ThreadPoolExecutor

-robot = werobot.WeRoBot(token=config.fetch(const.WECHAT_MP).get('token'))
+robot = werobot.WeRoBot(token=channel_conf(const.WECHAT_MP).get('token'))
 thread_pool = ThreadPoolExecutor(max_workers=8)
 cache = {}

@@ -15,14 +15,15 @@ def hello_world(msg):
    logger.info('[WX_Public] receive public msg: {}, userId: {}'.format(msg.content, msg.source))
    key = msg.content + '|' + msg.source
    if cache.get(key):
+        # request time
        cache.get(key)['req_times'] += 1
-    return WechatPublicAccount().handle(msg)
+    return WechatSubsribeAccount().handle(msg)


-class WechatPublicAccount(Channel):
+class WechatSubsribeAccount(Channel):
    def startup(self):
        logger.info('[WX_Public] Wechat Public account service start!')
-        robot.config['PORT'] = config.fetch(const.WECHAT_MP).get('port')
+        robot.config['PORT'] = channel_conf(const.WECHAT_MP).get('port')
        robot.run()

    def handle(self, msg, count=0):
--- a/channel/wechat/wechat_mp_service_channel.py
+++ b/channel/wechat/wechat_mp_service_channel.py
@@ -0,0 +1,36 @@
+import werobot
+from config import channel_conf
+from common import const
+from common.log import logger
+from channel.channel import Channel
+from concurrent.futures import ThreadPoolExecutor
+
+robot = werobot.WeRoBot(token=channel_conf(const.WECHAT_MP).get('token'))
+thread_pool = ThreadPoolExecutor(max_workers=8)
+
+@robot.text
+def hello_world(msg):
+    logger.info('[WX_Public] receive public msg: {}, userId: {}'.format(msg.content, msg.source))
+    return WechatServiceAccount().handle(msg)
+
+
+class WechatServiceAccount(Channel):
+    def startup(self):
+        logger.info('[WX_Public] Wechat Public account service start!')
+        robot.config['PORT'] = channel_conf(const.WECHAT_MP).get('port')
+        robot.config["APP_ID"] = "YOUR APP ID"
+        robot.config["APP_SECRET"] = "YOUR APP SECRET"
+        robot.run()
+
+    def handle(self, msg, count=0):
+        context = {}
+        context['from_user_id'] = msg.source
+        thread_pool.submit(self._do_send, msg.content, context)
+        return "正在思考中..."
+
+
+    def _do_send(self, query, context):
+        reply_text = super().build_reply_content(query, context)
+        logger.info('[WX_Public] reply content: {}'.format(reply_text))
+        client = robot.client
+        client.send_text_message(context['from_user_id'], reply_text)
--- a/common/const.py
+++ b/common/const.py
@@ -1,6 +1,7 @@
 # channel
 WECHAT = "wechat"
 WECHAT_MP = "wechat_mp"
+WECHAT_MP_SERVICE = "wechat_mp_service"

 # model
 OPEN_AI = "openai"
--- a/config-template.json
+++ b/config-template.json
@@ -1,23 +1,26 @@
 {
-  "channel": "wechat",
-  "bot": "openai",
-
-  "openai": {
-    "api_key": "YOUR API KEY",
-    "conversation_max_tokens": 1000,
-    "character_desc": "你是ChatGPT, 一个由OpenAI训练的大型语言模型, 你旨在回答并解决人们的任何问题，并且可以使用多种语言与人交流。"
+  "model": {
+    "type" : "openai",
+    "openai": {
+      "api_key": "YOUR API KEY",
+      "conversation_max_tokens": 1000,
+      "character_desc": "你是ChatGPT, 一个由OpenAI训练的大型语言模型, 你旨在回答并解决人们的任何问题，并且可以使用多种语言与人交流。"
+    }
  },
-
-  "wechat": {
+  "channel": {
+    "type": "wechat_mp",
    "single_chat_prefix": ["bot", "@bot"],
    "single_chat_reply_prefix": "[bot] ",
    "group_chat_prefix": ["@bot"],
-    "group_name_white_list": ["ALL_GROUP"],
-    "image_create_prefix": ["画", "看", "找一张"]
-  },
+    "group_name_white_list": ["ChatGPT测试群"],
+    "image_create_prefix": ["画", "看", "找一张"],

-  "wechat_mp": {
-    "token": "YOUR TOKEN",
-    "port": "8088"
+    "wechat": {
+    },
+
+    "wechat_mp": {
+      "token": "YOUR TOKEN",
+      "port": "8088"
+    }
  }
 }
--- a/config.py
+++ b/config.py
@@ -28,5 +28,25 @@ def read_file(path):
 def conf():
    return config

-def fetch(model):
-    return config.get(model)
+
+def model_conf(model_type):
+    return config.get('model').get(model_type)
+
+def model_conf_val(model_type, key):
+    val = config.get('model').get(model_type).get(key)
+    if not val:
+        # common default config
+        return config.get('model').get(key)
+    return val
+
+
+def channel_conf(channel_type):
+    return config.get('channel').get(channel_type)
+
+
+def channel_conf_val(channel_type, key, default=None):
+    val = config.get('channel').get(channel_type).get(key)
+    if not val:
+        # common default config
+        return config.get('channel').get(key, default)
+    return val
--- a/docs/images/wx_mp_config.png
+++ b/docs/images/wx_mp_config.png
--- a/model/openai/open_ai_model.py
+++ b/model/openai/open_ai_model.py
@@ -1,7 +1,7 @@
 # encoding:utf-8

 from model.model import Model
-from config import fetch
+from config import model_conf
 from common import const
 from common.log import logger
 import openai
@@ -12,7 +12,7 @@ user_session = dict()
 # OpenAI对话模型API (可用)
 class OpenAIModel(Model):
    def __init__(self):
-        openai.api_key = fetch(const.OPEN_AI).get('api_key')
+        openai.api_key = model_conf(const.OPEN_AI).get('api_key')


    def reply(self, query, context=None):
@@ -103,7 +103,7 @@ class Session(object):
        :param user_id: from user id
        :return: query content with conversaction
        '''
-        prompt = fetch(const.OPEN_AI).get("character_desc", "")
+        prompt = model_conf(const.OPEN_AI).get("character_desc", "")
        if prompt:
            prompt += "<|endoftext|>\n\n\n"
        session = user_session.get(user_id, None)
@@ -117,7 +117,7 @@ class Session(object):

    @staticmethod
    def save_session(query, answer, user_id):
-        max_tokens = fetch(const.OPEN_AI).get("conversation_max_tokens")
+        max_tokens = model_conf(const.OPEN_AI).get("conversation_max_tokens")
        if not max_tokens:
            # default 3000
            max_tokens = 1000