参考https://github.com/chatopera/clause-py-demo

前提：安装完毕 Docker、Docker Compose 和 Python3.x

git clone https://github.com/chatopera/clause-py-demo.git
cd clause-py-demo
docker-compose pull

我使用了sudo

创新python环境，并使用

(env) xuehp@haomeiya002:~/git/clause-py-demo$
sudo pip3 install thrift
sudo pip3 install absl-py
sudo pip3 install clause
sudo pip3 install -r requirements.txt
export DOCKER_HOST=8056
sudo docker-compose up

等待，这样服务就运行起来了。

注意：

1. bot.py 是一个示例程序

2. 加载文件profile.json，这个文件定义了一个机器人，例如：点餐机器人

3. 处理同义词，识别意图。这些在文件profile.json中已经定义。文件见后文

4. 和服务端交互，继续完成以下操作：训练机器人、开始对话。如下图

5. 运行bot.py失败，ConnectionResetError: [Errno 104] Connection reset by peer

   profile.json文件，很重要。定义了一个机器人，意图、交互方式（说法）、槽位、槽值的抽取方式
   {
   "chatbotID": "bot007",
   "dicts": [
   {
   "name": "food",
   "type": "vocab",
   "dictwords": [
   {
   "word": "汉堡",
   "synonyms": "汉堡包;漢堡;漢堡包"
   },
   {
   "word": "蕃茄酱"
   }
   ]
   },
   {
   "name": "phoneNumber",
   "type": "regex",
   "patterns": ["1[0-9]{9}"]
   }
   ],
   "intents": [
   {
   "name": "take_out",
   "description": "下外卖订单",
   "slots": [
   {
   "name": "time",
   "dictname": "@TIME",
   "requires": true,
   "question": "您想什么时候送到？"
   },
   {
   "name": "loc",
   "dictname": "@LOC",
   "requires": true,
   "question": "您希望该订单送到哪里？"
   },
   {
   "name": "food",
   "dictname": "food",
   "requires": true,
   "question": "您需要什么食物?"
   },
   {
   "name": "phone",
   "dictname": "phoneNumber",
   "requires": true,
   "question": "您的手机号是什么？"
   }
   ],
   "utters": [
   {
   "utterance": "我想订一份{food}"
   },
   {
   "utterance": "我想点外卖"
   },
   {
   "utterance": "我想点一份外卖，{time}用餐"
   },
   {
   "utterance": "我想点一份{food}，送到{loc}"
   }
   ]
   }
   ]
   }