https://github.com/lich0821/shadiaorobot
WeChatFerry 群的沙雕机器人
https://github.com/lich0821/shadiaorobot
rasa robot wechat
Last synced: 8 months ago
JSON representation
WeChatFerry 群的沙雕机器人
- Host: GitHub
- URL: https://github.com/lich0821/shadiaorobot
- Owner: lich0821
- License: mit
- Created: 2023-11-05T12:56:38.000Z (over 2 years ago)
- Default Branch: master
- Last Pushed: 2023-11-19T14:10:01.000Z (over 2 years ago)
- Last Synced: 2025-05-05T14:52:29.232Z (about 1 year ago)
- Topics: rasa, robot, wechat
- Language: Python
- Homepage:
- Size: 9.55 MB
- Stars: 5
- Watchers: 2
- Forks: 5
- Open Issues: 0
-
Metadata Files:
- Readme: README.MD
- License: LICENSE
Awesome Lists containing this project
README
# 群主的沙雕机器人
基于 [Rasa](https://github.com/rasaHQ/rasa) 的一个机器人示例。
## 环境搭建
### 安装 Python
官方说支持 `3.7`，`3.8`，`3.9` 和 `3.10`，但 `3.7` 太旧了；`3.10` 又要求蛮多；建议 `3.8` 或者 `3.9`。
### 创建虚拟环境
```sh
python -m venv .env
# 激活虚拟环境
source .env/bin/activate
```
### 安装 RASA 开源版
```sh
# 更新 pip
pip install -U pip
# 安装 RASA
pip install rasa
```
## 快速开始
### 克隆项目
```sh
git clone
```
### 安装依赖
```sh
pip install -r requirements.txt
```
### 训练模型
```sh
rasa train
# 训练时遇到网络报错的话，需要科学一下
```
### 运行 Action 服务
#### 闲聊
闲聊模型需要先训练。
```sh
# 切换到 actions 目录
cd actions/
python chitchat_helper.py -m train
```
如果没有 GPU，训练可能比较费劲，可以从 [release](https://github.com/lich0821/ShaDiaoRobot/releases/latest) 处下载训练好的模型（虽然也训练得不怎么样），解压到 `actions/chitchat/model_data`：
```
actions/chitchat/model_data
├── checkpoint
├── ckpt-10.data-00000-of-00001
└── ckpt-10.index
```
Tensorflow 的兼容性不行，注意版本：
> tensorflow 2.12.0
> tensorflow-estimator 2.12.0
> tensorflow-hub 0.13.0
> tensorflow-io-gcs-filesystem 0.32.0
> tensorflow-text 2.12.0
#### 天气
去[知心天气](https://www.seniverse.com/)注册一个账号，可以免费使用天气接口。
```sh
# 设置 KEY
export WEATHER_KEY=your_key
rasa run actions
```
### 启动 RASA API
```sh
rasa run --enable-api
```
### 验证
* cURL
```sh
curl --location 'http://localhost:5005/webhooks/rest/webhook' \
--data '{
"sender": "user1",
"message": "明天北京天气怎么样？"
}'
```
* Python
```py
import requests
url = "http://localhost:5005/webhooks/rest/webhook"
payload = "{\n \"sender\": \"user1\",\n \"message\": \"明天北京天气怎么样？\"\n}"
headers = {}
response = requests.request("POST", url, headers=headers, data=payload)
print(response.text)
```
## 从零创建项目
忽略一个 `SQLAlchemy` 告警：
```sh
export SQLALCHEMY_SILENCE_UBER_WARNING=1
```
### 初始化项目
```sh
rasa init
```
```txt
Welcome to Rasa! 🤖
To get started quickly, an initial project will be created.
If you need some help, check out the documentation at https://rasa.com/docs/rasa.
Now let's start! 👇🏽
? Please enter a path where the project will be created [default: current directory]
? Directory 'RasaDemo' is not empty. Continue? Yes
……
Created project directory at 'RasaDemo'.
Finished creating project structure.
? ? Do you want to train an initial model? 💪🏽 No
No problem 👍🏼. You can also train a model later by going to the project directory and running 'rasa train'.
```
这样就有一个示例项目；也可以选择训练跑跑。然后基示例项目改。
Once you've done the following, you can train your bot and try it out!
* Add RulePolicy to your policies and ResponseSelector to your pipeline in config.yml
* Add at least one rule for responding to FAQs/chitchat
* Add examples for your FAQs/chitchat intents
* Add responses for your FAQs/chitchat intents
* Update the intents in your domain
## 调试
```sh
rasa shell --debug
```