rasa中文对话系统构建

【关于 rasa中文对话系统构建】那些你不知道的事

作者：杨夕

项目地址：https://github.com/km1994/nlp_paper_study

个人介绍：大佬们好，我叫杨夕，该项目主要是本人在研读顶会论文和复现经典论文过程中，所见、所思、所想、所闻，可能存在一些理解错误，希望大佬们多多指正。

学习项目：https://github.com/zqhZY/_rasa_chatbot

一、安装 Rasa 内容

温馨提示：由于安装 Rasa 过程中，会安装各种乱七八糟的依赖库（eg：tensorflow 2.0，...），导致安装失败，所以建议用 conda ，新建一个 conda 环境，然后在该环境上面开发。

1.1 安装中文版本的 Rasa-nlu

  $ git clone https://github.com/crownpku/Rasa_NLU_Chi.git
  $ cd rasa_nlu
  $ pip install -r requirements.txt
  $ python setup.py install

1.2 安装 rasa

  pip --default-timeout=500 install -U rasa
  pip install --upgrade --ignore-installed tensorflow

安装过程中遇到的问题

问题一：

  ERROR: After October 2020 you may experience errors when installing or updating packages. This is because pip will change the way that it resolves dependency conflicts.

  We recommend you use --use-feature=2020-resolver to test your packages with the new resolver before it becomes the default.

然後使用者都要自己去解相依性問題 … pip 從 2020/10 月後會改用 2020-resolver 這個 resolver，pip 會自己去解決相依性問題，在官方還沒預設採用 2020-resolver 之前要用 --use-feature=2020-resolver 手動指定 resolver

假設升級 aws-cdk.aws-sns-subscriptions 則有一連串也需要被更新的套件也會被拉上來

解决方法一

  python -m pip install --upgrade pip
  pip install --use-feature=2020-resolver --upgrade aws-cdk.aws-sns-subscriptions

然后重新安装

解决方法二

解决方法: 在pip命令中加入–use-feature=2020-resolver参数就可以了, 比如pip install xxx --use-feature=2020-resolver

1.3 sklearn 和 MITIE 库安装

  pip install -U scikit-learn sklearn-crfsuite
  pip install git+https://github.com/mit-nlp/MITIE.git

下载中文词向量 total_word_feature_extractor_zh.dat https://mega.nz/#!EWgTHSxR!NbTXDAuVHwwdP2-Ia8qG7No-JUsSbH5mNQSRDsjztSA

注：MITIE 库比较大，所以这种安装方式容易出现问题，所以我用另一种安装方式

  $ git clone https://github.com/mit-nlp/MITIE.git
  $ cd MITIE/
  $ python setup.py install

安装结果

  Compiling src/text_feature_extraction.cpp
  Compiling ../dlib/dlib/threads/multithreaded_object_extension.cpp
  Compiling ../dlib/dlib/threads/threaded_object_extension.cpp
  Compiling ../dlib/dlib/threads/threads_kernel_1.cpp
  Compiling ../dlib/dlib/threads/threads_kernel_2.cpp
  Compiling ../dlib/dlib/threads/threads_kernel_shared.cpp
  Compiling ../dlib/dlib/threads/thread_pool_extension.cpp
  Compiling ../dlib/dlib/misc_api/misc_api_kernel_1.cpp
  Compiling ../dlib/dlib/misc_api/misc_api_kernel_2.cpp
  Linking libmitie.so
  Making libmitie.a
  Build Complete
  make[1]: Leaving directory `/web/workspace/yangkm/python_wp/nlu/DSWp/MITIE/mitielib'
  running build_py
  creating build
  creating build/lib
  creating build/lib/mitie
  copying mitielib/__init__.py -> build/lib/mitie
  copying mitielib/mitie.py -> build/lib/mitie
  copying mitielib/libmitie.so -> build/lib/mitie
  running install_lib
  copying build/lib/mitie/__init__.py -> /home/amy/.conda/envs/yangkm/lib/python3.6/site-packages/mitie
  copying build/lib/mitie/mitie.py -> /home/amy/.conda/envs/yangkm/lib/python3.6/site-packages/mitie
  copying build/lib/mitie/libmitie.so -> /home/amy/.conda/envs/yangkm/lib/python3.6/site-packages/mitie
  byte-compiling /home/amy/.conda/envs/yangkm/lib/python3.6/site-packages/mitie/__init__.py to __init__.cpython-36.pyc
  byte-compiling /home/amy/.conda/envs/yangkm/lib/python3.6/site-packages/mitie/mitie.py to mitie.cpython-36.pyc
  running install_egg_info
  Writing /home/amy/.conda/envs/yangkm/lib/python3.6/site-packages/mitie-0.7.0-py3.6.egg-info

注：会存在一些 warning 警告，对结果影响不大

1.4 安装 rasa_core

pip install rasa_core==0.9.0

注：

  # NameVersion   Build  Channel
  absl-py   0.9.0 <pip>
  aiofiles  0.5.0 <pip>
  aiohttp   3.6.2 <pip>
  APScheduler   3.6.3 <pip>
  astor 0.8.1 <pip>
  astunparse1.6.3 <pip>
  async-generator   1.10  <pip>
  async-timeout 3.0.1 <pip>
  attrs 19.3.0<pip>
  Automat   20.2.0<pip>
  bleach1.5.0 <pip>
  boto3 1.15.1<pip>
  botocore  1.18.1<pip>
  bz2file   0.98  <pip>
  cachetools4.1.1 <pip>
  certifi   2020.6.20 <pip>
  certifi   2016.2.28py36_0https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
  cffi  1.14.3<pip>
  chardet   3.0.4 <pip>
  click 7.1.2 <pip>
  cloudpickle   1.3.0 <pip>
  colorclass2.2.0 <pip>
  coloredlogs   10.0  <pip>
  colorhash 1.0.2 <pip>
  ConfigArgParse1.2.3 <pip>
  constantly15.1.0<pip>
  contextvars   2.4   <pip>
  cryptography  3.1   <pip>
  cycler0.10.0<pip>
  dataclasses   0.7   <pip>
  decorator 4.4.2 <pip>
  dill  0.3.2 <pip>
  dnspython 1.16.0<pip>
  docopt0.6.2 <pip>
  docutils  0.16  <pip>
  dopamine-rl   3.1.7 <pip>
  fakeredis 0.10.3<pip>
  fbmessenger   6.0.0 <pip>
  Flask 1.1.2 <pip>
  Flask-Cors3.0.9 <pip>
  flax  0.2.1 <pip>
  future0.16.0<pip>
  gast  0.2.2 <pip>
  gevent1.5.0 <pip>
  gin-config0.3.0 <pip>
  google-api-core   1.22.2<pip>
  google-api-python-client  1.12.1<pip>
  google-auth   1.21.2<pip>
  google-auth-httplib2  0.0.4 <pip>
  google-auth-oauthlib  0.4.1 <pip>
  google-pasta  0.2.0 <pip>
  googleapis-common-protos  1.52.0<pip>
  GPUtil1.3.0 <pip>
  graphviz  0.8.4 <pip>
  greenlet  0.4.16<pip>
  grpcio1.32.0<pip>
  gunicorn  20.0.4<pip>
  gym   0.17.2<pip>
  h11   0.8.1 <pip>
  h23.2.0 <pip>
  h5py  2.10.0<pip>
  hpack 3.0.0 <pip>
  hstspreload   2020.9.15 <pip>
  html5lib  0.9999999 <pip>
  httpcore  0.3.0 <pip>
  httplib2  0.18.1<pip>
  httptools 0.1.1 <pip>
  httpx 0.9.3 <pip>
  humanfriendly 8.2   <pip>
  hyperframe5.2.0 <pip>
  hyperlink 17.3.1<pip>
  idna  2.10  <pip>
  idna  2.7   <pip>
  idna-ssl  1.1.0 <pip>
  immutables0.14  <pip>
  importlib-metadata1.7.0 <pip>
  incremental   17.5.0<pip>
  itsdangerous  1.1.0 <pip>
  jax   0.1.77<pip>
  jaxlib0.1.55<pip>
  jieba 0.42.1<pip>
  Jinja22.11.2<pip>
  jmespath  0.10.0<pip>
  joblib0.16.0<pip>
  jsonpickle1.4.1 <pip>
  jsonschema3.2.0 <pip>
  kafka-python  1.4.7 <pip>
  Keras 2.1.5 <pip>
  Keras-Applications1.0.8 <pip>
  Keras-Preprocessing   1.1.0 <pip>
  kfac  0.2.2 <pip>
  kiwisolver1.2.0 <pip>
  klein 17.10.0   <pip>
  Markdown  3.2.2 <pip>
  MarkupSafe1.1.1 <pip>
  matplotlib3.2.2 <pip>
  mattermostwrapper 2.2   <pip>
  mesh-tensorflow   0.1.16<pip>
  mitie 0.7.0 <pip>
  mpmath1.1.0 <pip>
  msgpack   1.0.0 <pip>
  msgpack-python0.5.4 <pip>
  multidict 4.6.1 <pip>
  networkx  2.4   <pip>
  numpy 1.18.5<pip>
  numpy 1.19.2<pip>
  oauth2client  4.1.3 <pip>
  oauthlib  3.1.0 <pip>
  opencv-python 4.4.0.42  <pip>
  openssl   1.0.2l0https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
  opt-einsum3.3.0 <pip>
  packaging 20.4  <pip>
  pandas1.1.2 <pip>
  pi0.1.2 <pip>
  pika  1.1.0 <pip>
  Pillow7.2.0 <pip>
  pip   20.2.3<pip>
  pip   9.0.1py36_1https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
  promise   2.3   <pip>
  prompt-toolkit2.0.10<pip>
  protobuf  3.13.0<pip>
  psycopg2-binary   2.8.6 <pip>
  pyasn10.4.8 <pip>
  pyasn1-modules0.2.8 <pip>
  pycparser 2.20  <pip>
  pydot 1.4.1 <pip>
  pygame1.9.6 <pip>
  pyglet1.5.0 <pip>
  PyHamcrest2.0.2 <pip>
  PyJWT 1.7.1 <pip>
  pykwalify 1.7.0 <pip>
  pymongo   3.8.0 <pip>
  pyparsing 2.4.7 <pip>
  pypng 0.0.20<pip>
  pyrsistent0.17.3<pip>
  PySocks   1.7.1 <pip>
  python3.6.2 0https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
  python-crfsuite   0.9.7 <pip>
  python-dateutil   2.8.1 <pip>
  python-engineio   3.12.1<pip>
  python-socketio   4.5.1 <pip>
  python-telegram-bot   11.1.0<pip>
  pytz  2019.3<pip>
  PyYAML5.1   <pip>
  pyzmq 17.1.0<pip>
  questionary   1.5.2 <pip>
  rasa  1.10.12   <pip>
  rasa-core 0.9.0 <pip>
  rasa-sdk  1.10.2<pip>
  readline  6.2   2https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
  redis 3.5.3 <pip>
  regex 2020.6.8  <pip>
  requests  2.24.0<pip>
  requests-async0.5.0 <pip>
  requests-oauthlib 1.3.0 <pip>
  requests-toolbelt 0.9.1 <pip>
  rfc3986   1.4.0 <pip>
  rocketchat-API0.6.36<pip>
  rsa   4.6   <pip>
  ruamel.yaml   0.16.12   <pip>
  ruamel.yaml.clib  0.2.2 <pip>
  s3transfer0.3.3 <pip>
  sanic 19.12.2   <pip>
  Sanic-Cors0.10.0.post3  <pip>
  sanic-jwt 1.4.1 <pip>
  Sanic-Plugins-Framework   0.9.4 <pip>
  scikit-learn  0.22.2.post1  <pip>
  scipy 1.4.1 <pip>
  setuptools36.4.0   py36_1https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
  setuptools50.3.0<pip>
  simplejson3.13.2<pip>
  six   1.15.0<pip>
  sklearn-crfsuite  0.3.6 <pip>
  slackclient   2.9.0 <pip>
  sniffio   1.1.0 <pip>
  SQLAlchemy1.3.19<pip>
  sqlite3.13.00https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
  sympy 1.6.2 <pip>
  tabulate  0.8.7 <pip>
  tensor2tensor 1.14.1<pip>
  tensorboard   1.8.0 <pip>
  tensorboard   2.1.1 <pip>
  tensorboard-plugin-wit1.7.0 <pip>
  tensorflow2.1.1 <pip>
  tensorflow1.8.0 <pip>
  tensorflow-addons 0.7.1 <pip>
  tensorflow-cpu1.15.0<pip>
  tensorflow-datasets   3.2.1 <pip>
  tensorflow-estimator  2.1.0 <pip>
  tensorflow-gan2.0.0 <pip>
  tensorflow-hub0.8.0 <pip>
  tensorflow-metadata   0.24.0<pip>
  tensorflow-probability0.7.0 <pip>
  termcolor 1.1.0 <pip>
  terminaltables3.1.0 <pip>
  tf-slim   1.1.0 <pip>
  threadpoolctl 2.1.0 <pip>
  tk8.5.180https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
  tqdm  4.45.0<pip>
  twilio6.26.3<pip>
  Twisted   20.3.0<pip>
  typing3.6.2 <pip>
  typing-extensions 3.7.4.3   <pip>
  tzlocal   2.1   <pip>
  ujson 2.0.3 <pip>
  uritemplate   3.0.1 <pip>
  urllib3   1.25.10   <pip>
  urllib3   1.24.3<pip>
  uvloop0.14.0<pip>
  wcwidth   0.2.5 <pip>
  webexteamssdk 1.3   <pip>
  websocket-client  0.54.0<pip>
  websockets8.1   <pip>
  Werkzeug  1.0.1 <pip>
  wheel 0.29.0   py36_0https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
  wheel 0.35.1<pip>
  wrapt 1.12.1<pip>
  xz5.2.3 0https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
  yarl  1.5.1 <pip>
  zipp  3.1.0 <pip>
  zlib  1.2.110https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
  zope.interface5.1.0 <pip>

二、新建项目

2.1 安装rasa，自动生成rasa项目所需文件

  # 安装rasa, 由于网络问题，延长超时时间
  # 如果仍然超时异常，多执行几次
  $ mkdir rasa_zh
  $ cd rasa_zh
  $ rasa init --no-prompt # 生成文件

如果出现 No matching distribution found for tensorflow>=2.1.0异常，可以通过执行pip3 install --upgrade tensorflow rasa命令解决。

2.2 测试助手

  sudo rasa shell

三、Rasa使用–构建简单聊天机器人

3.1 Rasa工作原理

首先，将用户输入的Message传递到Interpreter(Rasa NLU模块)，该模块负责识别Message中的"意图(intent)“和提取所有"实体”(entity)数据；
其次，Rasa Core会将Interpreter提取到的意图和识别传给Tracker对象，该对象的主要作用是跟踪会话状态(conversation state)；
第三，利用policy记录Tracker对象的当前状态，并选择执行相应的action，其中，这个action是被记录在Track对象中的；
最后，将执行action返回的结果输出即完成一次人机交互。

3.2 构建NLU样本

3.2.1 data/nlu.md

NLU模型训练样本数据：

## intent:greet
- 你好
- 你好啊
- 早上好
- 晚上好
- hello
- hi
- 嗨
- 嗨喽
- 见到你很高兴
- 嘿
- 早
- 上午好
- hello哈喽
- 哈喽哈喽
- hello hello
- 喂喂

## intent:goodbye
- goodbye
- bye
- bye bye
- 88
- 886
- 再见
- 拜
- 拜拜
- 拜拜，下次再聊
- 下次见
- 回头见
- 下次再见
- 下次再聊
- 有空再聊
- 先这样吧
- 好了，就说这么多了
- 好了，先这样
- 没事

## intent:whoareyou
- 你是谁
- 我知道你吗
- 谁
- 我认识你吗
- 这是谁啊
- 是谁
- 请问你是谁
- 请问我认识你吗
- 你是哪位
- 你是？
- 是谁？
- 可以告诉我你的名字吗
- 你叫什么名字

## intent:whattodo
- 你支持什么功能
- 你有什么功能
- 你能干什么
- 你能做什么

## intent:thanks
- 谢谢
- thanks
- thank you
- 真的太感谢你了，帮了我大忙
- 谢谢你帮了我大忙
- 你帮了我大忙，谢谢你小智
- 非常感谢
- 谢了

## intent:deny
- 不
- no
- 不可以
- 不是的
- 不认同
- 否定
- 不是这样子的
- 我不同意你的观点
- 不同意
- 不好
- 你长得很美，就不要想得太美。
- 拒绝
- 不行

## intent:affirm
- 是的
- 当然
- 好的
- ok
- 嗯
- 可以
- 你可以这么做
- 你做得可以啊
- 同意
- 听起来不错
- 是这样的
- 的确是这样子的
- 我同意你的观点
- 对的
- 好滴
- 行
- 还行
- 当然可以

## intent: request_weather
- 天气
- 查询天气
- 帮我查天气信息
- 我想知道[明天](date-time)的天气
- [星期一](date-time)的天气
- [今天](date-time)的天气怎么样
- 帮我查下[后天](date-time)的天气
- 查下[广州](address)的天气怎么样
-  [长沙](address)的天气
- [深圳](address)[明天](date-time)的天气
- 查下[今天](date-time)[上海](address)的天气
- 帮我查查[佛山](address)这[周六](date-time)的天气

3.3 构建Core样本

3.3.1 data/stories.md

## greet
* greet
- utter_answer_greet

## say affirm  with greet
* greet
- utter_answer_greet
* affirm
- utter_answer_affirm

## say affirm 
* affirm
- utter_answer_affirm

## say no with greet
* greet
- utter_answer_greet
* deny
- utter_answer_deny

## say no 
* deny
- utter_answer_deny


## say goodbye
* goodbye
- utter_answer_goodbye

## thanks with greet
* greet
- utter_answer_greet
* thanks
- utter_answer_thanks

## thanks
* thanks
- utter_answer_thanks

## who are you with greet
* greet
- utter_answer_greet
* whoareyou
- utter_answer_whoareyou

## who are you
* whoareyou
- utter_answer_whoareyou

## who are you with greet
* greet
- utter_answer_greet
* whoareyou
- utter_answer_whoareyou

## what to do
* whattodo
- utter_answer_whattodo

## what to do with greet
* greet
- utter_answer_greet
* whattodo
- utter_answer_whattodo

## happy path
* request_weather
- weather_form
- form{"name": "weather_form"}
- form{"name": null}

3.3.2 domain.yml

intents:
  - affirm
  - deny
  - greet
  - goodbye
  - thanks
  - whoareyou
  - whattodo
  - request_weather

slots:
  date-time:
type: unfeaturized
  address:
type: unfeaturized

entities:
  - date-time
  - address

actions:
  - utter_answer_affirm
  - utter_answer_deny
  - utter_answer_greet
  - utter_answer_goodbye
  - utter_answer_thanks
  - utter_answer_whoareyou
  - utter_answer_whattodo
  - utter_ask_date-time
  - utter_ask_address
  - action_default_fallback

forms:
  - weather_form

responses:
  utter_answer_affirm:
- text: "嗯嗯，好的！"
- text: "嗯嗯，很开心能够帮您解决问题~"
- text: "嗯嗯，还需要什么我能够帮助您的呢？"

  utter_answer_greet:
- text: "您好！请问我可以帮到您吗？"
- text: "您好！很高兴为您服务。请说出您要查询的功能？"

  utter_answer_goodbye:
- text: "再见"
- text: "拜拜"
- text: "虽然我有万般舍不得，但是天下没有不散的宴席~祝您安好！"
- text: "期待下次再见！"
- text: "嗯嗯，下次需要时随时记得我哟~"
- text: "see you!"

  utter_answer_deny:
- text: "主人，您不开心吗？不要离开我哦"
- text: "怎么了，主人？"

  utter_answer_thanks:
- text: "嗯呢。不用客气~"
- text: "这是我应该做的，主人~"
- text: "嗯嗯，合作愉快！"

  utter_answer_whoareyou:
- text: "您好！我是小蒋呀，您的AI智能助理"

  utter_answer_whattodo:
- text: "您好！很高兴为您服务，我目前只支持查询天气哦。"

  utter_ask_date-time:
- text: "请问您要查询哪一天的天气？"

  utter_ask_address:
- text: "请问您要查下哪里的天气？"

  utter_default:
- text: "没听懂，请换种说法吧~"

3.4 训练NLU和CORE模型

3.4.1 config.yml

训练NLU和Core模型配置文件：

language: "zh"

pipeline:
- name: "MitieNLP"
  model: "data/total_word_feature_extractor_zh.dat"
- name: "JiebaTokenizer"
- name: "MitieEntityExtractor"
- name: "EntitySynonymMapper"
- name: "RegexFeaturizer"
- name: "MitieFeaturizer"
- name: "SklearnIntentClassifier"

policies:
  - name: KerasPolicy
    epochs: 500
    max_history: 5
  - name: FallbackPolicy
    fallback_action_name: 'action_default_fallback'
  - name: MemoizationPolicy
    max_history: 5
  - name: FormPolicy

3.4.2 模型训练

验证

$ python -m rasa data validate
/home/amy/.conda/envs/rasa/lib/python3.6/site-packages/rasa/core/domain.py:151: FutureWarning: No tracker session configuration was found in the loaded domain. Domains without a session config will automatically receive a session expiration time of 60 minutes in Rasa version 2.0 if not configured otherwise.
  session_config = cls._get_session_config(data.get(SESSION_CONFIG_KEY, {}))
2020-09-25 09:15:00 INFO     rasa.validator  - Validating intents...
2020-09-25 09:15:00 INFO     rasa.validator  - Validating uniqueness of intents and stories...
2020-09-25 09:15:00 INFO     rasa.validator  - Validating utterances...
2020-09-25 09:15:00 INFO     rasa.validator  - Story structure validation...
Processed Story Blocks: 100%|█████████████████████████████████████| 17/17 [00:00<00:00, 4121.57it/s, # trackers=1]
2020-09-25 09:15:00 INFO     rasa.core.training.story_conflict  - Considering the preceding 6 turns for conflict analysis.
2020-09-25 09:15:00 INFO     rasa.validator  - No story structure conflicts found.

运行

当所有样本和配置文件准备好后，接下来就是训练模型了，打开Pycharm命令终端执行下面的命令，该命令会同时训练NLU和Core模型，具体如下：

python -m rasa train --config config.yml --domain domain.yml --data data/

参数介绍

usage: rasa train [-h] [-v] [-vv] [--quiet] [--data DATA [DATA ...]]
  [-c CONFIG] [-d DOMAIN] [--out OUT]
  [--augmentation AUGMENTATION] [--DEBUG -plots]
  [--dump-stories] [--fixed-model-name FIXED_MODEL_NAME]
  [--persist-nlu-data] [--force]
  {core,nlu} ...

positional arguments:
  {core,nlu}
core指定训练的模型为core模型
nlu 指定选了的模型为nlu模型

optional arguments:
  -h, --help帮助信息；
  --data指定NLU和Core模型所有样本文件，默认为data目录；
  -c 或--config 指定policy和nlu pipeline配置文件，默认为根目录下config.ym；
  -d 或--domain 指定domain.yml文件，默认为根目录下domain.yml；
  --out 指定模型文件输出路径，默认为自定生成models；
  --augmentation指定训练时需要多少数据augmentation(扩展)，默认为50；
  --DEBUG -plots 一般不用
  --dump-stories是否开启将flattened stories保存到文件，默认为false；
  --fixed-model-name指定生成的模型文件名称，默认none
  --persist-nlu-data是否一定要将nlu训练数据保存到模型，默认为false；
  --force   是否强化模型当训练数据没有变化时，默认为false

Python Logging Options:
  -v, --verbose 开启打印日志；
  -vv, --DEBUG   开启调试模式；
  --quiet   设置日志打印级别为WARNING；

四、配置Http和Action

4.1 credentials.yml

credentials.yml为配置连接到其他服务的详细(认证)信息，当我们需要通过Http的形式访问Rasa Server时，就需要在该文件中配置rest:。rest通道将为您提供一个rest端点（即Rasa Server），用于向其发送消息，响应该请求将发送回bots消息。根据这个文档的说明，当我们请求Rasa Server的URL应为：

http://rasaServerIP:rasaServerPort/webhooks/rest/webhook

该文件内容如下：

# This file contains the credentials for the voice & chat platforms
# which your bot is using.
# https://rasa.com/docs/rasa/user-guide/messaging-and-voice-channels/

rest:
#  # you don't need to provide anything here - this channel doesn't
#  # require any credentials


#facebook:
#  verify: "<verify>"
#  secret: "<your secret>"
#  page-access-token: "<your page access token>"

#slack:
#  slack_token: "<your slack token>"
#  slack_channel: "<the slack channel>"

#socketio:
#  user_message_evt: <event name for user message>
#  bot_message_evt: <event name for but messages>
#  session_persistence: <true/false>

#mattermost:
#  url: "https://<mattermost instance>/api/v4"
#  token: "<bot token>"
#  webhook_url: "<callback URL>"

# This entry is needed if you are using Rasa X. The entry represents credentials 
# for the Rasa X "channel", i.e. Talk to your bot and Share with guest testers.
rasa:
  url: "http://localhost:5002/api"

4.2 endpoints.yml

如果希望rasa server(注：指rasa core)能够连接到其他web，我们可以再endpoints.yml这个文件中进行配置，比如为了实现custom action，我们就需要在该文件中对action server进行配置，又比如我们将nlu模块放到其他的web项目中，就需要在该文件中配置nlu server等等。endpoints.yml文件内容如下：

# 指定action server的url
# 当然，也可以将action server单独实现在一个web server项目中
# 那么这个url为"https://yourWebIp:yourWebPort/webhook“
action_endpoint:
 url: "http://localhost:5055/webhook"
 
# 配置nlu(单独创建一个web项目):
#  url: "http://10.0.0.153:5000/"

# 配置tracker信息存储服务器
#tracker_store:
#    type: mongod
#    url: mongodb://localhost:27017
#    db: rasa
#    username:
#    password:
#    auth_source: rasa

4.3 action.py

当Rasa NLU识别到用户输入Message的意图后，Rasa Core对话管理模块就会对其作出回应，而完成这个回应的模块就是action。Rasa Core支持三种action，即default actions、utter actions以及 custom action，这部分我们将在后面详细讲解。另外，本项目为了测试，还对接了图灵机器人和心知天气的API，代码就没有给出了，感兴趣的可以再文末Github源码中查看，这里只给出default action和custom action代码。具体如下：

from typing import Dict, Text, Any, List
from rasa_sdk import Tracker, Action
from rasa_sdk.events import UserUtteranceReverted, Restarted
from rasa_sdk.executor import CollectingDispatcher
from rasa_sdk.forms import FormAction
from requests import (
  ConnectionError,
  HTTPError,
  TooManyRedirects,
  Timeout
)

# action weather_form
class WeatherForm(FormAction):
def name(self) -> Text:
"""Unique identifier of the form"""
return "weather_form"

@staticmethod
def required_slots(tracker: Tracker) -> List[Text]:
"""A list of required slots that the form has to fill"""
return ["date-time", "address"]

def submit(
self,
dispatcher: CollectingDispatcher,
tracker: Tracker,
domain: Dict[Text, Any],
) -> List[Dict]:
"""Define what the form has to do
after all required slots are filled"""
address = tracker.get_slot('address')
date_time = tracker.get_slot('date-time')
print(f"action_default_fallback->address:{address}")
print(f"action_default_fallback->date_time:{date_time}")
dispatcher.utter_message("正在为你查询 {} {}的天气 ing".format(address,date_time))
return [Restarted()]
   
# action_default_fallback
class ActionDefaultFallback(Action):
"""Executes the fallback action and goes back to the previous state
of the dialogue"""
def name(self):
return 'action_default_fallback'

def run(self, dispatcher, tracker, domain):
text = tracker.latest_message.get('text')
print(f"action_default_fallback->text:{text}")
dispatcher.utter_template('utter_default', tracker, silent_fail=True)
return [UserUtteranceReverted()]

五、启动对话

5.1 通过 post 请求查询

（1）启动Rasa服务

输入下面命令：

# 启动rasa服务
# 该服务实现自然语言理解(NLU)和对话管理(Core)功能
# 注：该服务的--port默认为5005，如果使用默认则可以省略

$ python -m rasa run --port 5005 --endpoints endpoints.yml --credentials credentials.yml --DEBUG 
2020-09-21 20:59:49 DEBUG rasa.model  - Extracted model to '/tmp/tmpi1ucn6wd'.
2020-09-21 20:59:49 DEBUG sanic.root  - CORS: Configuring CORS with resources: {'/*': {'origins': [''], 'methods': 'DELETE, GET, HEAD, OPTIONS, PATCH, POST, PUT', 'allow_headers': ['.*'], 'expose_headers': 'filename', 'supports_credentials': True, 'max_age': None, 'send_wildcard': False, 'automatic_options': True, 'vary_header': True, 'resources': {'/*': {'origins': ''}}, 'intercept_exceptions': True, 'always_send': True}}
2020-09-21 20:59:49 DEBUG rasa.core.utils  - Available web server routes: 
/webhooks/rasa GETcustom_webhook_RasaChatInput.health
/webhooks/rasa/webhook POST   custom_webhook_RasaChatInput.receive
/webhooks/rest GETcustom_webhook_RestInput.health
/webhooks/rest/webhook POST   custom_webhook_RestInput.receive
/  GEThello
2020-09-21 20:59:49 INFO root  - Starting Rasa server on http://localhost:5005
2020-09-21 20:59:49 DEBUG rasa.core.utils  - Using the default number of Sanic workers (1).
2020-09-21 20:59:49 INFO root  - Enabling coroutine DEBUG ging. Loop id 77433480.
2020-09-21 20:59:49 DEBUG rasa.model  - Extracted model to '/tmp/tmprhshce_e'.
2020-09-21 20:59:52 INFO rasa.nlu.components  - Added 'MitieNLP' to component cache. Key 'MitieNLP-/web/workspace/yangkm/python_wp/nlu/DSwP/rasa_zh_my/data/total_word_feature_extractor_zh.dat'.
2020-09-21 20:59:52 DEBUG rasa.core.tracker_store  - Connected to InMemoryTrackerStore.
2020-09-21 20:59:52 DEBUG rasa.core.lock_store  - Connected to lock store 'InMemoryLockStore'.
2020-09-21 20:59:52 DEBUG rasa.model  - Extracted model to '/tmp/tmpxex779pr'.
2020-09-21 20:59:52 DEBUG pykwalify.compat  - Using yaml library: /home/amy/.conda/envs/rasa/lib/python3.6/site-packages/ruamel/yaml/__init__.py
/home/amy/.conda/envs/rasa/lib/python3.6/site-packages/rasa/core/policies/keras_policy.py:265: FutureWarning: 'KerasPolicy' is deprecated and will be removed in version 2.0. Use 'TEDPolicy' instead.
  current_epoch=meta["epochs"],
2020-09-21 20:59:55 INFO rasa.core.policies.ensemble  - MappingPolicy not included in policy ensemble. Default intents 'restart and back will not trigger actions 'action_restart' and 'action_back'.
2020-09-21 20:59:55 DEBUG rasa.core.nlg.generator  - Instantiated NLG to 'TemplatedNaturalLanguageGenerator'.
2020-09-21 20:59:55 INFO root  - Rasa server is up and running.

参数说明

usage: rasa run [-h] [-v] [-vv] [--quiet] [-m MODEL] [--log-file LOG_FILE]
[--endpoints ENDPOINTS] [-p PORT] [-t AUTH_TOKEN]
[--cors [CORS [CORS ...]]] [--enable-api]
[--remote-storage REMOTE_STORAGE]
[--ssl-certificate SSL_CERTIFICATE]
[--ssl-keyfile SSL_KEYFILE] [--ssl-ca-file SSL_CA_FILE]
[--ssl-password SSL_PASSWORD] [--credentials CREDENTIALS]
[--connector CONNECTOR] [--jwt-secret JWT_SECRET]
[--jwt-method JWT_METHOD]
{actions} ... [model-as-positional-argument]

positional arguments:
  {actions}
actions 运行action server
  model-as-positional-argument 

optional arguments:
  -h, --help  
  						显示帮助信息；
  -m MODEL, --model MODEL 
  						指定训练好的模型路径，默认使用models目录。
  						如果模型存储路径改变，则需要该参数指定；
  --log-file LOG_FILE 
  						指定保存logs文件，默认为None；
  --endpoints ENDPOINTS   
  						指定endpoints.yml文件路径，默认为None;

Python Logging Options:
  -v, --verbose 
  						设置日志等级为INFO;
  -vv, --DEBUG   
  						开启调试
  --quiet   
  						设置日志等级为WARNING，默认为None;

Server Settings:
  -p PORT, --port PORT  
  						设置运行rasa serve的端口号，默认为5005；
  -t AUTH_TOKEN, --auth-token AUTH_TOKEN
开启token身份验证，默认为None;
  --cors [CORS [CORS ...]]
  --enable-api  
  						启动web服务器API通道，默认值:False；
  --remote-storage REMOTE_STORAGE
设置rasa模型远程位置，如果有的话；
  --ssl-certificate SSL_CERTIFICATE
设置SSL证书，默认为None;
  --ssl-keyfile SSL_KEYFILE
设置SSL密钥文件，默认为None;
  --ssl-ca-file SSL_CA_FILE
设置CA文件便于SSL证书验证，默认为None;
  --ssl-password SSL_PASSWORD
  						设置SSL密钥文件密码，默认为None;

Channels:
  --credentials CREDENTIALS
指定credentials.yml文件路径；
  --connector CONNECTOR
Service to connect to. (default: None)

JWT Authentication:
  --jwt-secret JWT_SECRET
指定jwt-secret；
  --jwt-method JWT_METHOD
指定jwt-method；

（2）启动Action服务

# 启动action服务
# 注：该服务的--port默认为5055，如果使用默认则可以省略
$ python -m rasa run actions --port 5055 --actions actions --DEBUG 
2020-09-21 21:39:28 INFO rasa_sdk.endpoint  - Starting action endpoint server...
2020-09-21 21:39:28 INFO rasa_sdk.executor  - Registered function for 'action_default_fallback'.
2020-09-21 21:39:28 INFO rasa_sdk.executor  - Registered function for 'weather_form'.
2020-09-21 21:39:28 INFO rasa_sdk.endpoint  - Action endpoint is up and running on http://localhost:5055
2020-09-21 21:39:28 DEBUG rasa_sdk.utils  - Using the default number of Sanic workers (1).

参数说明

usage: rasa run actions [-h] [-v] [-vv] [--quiet] [-p PORT]
[--cors [CORS [CORS ...]]] [--actions ACTIONS]
[--ssl-keyfile SSL_KEYFILE]
[--ssl-certificate SSL_CERTIFICATE]
[--ssl-password SSL_PASSWORD]

optional arguments:
  -h, --help
  						显示帮助信息
  -p PORT, --port PORT  
  						指定action server的端口号，默认为5055；
  --cors [CORS [CORS ...]]
开启CORS;
  --actions ACTIONS 
  						指定action.py等文件所在包路径；
  --ssl-certificate SSL_CERTIFICATE
设置SSL证书，默认为None;
  --ssl-keyfile SSL_KEYFILE
设置SSL密钥文件，默认为None;
  --ssl-ca-file SSL_CA_FILE
设置CA文件便于SSL证书验证，默认为None;

Python Logging Options:
  -v, --verbose 
  						设置日志等级为INFO;
  -vv, --DEBUG   
  						开启调试
  --quiet   
  						设置日志等级为WARNING，默认为None;

（3）效果演示

另外开启一个窗口，输入：

$ curl -X POST localhost:5005/webhooks/rest/webhook -d '{"message":"询深圳周五的天气","sender":"1"}' | python -m json.tool
  % Total% Received % Xferd  Average Speed   TimeTime Time  Current
 Dload  Upload   Total   SpentLeft  Speed
100   155  100   104  10051   2355   1155 --:--:-- --:--:-- --:--:--  2363
[
{
"recipient_id": "1",
"text": "\u6682\u4e0d\u652f\u6301\u67e5\u8be2 ['\u6df1\u5733'] \u7684\u5929\u6c14"
}
]

Rasa Server 反映

2020-09-21 21:40:06 DEBUG rasa.core.lock_store  - Issuing ticket for conversation '1'.
2020-09-21 21:40:06 DEBUG rasa.core.lock_store  - Acquiring lock for conversation '1'.
2020-09-21 21:40:06 DEBUG rasa.core.lock_store  - Acquired lock for conversation '1'.
2020-09-21 21:40:06 DEBUG rasa.core.tracker_store  - Recreating tracker for id '1'
2020-09-21 21:40:06 DEBUG rasa.core.processor  - Received user message '询深圳周五的天气' with intent '{'name': 'request_weather', 'confidence': 0.49303004566904013}' and entities '[{'entity': 'address', 'value': '深圳', 'start': 1, 'end': 3, 'confidence': None, 'extractor': 'MitieEntityExtractor'}, {'entity': 'date-time', 'value': '周五', 'start': 3, 'end': 5, 'confidence': None, 'extractor': 'MitieEntityExtractor'}]'
2020-09-21 21:40:06 DEBUG rasa.core.processor  - Current slot values: 
	address: 深圳
	date-time: 周五
	requested_slot: None
2020-09-21 21:40:06 DEBUG rasa.core.processor  - Logged UserUtterance - tracker now has 38 events.
2020-09-21 21:40:06 DEBUG rasa.core.policies.fallback  - NLU confidence threshold met, confidence of fallback action set to core threshold (0.3).
2020-09-21 21:40:06 DEBUG rasa.core.policies.memoization  - Current tracker state [None, None, None, {}, {'prev_action_listen': 1.0, 'entity_date-time': 1.0, 'entity_address': 1.0, 'intent_request_weather': 1.0}]
2020-09-21 21:40:06 DEBUG rasa.core.policies.memoization  - There is no memorised next action
2020-09-21 21:40:06 DEBUG rasa.core.policies.form_policy  - There is no active form
2020-09-21 21:40:06 DEBUG rasa.core.policies.ensemble  - Predicted next action using policy_0_KerasPolicy
2020-09-21 21:40:06 DEBUG rasa.core.processor  - Predicted next action 'weather_form' with confidence 0.99.
2020-09-21 21:40:06 DEBUG rasa.core.actions.action  - Calling action endpoint to run action 'weather_form'.
2020-09-21 21:40:06 DEBUG rasa.core.processor  - Action 'weather_form' ended with events '[BotUttered('暂不支持查询 ['深圳'] 的天气', {"elements": null, "quick_replies": null, "buttons": null, "attachment": null, "image": null, "custom": null}, {}, 1600695606.1912577), <rasa.core.events.Form object at 0x7f46445940b8>, <rasa.core.events.SlotSet object at 0x7f4644594278>, <rasa.core.events.SlotSet object at 0x7f4644594240>, <rasa.core.events.SlotSet object at 0x7f4644594160>, <rasa.core.events.SlotSet object at 0x7f46445941d0>, <rasa.core.events.Restarted object at 0x7f46445944a8>, <rasa.core.events.Form object at 0x7f4644594128>, <rasa.core.events.SlotSet object at 0x7f46445944e0>]'.
2020-09-21 21:40:06 DEBUG rasa.core.processor  - Current slot values: 
	address: None
	date-time: None
	requested_slot: None
2020-09-21 21:40:06 DEBUG rasa.core.processor  - Predicted next action 'action_listen' with confidence 1.00.
2020-09-21 21:40:06 DEBUG rasa.core.processor  - Action 'action_listen' ended with events '[]'.
2020-09-21 21:40:06 DEBUG rasa.core.lock_store  - Deleted lock for conversation '1'.

Action Server 反应

5.2 聊天窗式对话

（一）启动Action服务

# 启动action服务
# 注：该服务的--port默认为5055，如果使用默认则可以省略
$ python -m rasa run actions --port 5055 --actions actions --debug
2020-09-21 21:39:28 INFO rasa_sdk.endpoint  - Starting action endpoint server...
2020-09-21 21:39:28 INFO rasa_sdk.executor  - Registered function for 'action_default_fallback'.
2020-09-21 21:39:28 INFO rasa_sdk.executor  - Registered function for 'weather_form'.
2020-09-21 21:39:28 INFO rasa_sdk.endpoint  - Action endpoint is up and running on http://localhost:5055
2020-09-21 21:39:28 DEBUG rasa_sdk.utils  - Using the default number of Sanic workers (1).

参数说明

usage: rasa run actions [-h] [-v] [-vv] [--quiet] [-p PORT]
[--cors [CORS [CORS ...]]] [--actions ACTIONS]
[--ssl-keyfile SSL_KEYFILE]
[--ssl-certificate SSL_CERTIFICATE]
[--ssl-password SSL_PASSWORD]

optional arguments:
  -h, --help
  						显示帮助信息
  -p PORT, --port PORT  
  						指定action server的端口号，默认为5055；
  --cors [CORS [CORS ...]]
开启CORS;
  --actions ACTIONS 
  						指定action.py等文件所在包路径；
  --ssl-certificate SSL_CERTIFICATE
设置SSL证书，默认为None;
  --ssl-keyfile SSL_KEYFILE
设置SSL密钥文件，默认为None;
  --ssl-ca-file SSL_CA_FILE
设置CA文件便于SSL证书验证，默认为None;

Python Logging Options:
  -v, --verbose 
  						设置日志等级为INFO;
  -vv, --DEBUG   
  						开启调试
  --quiet   
  						设置日志等级为WARNING，默认为None;

（二）开启对话窗口

  $ rasa shell
  2020-09-22 10:05:42 INFO root  - Connecting to channel 'cmdline' which was specified by the '--connector' argument. Any other channels will be ignored. To connect to all given channels, omit the '--connector' argument.
  2020-09-22 10:05:42 INFO root  - Starting Rasa server on http://localhost:5005
  2020-09-22 10:05:46 INFO rasa.nlu.components  - Added 'MitieNLP' to component cache. Key 'MitieNLP-/web/workspace/yangkm/python_wp/nlu/DSwP/rasa_zh_my/data/total_word_feature_extractor_zh.dat'.
  /home/amy/.conda/envs/rasa/lib/python3.6/site-packages/rasa/core/policies/keras_policy.py:265: FutureWarning: 'KerasPolicy' is deprecated and will be removed in version 2.0. Use 'TEDPolicy' instead.
current_epoch=meta["epochs"],
  2020-09-22 10:05:50 INFO rasa.core.policies.ensemble  - MappingPolicy not included in policy ensemble. Default intents 'restart and back will not trigger actions 'action_restart' and 'action_back'.
  2020-09-22 10:05:50 INFO root  - Rasa server is up and running.
  Bot loaded. Type a message and press enter (use '/stop' to exit): 
  Your input ->

（三）

场景一

对话内容

  Your input ->  查询天气  
  请问您要查询哪一天的天气？
  Your input ->  大后天 
  请问您要查下哪里的天气？
  Your input ->  哈尔滨 
  正在为你查询 哈尔滨 大后天的天气 ing

后台结果

2020-09-22 10:06:45 DEBUG rasa_sdk.forms  - Validating user input '{'intent': {'name': 'request_weather', 'confidence': 0.6752573663579792}, 'entities': [], 'intent_ranking': [{'name': 'request_weather', 'confidence': 0.6752573663579792}, {'name': 'deny', 'confidence': 0.08473735571280376}, {'name': 'affirm', 'confidence': 0.07509502417330365}, {'name': 'greet', 'confidence': 0.05720770660563366}, {'name': 'thanks', 'confidence': 0.03785058670945419}, {'name': 'goodbye', 'confidence': 0.03603811968623011}, {'name': 'whattodo', 'confidence': 0.02145608716337933}, {'name': 'whoareyou', 'confidence': 0.01235775359121649}], 'text': '查询天气'}'
2020-09-22 10:06:45 DEBUG rasa_sdk.forms  - Validating extracted slots: {}
2020-09-22 10:06:45 DEBUG rasa_sdk.forms  - Request next slot 'date-time'
2020-09-22 10:06:45 DEBUG rasa_sdk.executor  - Finished running 'weather_form'
2020-09-22 10:06:51 DEBUG rasa_sdk.executor  - Received request to run 'weather_form'
2020-09-22 10:06:51 DEBUG rasa_sdk.forms  - The form '{'name': 'weather_form', 'validate': True, 'rejected': False, 'trigger_message': {'intent': {'name': 'request_weather', 'confidence': 0.6752573663579792}, 'entities': [], 'intent_ranking': [{'name': 'request_weather', 'confidence': 0.6752573663579792}, {'name': 'deny', 'confidence': 0.08473735571280376}, {'name': 'affirm', 'confidence': 0.07509502417330365}, {'name': 'greet', 'confidence': 0.05720770660563366}, {'name': 'thanks', 'confidence': 0.03785058670945419}, {'name': 'goodbye', 'confidence': 0.03603811968623011}, {'name': 'whattodo', 'confidence': 0.02145608716337933}, {'name': 'whoareyou', 'confidence': 0.01235775359121649}], 'text': '查询天气'}}' is active
2020-09-22 10:06:51 DEBUG rasa_sdk.forms  - Validating user input '{'intent': {'name': 'request_weather', 'confidence': 0.9364390143513037}, 'entities': [{'entity': 'date-time', 'value': '大后天', 'start': 0, 'end': 3, 'confidence': None, 'extractor': 'MitieEntityExtractor'}], 'intent_ranking': [{'name': 'request_weather', 'confidence': 0.9364390143513037}, {'name': 'goodbye', 'confidence': 0.02844964603391187}, {'name': 'greet', 'confidence': 0.01407263502906878}, {'name': 'affirm', 'confidence': 0.0070474768555266265}, {'name': 'thanks', 'confidence': 0.006487673729264043}, {'name': 'whattodo', 'confidence': 0.0032935182462639516}, {'name': 'deny', 'confidence': 0.0027266823314896858}, {'name': 'whoareyou', 'confidence': 0.0014833534231716052}], 'text': '大后天'}'
2020-09-22 10:06:51 DEBUG rasa_sdk.forms  - Trying to extract requested slot 'date-time' ...
2020-09-22 10:06:51 DEBUG rasa_sdk.forms  - Got mapping '{'type': 'from_entity', 'entity': 'date-time', 'intent': [], 'not_intent': [], 'role': None, 'group': None}'
2020-09-22 10:06:51 DEBUG rasa_sdk.forms  - Successfully extracted '大后天' for requested slot 'date-time'
2020-09-22 10:06:51 DEBUG rasa_sdk.forms  - Validating extracted slots: {'date-time': '大后天'}
2020-09-22 10:06:51 DEBUG rasa_sdk.forms  - Request next slot 'address'
2020-09-22 10:06:51 DEBUG rasa_sdk.executor  - Finished running 'weather_form'
2020-09-22 10:07:13 DEBUG rasa_sdk.executor  - Received request to run 'weather_form'
2020-09-22 10:07:13 DEBUG rasa_sdk.forms  - The form '{'name': 'weather_form', 'validate': True, 'rejected': False, 'trigger_message': {'intent': {'name': 'request_weather', 'confidence': 0.6752573663579792}, 'entities': [], 'intent_ranking': [{'name': 'request_weather', 'confidence': 0.6752573663579792}, {'name': 'deny', 'confidence': 0.08473735571280376}, {'name': 'affirm', 'confidence': 0.07509502417330365}, {'name': 'greet', 'confidence': 0.05720770660563366}, {'name': 'thanks', 'confidence': 0.03785058670945419}, {'name': 'goodbye', 'confidence': 0.03603811968623011}, {'name': 'whattodo', 'confidence': 0.02145608716337933}, {'name': 'whoareyou', 'confidence': 0.01235775359121649}], 'text': '查询天气'}}' is active
2020-09-22 10:07:13 DEBUG rasa_sdk.forms  - Validating user input '{'intent': {'name': 'request_weather', 'confidence': 0.2875610430157071}, 'entities': [{'entity': 'address', 'value': '哈尔滨', 'start': 0, 'end': 3, 'confidence': None, 'extractor': 'MitieEntityExtractor'}], 'intent_ranking': [{'name': 'request_weather', 'confidence': 0.2875610430157071}, {'name': 'greet', 'confidence': 0.2872763354473037}, {'name': 'goodbye', 'confidence': 0.12110821662968022}, {'name': 'affirm', 'confidence': 0.0787910767204003}, {'name': 'thanks', 'confidence': 0.0736496158800513}, {'name': 'whattodo', 'confidence': 0.06899456035142955}, {'name': 'deny', 'confidence': 0.05765607701548124}, {'name': 'whoareyou', 'confidence': 0.024963074939946338}], 'text': '哈尔滨'}'
2020-09-22 10:07:13 DEBUG rasa_sdk.forms  - Trying to extract requested slot 'address' ...
2020-09-22 10:07:13 DEBUG rasa_sdk.forms  - Got mapping '{'type': 'from_entity', 'entity': 'address', 'intent': [], 'not_intent': [], 'role': None, 'group': None}'
2020-09-22 10:07:13 DEBUG rasa_sdk.forms  - Successfully extracted '哈尔滨' for requested slot 'address'
2020-09-22 10:07:13 DEBUG rasa_sdk.forms  - Validating extracted slots: {'address': '哈尔滨'}
2020-09-22 10:07:13 DEBUG rasa_sdk.forms  - No slots left to request, all required slots are filled:
	date-time: 大后天
	address: 哈尔滨
2020-09-22 10:07:13 DEBUG rasa_sdk.forms  - Submitting the form 'weather_form'
action_default_fallback->address:哈尔滨
action_default_fallback->date_time:大后天
2020-09-22 10:07:13 DEBUG rasa_sdk.forms  - Deactivating the form 'weather_form'
2020-09-22 10:07:13 DEBUG rasa_sdk.executor  - Finished running 'weather_form'

场景二

对话内容

  Your input ->  明天的天气 
  请问您要查下哪里的天气？
  Your input ->  深圳   
  正在为你查询 深圳 明天的天气 ing
  Your input ->  深圳的天气 
  请问您要查询哪一天的天气？
  Your input ->  后台   
  正在为你查询 深圳 后台的天气 ing

后台结果

2020-09-22 10:07:23 DEBUG     rasa_sdk.forms  - Activated the form 'weather_form'
2020-09-22 10:07:23 DEBUG     rasa_sdk.forms  - Validating pre-filled required slots: {'date-time': '明天'}
2020-09-22 10:07:23 DEBUG     rasa_sdk.forms  - Validating user input '{'intent': {'name': 'request_weather', 'confidence': 0.859133506308341}, 'entities': [{'entity': 'date-time', 'value': '明天', 'start': 0, 'end': 2, 'confidence': None, 'extractor': 'MitieEntityExtractor'}], 'intent_ranking': [{'name': 'request_weather', 'confidence': 0.859133506308341}, {'name': 'deny', 'confidence': 0.03325788518527375}, {'name': 'affirm', 'confidence': 0.03310349827249277}, {'name': 'greet', 'confidence': 0.02768090617777744}, {'name': 'thanks', 'confidence': 0.018412903569335396}, {'name': 'goodbye', 'confidence': 0.016649633914462992}, {'name': 'whattodo', 'confidence': 0.005943813262005036}, {'name': 'whoareyou', 'confidence': 0.005817853310311661}], 'text': '明天的天气'}'
2020-09-22 10:07:23 DEBUG     rasa_sdk.forms  - Extracted '明天' for extra slot 'date-time'.
2020-09-22 10:07:23 DEBUG     rasa_sdk.forms  - Validating extracted slots: {'date-time': '明天'}
2020-09-22 10:07:23 DEBUG     rasa_sdk.forms  - Request next slot 'address'
2020-09-22 10:07:23 DEBUG     rasa_sdk.executor  - Finished running 'weather_form'
2020-09-22 10:07:26 DEBUG     rasa_sdk.executor  - Received request to run 'weather_form'
2020-09-22 10:07:26 DEBUG     rasa_sdk.forms  - The form '{'name': 'weather_form', 'validate': True, 'rejected': False, 'trigger_message': {'intent': {'name': 'request_weather', 'confidence': 0.859133506308341}, 'entities': [{'entity': 'date-time', 'value': '明天', 'start': 0, 'end': 2, 'confidence': None, 'extractor': 'MitieEntityExtractor'}], 'intent_ranking': [{'name': 'request_weather', 'confidence': 0.859133506308341}, {'name': 'deny', 'confidence': 0.03325788518527375}, {'name': 'affirm', 'confidence': 0.03310349827249277}, {'name': 'greet', 'confidence': 0.02768090617777744}, {'name': 'thanks', 'confidence': 0.018412903569335396}, {'name': 'goodbye', 'confidence': 0.016649633914462992}, {'name': 'whattodo', 'confidence': 0.005943813262005036}, {'name': 'whoareyou', 'confidence': 0.005817853310311661}], 'text': '明天的天气'}}' is active
2020-09-22 10:07:26 DEBUG     rasa_sdk.forms  - Validating user input '{'intent': {'name': 'request_weather', 'confidence': 0.5679284847462811}, 'entities': [{'entity': 'address', 'value': '深圳', 'start': 0, 'end': 2, 'confidence': None, 'extractor': 'MitieEntityExtractor'}], 'intent_ranking': [{'name': 'request_weather', 'confidence': 0.5679284847462811}, {'name': 'greet', 'confidence': 0.1836000885202259}, {'name': 'thanks', 'confidence': 0.06674624179701746}, {'name': 'deny', 'confidence': 0.050381853263852465}, {'name': 'affirm', 'confidence': 0.044965754284358395}, {'name': 'goodbye', 'confidence': 0.04295301432924679}, {'name': 'whattodo', 'confidence': 0.03497770707600284}, {'name': 'whoareyou', 'confidence': 0.00844685598301393}], 'text': '深圳'}'
2020-09-22 10:07:26 DEBUG     rasa_sdk.forms  - Trying to extract requested slot 'address' ...
2020-09-22 10:07:26 DEBUG     rasa_sdk.forms  - Got mapping '{'type': 'from_entity', 'entity': 'address', 'intent': [], 'not_intent': [], 'role': None, 'group': None}'
2020-09-22 10:07:26 DEBUG     rasa_sdk.forms  - Successfully extracted '深圳' for requested slot 'address'
2020-09-22 10:07:26 DEBUG     rasa_sdk.forms  - Validating extracted slots: {'address': '深圳'}
2020-09-22 10:07:26 DEBUG     rasa_sdk.forms  - No slots left to request, all required slots are filled:
	date-time: 明天
	address: 深圳
2020-09-22 10:07:26 DEBUG     rasa_sdk.forms  - Submitting the form 'weather_form'
action_default_fallback->address:深圳
action_default_fallback->date_time:明天
2020-09-22 10:07:26 DEBUG     rasa_sdk.forms  - Deactivating the form 'weather_form'
2020-09-22 10:07:26 DEBUG     rasa_sdk.executor  - Finished running 'weather_form'
2020-09-22 10:07:33 DEBUG     rasa_sdk.executor  - Received request to run 'weather_form'
2020-09-22 10:07:33 DEBUG     rasa_sdk.forms  - There is no active form
2020-09-22 10:07:33 DEBUG     rasa_sdk.forms  - Activated the form 'weather_form'
2020-09-22 10:07:33 DEBUG     rasa_sdk.forms  - Validating pre-filled required slots: {'address': '深圳'}
2020-09-22 10:07:33 DEBUG     rasa_sdk.forms  - Validating user input '{'intent': {'name': 'request_weather', 'confidence': 0.6615162196934451}, 'entities': [{'entity': 'address', 'value': '深圳', 'start': 0, 'end': 2, 'confidence': None, 'extractor': 'MitieEntityExtractor'}], 'intent_ranking': [{'name': 'request_weather', 'confidence': 0.6615162196934451}, {'name': 'affirm', 'confidence': 0.0900704675472}, {'name': 'deny', 'confidence': 0.08134485744568894}, {'name': 'greet', 'confidence': 0.06553939061503906}, {'name': 'thanks', 'confidence': 0.046344364587568965}, {'name': 'goodbye', 'confidence': 0.0333352458887086}, {'name': 'whattodo', 'confidence': 0.014025037948307274}, {'name': 'whoareyou', 'confidence': 0.007824416274041972}], 'text': '深圳的天气'}'
2020-09-22 10:07:33 DEBUG     rasa_sdk.forms  - Extracted '深圳' for extra slot 'address'.
2020-09-22 10:07:33 DEBUG     rasa_sdk.forms  - Validating extracted slots: {'address': '深圳'}
2020-09-22 10:07:33 DEBUG     rasa_sdk.forms  - Request next slot 'date-time'
2020-09-22 10:07:33 DEBUG     rasa_sdk.executor  - Finished running 'weather_form'

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

rasa中文对话系统构建

【关于 rasa中文对话系统构建】那些你不知道的事

目录

一、安装 Rasa 内容

1.1 安装中文版本的 Rasa-nlu

1.2 安装 rasa

安装过程中遇到的问题

1.3 sklearn 和 MITIE 库安装

1.4 安装 rasa_core

二、新建项目

2.1 安装rasa，自动生成rasa项目所需文件

2.2 测试助手

三、Rasa使用–构建简单聊天机器人

3.1 Rasa工作原理

3.2 构建NLU样本

3.2.1 data/nlu.md

3.3 构建Core样本

3.3.1 data/stories.md

3.3.2 domain.yml

3.4 训练NLU和CORE模型

3.4.1 config.yml

3.4.2 模型训练

四、配置Http和Action

4.1 credentials.yml

4.2 endpoints.yml

4.3 action.py

五、启动对话

5.1 通过 post 请求查询

（1）启动Rasa服务

（2）启动Action服务

（3）效果演示

5.2 聊天窗式对话

（一）启动Action服务

（二）开启对话窗口

（三）

Clone this wiki locally

rasa中文对话系统构建

【关于 rasa中文对话系统构建】那些你不知道的事

目录

一、安装 Rasa 内容

1.1 安装 中文版本 的 Rasa-nlu

1.2 安装 rasa

安装过程中遇到的问题

1.3 sklearn 和 MITIE 库 安装

1.4 安装 rasa_core

二、新建项目

2.1 安装rasa，自动生成rasa项目所需文件

2.2 测试助手

三、Rasa使用–构建简单聊天机器人

3.1 Rasa工作原理

3.2 构建NLU样本

3.2.1 data/nlu.md

3.3 构建Core样本

3.3.1 data/stories.md

3.3.2 domain.yml

3.4 训练NLU和CORE模型

3.4.1 config.yml

3.4.2 模型训练

四、配置Http和Action

4.1 credentials.yml

4.2 endpoints.yml

4.3 action.py

五、启动对话

5.1 通过 post 请求 查询

（1）启动Rasa服务

（2）启动Action服务

（3） 效果演示

5.2 聊天窗 式 对话

（一）启动Action服务

（二） 开启对话窗口

（三）

Clone this wiki locally

1.1 安装中文版本的 Rasa-nlu

1.3 sklearn 和 MITIE 库安装

5.1 通过 post 请求查询

（3）效果演示

5.2 聊天窗式对话

（二）开启对话窗口