释放双眼，带上耳机，听听看~！

本文介绍了如何将Rasa聊天机器人应用到Python Turtle模块中，展示了用户通过语音或文本输入指令控制Python Turtle的功能。同时详细介绍了配置意图、训练数据、标记实体的过程，并配置了回复和动作的相关信息。

背景

为了展示Rasa聊天机器人的使用效果以及如何将Rasa应用到业务系统中(这里将Python Turtle模块作为业务系统)，用户将语音或者文本输入至Rasa，经过处理后调用Python Turtle的功能。

turtle库是Python语言中一个很流行的绘制图像的函数库，想象一个小乌龟，在一个横轴为x、纵轴为y的坐标系原点，(0,0)位置开始，它根据一组函数指令的控制，在这个平面坐标系中移动，从而在它爬行的路径上绘制了图形。

前进功能实现

效果

如何将Rasa聊天机器人应用到Python Turtle模块中

B站完整效果

配置意图、训练数据、标记实体

如何将Rasa聊天机器人应用到Python Turtle模块中

nlu.yml

- intent: forward
  examples: |
    - 前进
    - 前进一下
    - 向前进
    - 向前进一下
    - 前进[50](distance)
    - 前进[30](distance)像素
    - 前进[200](distance)像素
    - 向前走[25.5](distance)
    - 往前移动[39.0](distance)像素
    - 往前挪[39.0](distance)像素
    - [1号](turtle_no)乌龟前进
    - 向前进[二号](turtle_no)乌龟
    - [11号](turtle_no)乌龟前进[200](distance)像素
    - [五号](turtle_no)乌龟向前走[25.5](distance)
    - [9号](turtle_no)乌龟往前移动[39.0](distance)像素
    - 往前挪[39.0](distance)像素[六号](turtle_no)小乌龟
    - 乌龟[1](turtle_no)前进
    - 向前进乌龟[二](turtle_no)
    - 乌龟[11](turtle_no)前进[200](distance)像素
    - 乌龟[五](turtle_no)向前走[25.5](distance)
    - 乌龟[9](turtle_no)往前移动[39.0](distance)像素
    - 往前挪[39.0](distance)像素小乌龟[六](turtle_no)

定义’forward’意图、增加示例数据、标记实体

domain.yml

intents:
- forward

entities:
- distance
- turtle_no

slots:
  s_turtle_no:
    type: text
    mappings:
    - type: from_entity
      entity: turtle_no
  s_distance:
    type: float
    influence_conversation: false
    mappings:
    - type: from_entity
      entity: distance
      
responses:
...省略...

forms:
...省略...

在intents中增加意图’forward’
在entities中增加实体’distance’ ‘turtle_no’
在slots中增加slot(插槽)’s_turtle_no’ ‘s_distance’，并设置类型、mapping

配置回复和动作

domain.yml

intents:
...省略...
entities:
...省略...
slots:
...省略...
responses:
  utter_forwrad_success:
  - text: 前进完成
  utter_forwrad_fail:
  - text: 前进失败，{turtleno}号小乌龟不存在
forms:
...省略...
actions:
- action_forward

在responses中配置文本回复
在actions中配置动作即函数
可以在action_forward函数中直接将文本’前进完成’输出，也可以配置到responses下，然后在action_forward函数中输出utter_forwrad_success或者utter_forwrad_fail，这样有利于复用

编写对话脚本

stories.yml

- story: turtle forward
  steps:
  - intent: forward
  - action: action_forward

将意图(通过用户输入推断出意图)和动作(由bot执行一个函数)连接起来构成单轮对话

action.py函数定义

如何将Rasa聊天机器人应用到Python Turtle模块中

class ActionForward(Action):

    def name(self) -> Text:
        return "action_forward"

    def run(self, dispatcher: CollectingDispatcher,
            tracker: Tracker,
            domain: Dict[Text, Any]) -> List[Dict[Text, Any]]:

        slots = {"s_turtle_no":None,"s_distance":None}
        if len(g_TurtleDict) == 0:
            t = Util.create_new_turtle()
            g_TurtleDict[0] = t
        '''
        1. s_distance s_turtle_no 都没有指定 
        2. s_turtle_no 指定了
        3. s_distance 指定了
        4. s_distance s_turtle_no 都指定了
        '''
        isSuccess = True
        turtle_no = "" # 去掉xx号后的xx
        sturtle_no = tracker.get_slot('s_turtle_no')
        sdistance = tracker.get_slot('s_distance')
        if sturtle_no == None and sdistance == None:
            for t in g_TurtleDict.values():
                t.forward(MOVE_STEP)
            # mainloop()
        elif sturtle_no != None and sdistance == None:
            turtle_no = Util.get_real_turtle_no(sturtle_no)
            if turtle_no in g_TurtleDict.keys():
                g_TurtleDict[turtle_no].forward(MOVE_STEP)
                # mainloop()
            else:
                isSuccess = False
                dispatcher.utter_message(response="utter_forwrad_fail",turtleno=turtle_no)
        elif sturtle_no == None and sdistance != None:
            for t in g_TurtleDict.values():
                t.forward(float(sdistance))
            # mainloop()
        else:
            turtle_no = Util.get_real_turtle_no(sturtle_no)
            if turtle_no in g_TurtleDict.keys():
                g_TurtleDict[turtle_no].forward(float(sdistance))
                # mainloop()
            else:
                isSuccess = False
                dispatcher.utter_message(response="utter_forwrad_fail",turtleno=turtle_no)

        if isSuccess == True:
            dispatcher.utter_message(response="utter_forwrad_success")
        return [SlotSet(slot,value) for slot,value in slots.items()]

获取slot:s_turtle_no和s_distance的值，明确用户指定哪个小乌龟前进以及前进像素数
调用turtle库forward函数执行前进动作

运行turtle demo功能实现

效果

如何将Rasa聊天机器人应用到Python Turtle模块中

配置意图、训练数据、标记实体

nlu.yml

- intent: show_turtledemo
  examples: |
    - 展示小乌龟的demo
    - demo
    - turtle demo
    - turtledemo
    - show demos
    - run demos
    - 展示小乌龟的能力
- intent: select_turtledemo
  examples: |
    - bytedesign
    - chaos
    - clock

定义’show_turtledemo’意图、增加示例数据
定义’select_turtledemo’意图、增加示例数据，该意图用于回复Button格式数据，详见下节

domain.yml

intents:
- select_turtledemo
- show_turtledemo

entities:
...省略...
slots:
...省略...
responses:
...省略...
forms:
...省略...
actions:
...省略...

配置form

domain.yml

intents:
...省略...

entities:
- demo_type
slots:
  s_demo_type:
    type: text
    influence_conversation: true
    mappings:
    - type: from_entity
      entity: demo_type
      conditions:
      - active_loop: turtle_demo_form
responses:
...省略...

forms:
  turtle_demo_form:
    required_slots:
    - s_demo_type
    
actions:
...省略...

首先在domain.yml中forms中配置form名称:turtle_demo_form，需要的slot:s_demo_type
在entities中配置实体demo_type，用于标识想运行的demo
在s_demo_type中配置slots_demo_type，用户输入包含实体demo_type后，将s_demo_type赋值，并且只有在turtle_demo_form激活时才为s_demo_type赋值

配置form要求的slot回复和动作

domain.yml

intents:
...省略...
entities:
...省略...
slots:
...省略...
responses:
utter_ask_turtle_demo_form_s_demo_type:
  - buttons:
    - title: bytedesign
      payload: /select_turtledemo{{"demo_type":"bytedesign"}}
    - title: chaos
      payload: /select_turtledemo{{"demo_type":"chaos"}}
    - title: clock
      payload: /select_turtledemo{{"demo_type":"clock"}}
    - title: colormixer
      payload: /select_turtledemo{{"demo_type":"colormixer"}}
    - title: flyobjs
      payload: /select_turtledemo{{"demo_type":"flyobjs"}}
    - title: forest
      payload: /select_turtledemo{{"demo_type":"forest"}}
    - title: fractalcurves
      payload: /select_turtledemo{{"demo_type":"fractalcurves"}}
    - title: lindenmayer
      payload: /select_turtledemo{{"demo_type":"lindenmayer"}}
    - title: minimal_hanoi
      payload: /select_turtledemo{{"demo_type":"minimal_hanoi"}}
    - title: nim
      payload: /select_turtledemo{{"demo_type":"nim"}}
    - title: paint
      payload: /select_turtledemo{{"demo_type":"paint"}}
    - title: peach
      payload: /select_turtledemo{{"demo_type":"peach"}}
    - title: penrose
      payload: /select_turtledemo{{"demo_type":"penrose"}}
    - title: planet_and_moon
      payload: /select_turtledemo{{"demo_type":"planet_and_moon"}}
    - title: rosette
      payload: /select_turtledemo{{"demo_type":"rosette"}}
    - title: round_dance
      payload: /select_turtledemo{{"demo_type":"round_dance"}}
    - title: sorting_animate
      payload: /select_turtledemo{{"demo_type":"sorting_animate"}}
    - title: trees
      payload: /select_turtledemo{{"demo_type":"trees"}}
    - title: two_canvases
      payload: /select_turtledemo{{"demo_type":"two_canvases"}}
    - title: yinyang
      payload: /select_turtledemo{{"demo_type":"yinyang"}}
    text: 请选择想要运行的demo

forms:
...省略...
actions:
- action_run_demo

在responses中配置回复utter_ask_turtle_demo_form_s_demo_type，规则如下utter_ask_表示要求用户输入文本，turtle_demo_form表示form名称,s_demo_type标识form需要的slot名称
buttons表示回复的格式是button，title为按钮显示文本,payload为上一节定义的意图select_turtledemo，用户点击对应按钮后，会将实体demo_type赋值
在actions中配置action_run_demo，该在函数中调用对应turtle demo

编写对话脚本

stories.yml

- story: turtle demo 
  steps:
  - intent: show_turtledemo       //用户输入问题推断意图为运行turtle demo
  - action: turtle_demo_form      //激活turtle_demo_form
  - active_loop: turtle_demo_form //循环运行form
  - slot_was_set:                 //只有s_demo_type slot被填充后才能跳出循环
    - requested_slot: s_demo_type
  - active_loop: null
  - slot_was_set:
    - requested_slot: null
  - action: action_run_demo       //然后运行函数action_run_demo

详见注释

action.py函数定义

...省略...
from .turtledemos.rosette import *
from .turtledemos.round_dance import *
from .turtledemos.sorting_animate import *
from .turtledemos.tree import *
from .turtledemos.two_canvases import *
from .turtledemos.yinyang import *

class ActionRunDemo(Action):

    def name(self) -> Text:
        return "action_run_demo"

    def run(self, dispatcher: CollectingDispatcher,
            tracker: Tracker,
            domain: Dict[Text, Any]) -> List[Dict[Text, Any]]:
        slots = {
            "s_demo_type":None
        }
        screen = turtle.Screen()
        screen.setup (width=1280, height=720, startx=0, starty=0)
        demo_type = tracker.get_slot('s_demo_type')
        eval(demo_type)()
        mainloop()

        dispatcher.utter_message(text="demo运行完成")
        return [SlotSet(slot,value) for slot,value in slots.items()]

根据slot: s_demo_type值调用对应demo

本文为博主原创文章，未经博主允许不得转载。

本网站的内容主要来自互联网上的各种资源，仅供参考和信息分享之用，不代表本网站拥有相关版权或知识产权。如您认为内容侵犯您的权益，请联系我们，我们将尽快采取行动，包括删除或更正。

{{userData.name}}已认证

如何将Rasa聊天机器人应用到Python Turtle模块中

背景

前进功能实现

效果

配置意图、训练数据、标记实体

nlu.yml

domain.yml

配置回复和动作

domain.yml

编写对话脚本

stories.yml

action.py函数定义

运行turtle demo功能实现

效果

配置意图、训练数据、标记实体

nlu.yml

domain.yml

配置form

domain.yml

配置form要求的slot回复和动作

domain.yml

编写对话脚本

stories.yml

action.py函数定义

AI的发展水平和分类

ChatGPT国内镜像站评测

GeoSpy.ai

Globe Explorer

即梦Dreamina

Luma Dream Machine

Motionshop

Kling AI | Sora-Like Video Model

归档

{{userData.name}}已认证

背景

前进功能实现

效果

配置意图、训练数据、标记实体

nlu.yml

domain.yml

配置回复和动作

domain.yml

编写对话脚本

stories.yml

action.py函数定义

运行turtle demo功能实现

效果

配置意图、训练数据、标记实体

nlu.yml

domain.yml

配置form

domain.yml

配置form要求的slot回复和动作

domain.yml

编写对话脚本

stories.yml

action.py函数定义

AI的发展水平和分类

ChatGPT国内镜像站评测

4个强大的生成AI Python库，助您提升人工智能应用

Pandas AI: 强强联手，让数据分析更快速便捷

Hello world for Manim: 数学可视化和动画的Python库

Rasa实现百度UNIT智能客服教学机器人效果对比与配置