采用对话框流意图和查询修复

Question

我的聊天机器人是在对话框中创建的，现在我正在尝试从Python中访问它，以便在GUI中获取用户输入和显示输出(假设是一个基本的chatbot gui)。

我已经将我的Python环境连接到了对话框流和火药库，

下面是检测意图的代码；

#Detection of Dialogflow intents, project and input.
def detect_intent_texts(project_id, session_id, texts, language_code):
   #Returns the result of detect intent with texts as inputs, later can implement same `session_id` between requests allows continuation of the conversaion.
    import dialogflow_v2 as dialogflow
    session_client = dialogflow.SessionsClient()

    session = session_client.session_path(project_id, session_id)
#To ensure session path is correct - print('Session path: {}\n'.format(session))

    for text in texts:
        text_input = dialogflow.types.TextInput(text=text, language_code=language_code)
        query_input = dialogflow.types.QueryInput(text=text_input)
        response = session_client.detect_intent(session=session, query_input=query_input)

        print ('Chatbot:{}\n'.format(response.query_result.fulfillment_text))
detect_intent_texts("chat-8","abc",["Hey"],"en-us")

我需要在某种程度上说，如果这个意图被触发，从db中获取一些内容并显示给用户。

更新

我现在的代码已经全部完成了，这一切在我看来都是正确的，但它却抛出了一个我不明白的错误，这要感谢Sid8491到目前为止的帮助。

简而言之，我的问题是，我以前的代码允许我输入某些内容，聊天机器人响应，所有这些都在控制台中，但它有效.新代码应该允许我说“当这个意图被触发时，就这样做”。

import os, json
import sys
import dialogflow
from dialogflow_v2beta1 import *
import firebase_admin
import requests.packages.urllib3
requests.packages.urllib3.disable_warnings()
from firebase_admin import firestore
from firebase_admin import credentials
import requests.packages.urllib3
from Tkinter import *
from dialogflow_v2beta1 import agents_client
import Tkinter as tk
result = None
window = Tk()

def Response():
    #no need to use global here
    result = myText.get()
    displayText.configure(state='normal')
    displayText.insert(END, "User:"+ result + '\n')
    displayText.configure(state='disabled')

#Creating the GUI
myText = tk.StringVar()
window.resizable(False, False)
window.title("Chatbot")
window.geometry('400x400')
User_Input = tk.Entry(window, textvariable=myText, width=50).place(x=20, y=350)
subButton = tk.Button(window, text="Send", command=Response).place(x =350, y=350)
displayText = Text(window, height=20, width=40)
displayText.pack()
scroll = Scrollbar(window, command=displayText).pack(side=RIGHT)
window.mainloop()

#Initialize the firebase admin SDK
cred = credentials.Certificate('./file.json')
default_app = firebase_admin.initialize_app(cred)
db = firestore.client()

def getCourse():
    doc_ref = db.collection(u"Course_Information").document(u"CourseTypes")
    try:
        doc = doc_ref.get()
        return 'Document data: {}'.format(doc.to_dict())
    except google.cloud.exceptions.NotFound:
        return 'Not found'

def detect_intent_text(project_id, session_id, text, language_code):
    GOOGLE_APPLICATION_CREDENTIALS=".chat-8.json"
    session_client = dialogflow.SessionsClient(GOOGLE_APPLICATION_CREDENTIALS)

    session = session_client.session_path(project_id, session_id)

    text_input = dialogflow.types.TextInput(
        text=text, language_code=language_code)

    query_input = dialogflow.types.QueryInput(text=text_input)

    response = session_client.detect_intent(
        session=session, query_input=query_input)

queryText = [myText.get()]

res = detect_intent_text('chat-8', 'session-test', queryText, 'en')
intentName = res['query_result']['intent']['display_name']

if intentName == 'CourseEnquiry':
    reponse = getCourse()
    print json.dumps({
        'fulfillmentText': reponse,
    })
elif intentName == 'Greetings':
    print "Yo"

detect_intent_texts("chat-8","abc", queryText,"en-us")

但我知道这个错误：

C:\Users\chat\PycharmProjects\Chatbot\venv\Scripts\python.exe C:/Users/chat/PycharmProjects/Chatbot/venv/Chatbot.py
Traceback (most recent call last):
  File "C:/Users/chat/PycharmProjects/Chatbot/venv/Chatbot.py", line 65, in <module>
    res = detect_intent_text('chat-8', 'session-test', queryText, 'en')
  File "C:/Users/chat/PycharmProjects/Chatbot/venv/Chatbot.py", line 51, in detect_intent_text
    session_client = dialogflow.SessionsClient(GOOGLE_APPLICATION_CREDENTIALS)
  File "C:\Users\chat\PycharmProjects\Chatbot\venv\lib\site-packages\dialogflow_v2\gapic\sessions_client.py", line 109, in __init__
    self.sessions_stub = (session_pb2.SessionsStub(channel))
  File "C:\Users\chat\PycharmProjects\Chatbot\venv\lib\site-packages\dialogflow_v2\proto\session_pb2.py", line 1248, in __init__
    self.DetectIntent = channel.unary_unary(
AttributeError: 'str' object has no attribute 'unary_unary'

Process finished with exit code 1

Answer 1

是的，我认为你走在正确的轨道上。

您需要从从intentName获得的响应中提取dialogFlow或actionaName，并调用相应的函数，然后将响应发送回用户。

res = detect_intent_texts("chat-8","abc",["Hey"],"en-us")
action = res['queryResult']['action']

if action == 'getSomethingFromDb':
    reponse = someFunction(req)
    return json.dumps({
        'fulfillmentText': reponse,
    })
elif action == 'somethingElse':
    ....

如果您希望使用intentName而不是actionName来实现，那么您可以像下面这样提取intentName

intentName = res['query_result']['intent']['display_name']

编辑1：

例子-

import dialogflow
import os, json

def getCourse():
    doc_ref = db.collection(u"Course_Information").document(u"CourseTypes")
    try:
        doc = doc_ref.get()
        return 'Document data: {}'.format(doc.to_dict())
    except google.cloud.exceptions.NotFound:
        return 'Not found'

def detect_intent_text(project_id, session_id, text, language_code):
    GOOGLE_APPLICATION_CREDENTIALS="C:\\pyth_to_...\\cred.json"
    session_client = dialogflow.SessionsClient(GOOGLE_APPLICATION_CREDENTIALS)

    session = session_client.session_path(project_id, session_id)

    text_input = dialogflow.types.TextInput(
        text=text, language_code=language_code)

    query_input = dialogflow.types.QueryInput(text=text_input)

    response = session_client.detect_intent(
        session=session, query_input=query_input)

queryText = 'get courses of Python' # you will call some function to get text from your app

res = detect_intent_text('project_1234', 'session-test', queryText, 'en')
intentName = res['query_result']['intent']['display_name']

if intentName == 'getCourse':
    reponse = getCourse()
    return json.dumps({
        'fulfillmentText': reponse,
    })

试着上面的例子，并根据你对应用程序的需求进行修改。我的建议是，首先让DialogFlow在没有应用程序的情况下工作，然后将其与应用程序集成。否则，您将无法理解问题是发生在DialogFlow还是您的应用程序。

希望能帮上忙。