from flask import Flask, render_template, request
from dialogflow import get_suggestions

app = Flask(__name__)

@app.route('/ask_question',methods = ['POST', 'GET'])
def ask_question():
    if request.method == 'POST':
        # Get the issue description from the form data
        description = request.form.get('description')
        # ask dialogflow for suggestions
        articles = get_suggestions(description)
        # render the articles page
        return render_template('articles.html', articles=articles, description=description)

    return render_template('ask_question.html')

# ... before __main__

import os
import random
import dialogflow_v2 as dialogflow

# path to the Google API credentials file
os.environ["GOOGLE_APPLICATION_CREDENTIALS"]="google cloud credentials file"

# Google project ID from the agent settings page
PROJECT_ID = "google project id"
# the language your agent is trained on
LANGUAGE_CODE = "en-US"

session_client = dialogflow.SessionsClient()

def get_suggestions(text, session_id=None):
    # if no session, create a random session id (default behaviour)
    if session_id is None:
        session_id = str(random.randint(0, 10000))

    # get the dialogflow session
    session = session_client.session_path(PROJECT_ID, session_id)
    # create a new input using the text description
    text_input = dialogflow.types.TextInput(text=text, language_code=LANGUAGE_CODE)
    query_input = dialogflow.types.QueryInput(text=text_input)
    # call the Dialogflow API to ask for suggested articles given text in input
    response = session_client.detect_intent(session=session, query_input=query_input)
    result = response.query_result
    # if the matching intent is fallback, no suggested articles were found
    if result.intent.is_fallback or len(result.fulfillment_messages) == 0:
        return None
    # return the list of suggested articles as a list of dict
    articles = []
    for msg in result.fulfillment_messages:
        fields = msg.payload.fields
        articles.append({"url": fields["url"].string_value,
                    "title": fields["title"].string_value,
                    "confidence": result.intent_detection_confidence})

    return articles if len(articles)>0 else None

• Create a new Dialogflow session. A session is used to track the conversation with a user over multiple dialog turns. In this example you don’t need to track turns, therefore you will create a new session for each API call (session_id=None)
• Call the API to detect the intent for the given input.
• If the Fallback intent is returned, the agent didn’t recognized the customer description of the problem, so return None
• If at least one non-fallback intent is returned, add all fulfillment messages to the list to be returned. Here messages are the json objects you have added in the first step of this tutorial as an intent response.