How to create a free or coding style of AI chatbot?

Question

How to create a self-made AI chatbot for a hotel reservation for our system based on our given data from database only in provided area or country that can be accessed by other countries?

Answer 1

If using Microsoft Bot Framework:

1. Create a Bot Service:
   • Go to the Azure Portal.
   • Create a new resource and search for "Bot Services".
   • Set up your bot service with the necessary configurations.
2. Install Bot Framework SDK:
   • Install Bot Framework SDK for your chosen programming language. For Python, you can install it using pip:

          
          pip install botbuilder-core botbuilder-dialogs botbuilder-ai
          
     

Use a library to connect to your Azure SQL Database. For Python, you can use pyodbc or sqlalchemy.

import pyodbc

connection_string = 'DRIVER={ODBC Driver 17 for SQL Server};SERVER=tcp:<your_server>.database.windows.net;DATABASE=<your_database>;UID=<your_username>;PWD=<your_password>'

connection = pyodbc.connect(connection_string)

cursor = connection.cursor()

Create dialog flows to handle different user interactions like searching for hotels, making reservations...

from botbuilder.core import TurnContext, ActivityHandler

from botbuilder.dialogs import DialogSet, WaterfallDialog, WaterfallStepContext

class HotelReservationBot(ActivityHandler):

    def __init__(self, connection):

        self.connection = connection

        self.dialogs = DialogSet()

        self.dialogs.add(WaterfallDialog("main_dialog", [self.start_reservation, self.confirm_reservation]))

    async def on_message_activity(self, turn_context: TurnContext):

        dialog_context = await self.dialogs.create_context(turn_context)

        await dialog_context.continue_dialog()

        if dialog_context.active_dialog is None:

            await dialog_context.begin_dialog("main_dialog")

    async def start_reservation(self, step_context: WaterfallStepContext):

        # Logic to start reservation process

        return await step_context.prompt("prompt_text", "Welcome to Hotel Reservation. Please provide your destination.")

    async def confirm_reservation(self, step_context: WaterfallStepContext):

        # Logic to confirm reservation

        destination = step_context.result

        cursor = self.connection.cursor()

        cursor.execute("SELECT * FROM hotels WHERE location = ?", destination)

        hotels = cursor.fetchall()

        await step_context.context.send_activity(f"Found {len(hotels)} hotels in {destination}.")

        return await step_context.end_dialog()

To understand user inputs better, integrate NLU services like LUIS or Dialogflow.

from botbuilder.ai.luis import LuisApplication, LuisRecognizer

luis_app = LuisApplication("<app_id>", "<endpoint_key>", "https://<region>.api.cognitive.microsoft.com/")

luis_recognizer = LuisRecognizer(luis_app)

async def recognize_intent(self, turn_context: TurnContext):

    luis_result = await luis_recognizer.recognize(turn_context)

    return luis_result.properties["luisResult"]

Deploy your bot to Azure App Service for scalability and accessibility and register it with various channels (for example Web Chat, Skype, Facebook Messenger....).

Links to help you :

• Microsoft Bot Framework Documentation
• Azure SQL Database Documentation
• LUIS Documentation