Python Project With Source Code AI Mental Health Therapist Chatbot

Python Project With Source Code AI Mental Health Therapist Chatbot

AI Mental Health Therapist Chatbot involves using natural language processing (NLP) to understand and respond to users' emotional and mental health concerns.

AI Mental Health Therapist Chatbot

Dependencies:

• spaCy: NLP library for natural language understanding.
• textblob: Library for processing textual data.
• flask: Web framework for creating a web application.
• flask-ngrok: Used to expose a Flask web server to the internet for testing.

Installation:

You can install the required libraries using pip:

pip install spacy textblob flask flask-ngrok

Source Code:

from flask import Flask, request, jsonify
from textblob import TextBlob
import spacy
import random

app = Flask(__name__)

# Load English language model for spaCy
nlp = spacy.load("en_core_web_sm")

# Responses for different emotional states
responses = {
    "greeting": ["Hello! How are you feeling today?", "Hi there! How can I support you today?", "Hello, what's on your mind?"],
    "good": ["I'm glad to hear that!", "That's great to hear!", "Excellent!"],
    "bad": ["I'm sorry to hear that. Can you tell me more?", "It's okay to feel like this. What's been bothering you?", "I'm here to listen. What's been going on?"],
    "farewell": ["Take care of yourself. Goodbye!", "I'm here if you need to talk. Goodbye!", "Remember, I'm just a message away. Goodbye!"]
}

# Define endpoint for receiving messages
@app.route("/chat", methods=["POST"])
def chat():
    data = request.json
    message = data["message"]

    # Perform sentiment analysis with TextBlob
    analysis = TextBlob(message)
    if analysis.polarity > 0.2:
        mood = "good"
    elif analysis.polarity < -0.2:
        mood = "bad"
    else:
        mood = "neutral"

    # Determine response based on mood
    if "how are you" in message.lower() or "hello" in message.lower() or "hi" in message.lower():
        response = random.choice(responses["greeting"])
    elif mood == "good":
        response = random.choice(responses["good"])
    elif mood == "bad":
        response = random.choice(responses["bad"])
    elif "bye" in message.lower() or "goodbye" in message.lower():
        response = random.choice(responses["farewell"])
    else:
        response = "Tell me more about that."

    # Return response in JSON format
    return jsonify({"response": response})

if __name__ == "__main__":
    app.run(debug=True)

How it Works:

1. Dependencies: Imports necessary libraries (flask, textblob, spacy).
2. SpaCy Model: Loads the spaCy English model (en_core_web_sm) for natural language processing.
3. Responses: Defines a dictionary responses with predefined responses for different emotional states (greeting, good, bad, farewell).
4. Flask Web Server: Starts a Flask web server with a /chat endpoint to receive POST requests containing user messages.
5. Sentiment Analysis: Uses TextBlob to perform sentiment analysis on the user message and categorizes the sentiment as good, bad, or neutral.
6. Response Generation: Generates a response based on the detected sentiment or specific keywords in the user message.
7. JSON Response: Returns a JSON object containing the generated response.

How to Run:

• Save the code to a file (mental_health_chatbot.py).
• Run the Flask web server: python mental_health_chatbot.py.
• Use tools like Postman or send HTTP POST requests to http://localhost:5000/chat with JSON data {"message": "Hello, I'm feeling stressed."} to interact with the chatbot.

Notes:

• Expandability: This example uses static responses. For a more advanced application, integrate with mental health resources or therapy techniques.
• Security: Ensure proper validation and sanitization of user inputs, especially when dealing with sensitive topics like mental health.
• Deployment: For production deployment, consider hosting on a secure server and using HTTPS.

This basic AI Mental Health Therapist Chatbot demonstrates how to use NLP and sentiment analysis to provide supportive responses to users expressing emotional concerns.