Thrive Backend Documentation

This repository contains the backend code for Thrive, a platform that utilizes audio analysis to empower individuals to proactively manage their mental well-being.

Project Structure

• auth:
  • authService.js: Handles user authentication, signup, login, password reset, and other user management functionalities.
• server:
  • firebase.js: Initializes Firebase Admin SDK for database interaction.
• services:
  • audioHandler.js: Processes audio files, performs analysis using a machine learning model, and stores results in Firestore.
  • historyUser.js: Fetches analysis history for a user.
• routes:
  • routes.js: Defines API routes for authentication, user management, and audio analysis.
• app.py:
  • app.py: Flask API endpoint for audio analysis (ML model deployment).

Technologies Used

• Hapi.js: Web framework for building the API server.
• Firebase: Provides a NoSQL database (Firestore) and authentication services.
• Nodemailer: Enables sending emails for password reset.
• Flask: Used to create the API endpoint for the ML model.
• Librosa: Python library for audio analysis and feature extraction.
• TensorFlow Lite: Machine learning model for depression detection.

Key Features

• User Authentication: Secure signup, login, password reset, and logout functionalities.
• Audio Analysis: Uses a machine learning model to analyze audio recordings and predict potential depression.
• Personalized Suggestions: Provides tailored suggestions for self-care and potential resources based on the analysis.
• History Tracking: Allows users to track their past analysis results and progress.

GCP Architecture

Installation and Setup

1. Clone the repository:

   git clone https://github.com/Thrive-Depression-Speech-Analytics/thrive-backend.git

2. Install dependencies:

   npm install

3. Set up environment variables: Create a .env file and set the following environment variables:
   • FIREBASE_PROJECT_ID
   • FIREBASE_CLIENT_EMAIL
   • FIREBASE_PRIVATE_KEY
   • JWT_SECRET
4. Configure Firebase:
   • Create a Firebase project.
   • Enable Firestore and Authentication in your Firebase project.
   • Download a service account key file and place it in the server directory.
5. Configure Google Cloud Storage:
   • Create a Google Cloud Storage bucket for storing audio files.
   • Ensure your service account has permissions to access this bucket.
6. Start the server:

   npm start

Running the App.py for ML Model

1. Make sure you have TensorFlow Lite and librosa installed:

   pip install tensorflow-lite
   pip install librosa

2. Run the application:

   flask run 

API Endpoint Documentation

Authentication

POST /auth/register

• Description: Registers a new user.
• Request Body:

  {
      "username": "yourusername",
      "password": "yourpassword",
      "email": "your_email@example.com"
  }

• Response:

  {
      "message": "Signup successful",
      "userId": "user_id"
  }

• Status Codes:
  • 201: Successful signup.
  • 400: Invalid request body.
  • 409: Email already exists.
  • 500: Internal server error.

POST /auth/login

• Description: Logs in an existing user.
• Request Body:

  {
      "email": "your_email@example.com",
      "password": "yourpassword"
  }

• Response:

  {
      "message": "Login successful",
      "token": "jwt_token",
      "userId": "user_id"
  }

• Status Codes:
  • 200: Successful login.
  • 400: Invalid request body.
  • 401: Invalid password.
  • 404: Email not found.
  • 500: Internal server error.

POST /auth/forgot-password

• Description: Sends an OTP for password reset.
• Request Body:

  {
      "email": "your_email@example.com"
  }

• Response:

  {
      "message": "OTP sent successfully"
  }

• Status Codes:
  • 200: OTP sent successfully.
  • 400: Invalid request body.
  • 404: Email not found.
  • 500: Internal server error.

POST /auth/reset-password

• Description: Resets the user's password using an OTP.
• Request Body:

  {
      "email": "your_email@example.com",
      "otp": "your_otp",
      "newPassword": "your_new_password"
  }

• Response:

  {
      "message": "Password successfully reset"
  }

• Status Codes:
  • 200: Password reset successfully.
  • 400: Invalid request body.
  • 401: Invalid OTP.
  • 404: User not found.
  • 500: Internal server error.

POST /auth/logout

• Description: Logs out the current user.
• Request Body: None.
• Response:

  {
      "message": "Logged out successfully"
  }

• Status Codes:
  • 200: Successful logout.
  • 500: Internal server error.

User Management

PUT /users/{userId}/username

• Description: Updates the user's username.
• Request Body:

  {
      "newUsername": "your_new_username"
  }

• Response:

  {
      "message": "Username successfully changed"
  }

• Status Codes:
  • 200: Username updated successfully.
  • 400: Invalid request body.
  • 401: Unauthorized.
  • 404: User not found.
  • 500: Internal server error.

PUT /users/{userId}/password

• Description: Updates the user's password.
• Request Body:

  {
      "oldPassword": "your_old_password",
      "newPassword": "your_new_password"
  }

• Response:

  {
      "message": "Password successfully changed"
  }

• Status Codes:
  • 200: Password updated successfully.
  • 400: Invalid request body.
  • 401: Unauthorized.
  • 404: User not found.
  • 500: Internal server error.

GET /users/history

• Description: Retrieves the user's analysis history.
• Request Body: None.
• Response:

  [
      {
          "id": "analysis_id",
          "timestamp": "timestamp_in_firestore_format",
          "depressionStatus": "Potentially Depressed" | "Normal",
          "suggestions": "suggestion_text",
          "audioFileName": "audio_file_name.wav",
          "gcsUrl": "gcs_url"
      },
      // ... other history items
  ]

• Status Codes:
  • 200: History retrieved successfully.
  • 401: Unauthorized.
  • 500: Internal server error.

Audio Analysis

POST /audio/analyze

• Description: Analyzes an audio file to predict depression status.
• Request Body:

  {
      "audio": {
          "filename": "audio_file.wav",
          "_data": {
              // Stream data
          }
      }
  }
  

• Response:

  {
      "isDepressed": "depresi" | "normal",
      "confidence": "prediction_value",
      "analysisId": "analysis_id",
      "gcsUrl": "gcs_url",
      "suggestion": "random_suggestion_text"
  }

• Status Codes:
  • 200: Analysis successful.
  • 400: Invalid request body.
  • 401: Unauthorized.
  • 500: Internal server error.

Contributors

Name	Bangkit ID	Learning Path	University	Role
Muhammad Ridzki Nugraha	C193D4KY1133	Cloud Computing	Universitas Bina Sarana Informatika	Backend Engineer, DevOps, Cloud Engineer
Hari Handika Saputra	C193D4KY1176	Cloud Computing	Universitas Bina Sarana Informatika	Backend Engineer, DevOps, Cloud Engineer