Samadhaan

Introduction

Samadhaan is an AI/ML-driven grievance management system designed to facilitate the auto-categorization of grievance reports and assist citizens with common queries via Samadhaan App. This system aims to streamline the grievance handling process by providing an efficient way to categorize, track, and respond to grievances.

Problem Statements

• Problem Statement 1: Develop an AI/ML-driven system for topic clustering/modelling to enable auto-categorisation of received grievance reports.
• Problem Statement 2: Develop an AI/ML-driven Chatbot to assist citizens with filing grievances on the CPGRAMS portal.

Workflow Overview

1. Samadhaan App: Users sign in, enter grievances in their preferred language, and post them.
2. Samadhaan Server: Processes grievances, utilizing AI models for department prediction, severity assessment, and summarization.
3. Samadhaan E-portal: AI-enhanced grievances are made available to the respective departments for action.
4. Samadhaan App: Users receive notifications upon responses to their grievances.

REST API Endpoints

• Complaint Management: Endpoints for retrieving, creating, and updating complaints.
• User & Department Management: Endpoints for user and department details retrieval and creation.

REST API Endpoints Documentation

1. Get All Complaints

• Endpoint: /Complaints
• Method: GET
• Description: Retrieves a list of all complaints in the system.
• Request Parameters: None
• Response: Returns a JSON array of complaints.

2. Get Complaint by ID

• Endpoint: /Complaints/<path:id>
• Method: GET
• Description: Fetches a specific complaint by its unique ID.
• Request Parameters: id - The unique identifier of the complaint.
• Response: Returns a JSON object of the complaint details.

3. Get Complaints by User Code

• Endpoint: /Complaints/User/<user_code>
• Method: GET
• Description: Retrieves all complaints associated with a specific user, identified by a unique user code.
• Request Parameters: user_code - The unique code of the user.
• Response: Returns a JSON array of complaints linked to the specified user.

4. Get All Departments

• Endpoint: /Departments
• Method: GET
• Description: Fetches a list of all departments registered in the system.
• Request Parameters: None
• Response: Returns a JSON array of departments.

5. Get All Users

• Endpoint: /Users
• Method: GET
• Description: Retrieves a list of all users in the system.
• Request Parameters: None
• Response: Returns a JSON array of users.

6. Create a New User

• Endpoint: /Users
• Method: POST
• Description: Registers a new user in the system.
• Request Parameters: JSON object containing user details (e.g., name, email, user_code).
• Response: Returns a confirmation of the user creation with the user's details.

7. Get Department by ID

• Endpoint: /Departments/<int:id>
• Method: GET
• Description: Fetches details of a specific department by its unique ID.
• Request Parameters: id - The unique identifier of the department.
• Response: Returns a JSON object of the department's details.

8. Get User by ID

• Endpoint: /Users/<id>
• Method: GET
• Description: Retrieves details of a specific user by their unique ID.
• Request Parameters: id - The unique identifier of the user.
• Response: Returns a JSON object of the user's details.

9. Create or Update a Complaint

• Endpoint: /Complaints
• Method: POST, PUT
• Description: Allows for the creation of a new complaint or updating an existing one. Use POST for creating and PUT for updating.
• Request Parameters: JSON object containing complaint details. For PUT, include the complaint ID to update.
• Response: Returns a confirmation of the complaint creation or update with the complaint's details.

Notes:

• All GET requests might support query parameters for filtering, sorting, or pagination, which are not specified here.
• The POST and PUT methods expect requests to be made with a content type of application/json and the body of the request to contain a valid JSON object according to the endpoint's requirements.
• For the POST method on /Users and /Complaints, the server returns the created object with an assigned unique ID.
• Authentication and authorization mechanisms are not covered in this documentation but are essential for protecting sensitive endpoints and data.

Software Used

• App: Flutter
• Server: Python
• E-Portal: React Typescript and React Admin

Installation Instructions

App

• Prerequisites: Install Flutter and Dart

Server

• Prerequisites: Install Conda, Python, and Langchain

E-portal

• Prerequisites: Install React Typescript, React Admin, Node.js, and npm

Running the Application

App

• Launch Simulator (iOS and Android) and run the application.

E-Portal

• Execute npm run dev to start the E-Portal.

Server

• For Windows: Use start_samadhaan_server.bat script.
• For Mac: Activate conda environment and run samadhaan-server.py and grievance_enhancer.py.
• Models: a) Hugging Face Zephr and Samadhaan Fine Trained models are too large to be committed in github. Create models folder under server and download from hugging face (TheBloke/zephyr-7B-beta-GGUF) and save in models/hf-zephyr. b) Department Prediction Model: Send me an email to upload the copy of this model. Save in models/predict_dept folder

Contributing

We welcome contributions! Please read our contributing guide to learn how you can help improve Samadhaan.

License

MIT