Legal Metrology Compliance System

An AI-assisted compliance validation system for checking packaged commodities against the requirements of the Legal Metrology (Packaged Commodities) Rules, 2011.

The system allows users to upload package images, process them through a backend validation pipeline, evaluate applicable compliance rules, review validation results, and generate inspection reports.

Features

• User authentication using JWT
• Create and manage inspection sessions
• Upload packaged-commodity images
• Image-based validation
• Rule-based compliance evaluation
• Compliance score and overall status
• Field-level validation results
• Processed images with validation annotations
• Inspection result review
• Automated inspection report generation
• PDF report download
• PostgreSQL database
• REST API based frontend-backend integration

System Architecture

User
  │
  ▼
React Frontend
  │
  │ REST API + JWT
  ▼
FastAPI Backend
  │
  ├── Authentication
  ├── Validation API
  ├── Image API
  ├── Review API
  └── Report API
          │
          ▼
   Validation Pipeline
          │
          ├── Image Analysis
          ├── Rule Engine
          └── Processed Images
          │
          ▼
      PostgreSQL
          │
          ▼
    PDF Report

Project Structure

project-root/
│
├── frontend/
│   ├── src/
│   │   ├── components/
│   │   ├── pages/
│   │   ├── services/
│   │   ├── context/
│   │   └── ...
│   ├── .env
│   ├── package.json
│   └── README.md
│
├── backend/
│   ├── api/
│   │   └── v1/
│   │       ├── auth.py
│   │       ├── images.py
│   │       ├── validation.py
│   │       ├── review.py
│   │       └── reports.py
│   │
│   ├── core/
│   ├── db/
│   ├── models/
│   ├── services/
│   │   └── validation_pipeline.py
│   ├── ruleEngine/
│   ├── uploads/
│   ├── main.py
│   ├── requirements.txt
│   └── README.md
│
└── README.md

Application Workflow

1. Authentication

The user logs into the application through the frontend.

Login
  │
  ▼
POST /api/auth/login
  │
  ▼
JWT Access Token
  │
  ▼
GET /api/auth/me
  │
  ▼
Authenticated User

The JWT token is automatically attached to protected API requests.

2. Create Inspection

When a new inspection is started, the frontend creates a validation session.

POST /api/validation/

The backend creates a validation record with an initial status of PENDING.

The returned validation_id identifies the inspection throughout the workflow.

3. Upload Package Image

The selected package image is uploaded using:

POST /api/validation/{validation_id}/images

Supported image formats:

• JPEG
• PNG
• WebP

4. Process Validation

The frontend starts the validation pipeline using:

POST /api/validation/{validation_id}/process

The pipeline produces validation results, compliance status, validation score, processed images, and rule-level explanations.

5. Review Results

The frontend displays:

• Overall validation status
• Compliance score
• Product information
• Package information
• Individual validation results
• Applicable rules
• Original images
• Processed images

6. Generate Report

An inspection report can be generated using:

POST /api/validation/{validation_id}/report

The generated report can be retrieved using:

GET /api/validation/{validation_id}/report

Backend API

Authentication

Method	Endpoint	Description
POST	/api/auth/login	Authenticate user
GET	/api/auth/me	Get current authenticated user

Validation

Method	Endpoint	Description
POST	/api/validation/	Create validation
GET	/api/validation/	List validations
GET	/api/validation/{id}	Get validation
PUT	/api/validation/{id}	Update validation
DELETE	/api/validation/{id}	Delete validation
POST	/api/validation/{id}/process	Run validation pipeline

Images

Method	Endpoint	Description
POST	/api/validation/{id}/images	Upload image
GET	/api/validation/{id}/images	List images
GET	/api/validation/{id}/images/{image_id}	Get image
PUT	/api/validation/{id}/images/{image_id}	Update image
DELETE	/api/validation/{id}/images/{image_id}	Delete image

Reports

Method	Endpoint	Description
GET	/api/validation/{id}/report-data	Get report data
POST	/api/validation/{id}/report	Generate report
GET	/api/validation/{id}/report	Download PDF report

Review

The review API allows validation results to be reviewed and decisions/comments to be submitted for individual validation results.

Technology Stack

Frontend

• React
• Vite
• Tailwind CSS
• Axios
• React Router
• Lucide React

Backend

• Python
• FastAPI
• SQLAlchemy
• PostgreSQL
• JWT Authentication
• OpenCV

Validation

• Image analysis
• Rule-based compliance engine
• Legal Metrology compliance rules
• Processed image generation

Local Development

Prerequisites

• Node.js
• npm
• Python 3
• PostgreSQL

Backend Setup

cd backend
python -m venv venv
venv\Scripts\activate
pip install -r requirements.txt
uvicorn main:app --reload

Backend: http://localhost:8000

FastAPI documentation: http://localhost:8000/docs

Frontend Setup

cd frontend
npm install
npm run dev

Create .env:

VITE_API_BASE_URL=http://localhost:8000
VITE_USE_MOCK_API=false

Frontend: http://localhost:5173

Frontend → Backend Integration

The frontend communicates with the backend through REST APIs using Axios.

Authenticated requests include:

Authorization: Bearer <access_token>

Main frontend services include:

• api.js
• authService.js
• inspectionService.js

Inspection Data Flow

Create Validation
       │
       ▼
Upload Image
       │
       ▼
Start Processing
       │
       ▼
Validation Pipeline
       │
       ├── Image Analysis
       ├── Rule Evaluation
       └── Processed Image Generation
       │
       ▼
Validation Results
       │
       ▼
Inspection Result
       │
       ▼
Report Data
       │
       ▼
Generate PDF

Image Handling

Uploaded images are stored according to their validation session and exposed through /uploads/.

Typical structure:

uploads/
└── <validation_id>/
    └── <generated_image_name>.<extension>

Compliance Results

Each validation produces rule-level results containing information such as:

• Rule identifier
• Field being validated
• Validation status
• Reason or explanation
• Evidence information

The inspection also produces an overall compliance status and score.

Possible validation states include:

• PASS
• FAIL
• REVIEW
• PENDING

Reports

The report data contains information about:

• Inspection
• Product
• Package
• Images
• Validation results
• Applicable rules
• Inspection status
• Compliance score

The backend generates the final PDF report, which can be downloaded from the frontend.

Security

The application uses authenticated API requests. Protected backend resources verify the authenticated user before allowing access to validation and inspection data.

The frontend stores the JWT access token and automatically attaches it to API requests through an Axios interceptor.

Project Purpose

The system is designed to assist with the inspection of packaged commodities by combining image-based analysis, rule-based validation, persistent inspection records, and automated reporting.

The complete workflow is:

Image Upload → Validation → Review → Compliance Report