Smart Document Processing System

System that processes real-world business documents (invoices and purchase orders), extracts structured data, validates it, and presents it through a simple interface.

Features

• Document ingestion from PDF, CSV, TXT files
• Data extraction for key fields
• Validation engine with issue detection
• Review and approval workflow
• REST API with Swagger documentation
• Docker support
• Database migrations with go-migrate

Architecture

Backend

Follows Clean Architecture with a hexagonal architecture design, separating concerns into two layers:

• Domain — core business logic, models, and validation rules with no external dependencies
• Adapters — all external concerns including HTTP handlers, PostgreSQL persistence, file processing, and database migrations

Frontend

Built with a custom React CLI (@aditokmo/create-react-project) that builds a feature-based, pre-configured architecture. Development was done by a GitHub Copilot local agent configured with a custom ruleset to enforce consistency across the codebase.

Improvements I would make

• Refining the UI and cleaning up the React codebase, as the initial version was built with a free GitHub Copilot agent
• Adding rate limiter
• Adding support for image formats (that I will probably do in couple days)
• Supporting multi-file upload once the UX flow is properly planned out
• Continued testing to surface and fix any remaining bugs in the system if they exist
• Adding upload validation to prevent confusion around submitting multiple files before multi-upload is fully supported

AI Usage

• Building the frontend by combining my custom React CLI, which builds the full folder structure with pre-configured libraries, and GitHub Copilot Agents to generate code following that structure and the rules I defined through prompts
• Generating Swagger documentation and README docs
• Reviewing backend logic to catch anything I may have missed, and for occasional syntax guidance since I am still learning Go
• Setting up Docker and Makefile configuration to make it easier for others to run the project locally
• Generating Regex for data validation

API Endpoints

• POST /documents - Upload and process a document
• GET /documents - List all documents
• GET /documents/{id} - Get document details
• PUT /documents/{id} - Update document
• POST /documents/{id}/approve - Approve document
• POST /documents/{id}/reject - Reject document

Prerequisites

• Go 1.25.5 or higher
• Node.js 18.0 or higher
• pnpm 8.0 or higher
• Docker

Quick Start

# Clone repo
git clone https://github.com/aditokmo/smart-document-processing-system.git

# Enter project folder
cd smart-document-processing-system

# Install dependencies
make install

# Start all services (database, backend, frontend)
make dev

Then open:

• Frontend: http://localhost:5173
• Backend API: http://localhost:8080
• Swagger Docs: http://localhost:8080/swagger

Make Commands

• make install - Install dependencies
• make dev - Start all services in development mode
• make start - Start all services in production mode
• make stop - Stop all services
• make restart - Restart all services
• make clean - Stop and clean up