How kenken64/dynaform works
Project type and purpose
• This is a containerized, full-stack, microservices application for converting PDF forms to JSON, analyzing form fields with an LLM, and managing form data.
• Users upload forms via an Angular frontend; a Node.js/Express API orchestrates PDF-to-PNG conversion, LLM-based field extraction, and persistent storage in MongoDB.
File structure
• Top-level “dynaform” folder: Angular (TypeScript) app + nginx config → the browser UI.
• “describeImge” folder: Node.js/Express API (src/ → dist/), with Controllers/Services/Middleware/Routes → MVC pattern.
• “pdf-png” folder: Python Flask service to convert PDFs to PNG.
• “ollama” folder: Docker image hosting the LLM (qwen2.5vl) for multimodal inference.
• “mongodb” folder: Dockerfile, init-scripts, configs for MongoDB; security docs and scripts for role-based access.
• Dozens of start-and-setup shell/PowerShell scripts plus multiple docker-compose.*.yml files to orchestrate dev, prod, secure, macOS, debug stacks.
README insights
• Orchestrates five main containers via Docker Compose:
• Frontend (Angular + Nginx) at port 4201
• API (Node/Express) at port 3000
• PDF Converter (Flask) at port 5001
• LLM service (Ollama) at port 11434
• MongoDB at port 27017
• Emphasizes on-the-fly rebuilds, secure production setup (Docker secrets), and cross-platform scripts.
Elements
• Rectangle with folded corner for Frontend
• Rounded rectangles for microservices (API, Flask, Ollama)
• Blue for UI components
• Green for business logic/microservices
• Orange for data stores
• Annotations for build/orchestration scripts (e.g., docker-compose .yml files, setup scripts)
• Notes on secret management (Docker secrets, .env files)