How luislunaec/devradar works
Project type and purpose
• DevRadar is a full-stack web application with an auxiliary data-collection microservice.
• Purpose: Analyze the Ecuadorian IT job market by scraping job sites, storing listings, and exposing dashboards and AI-powered features (CV analysis, comparison, statistics) via a web UI.
File structure
• Top-level directories:
• backend: FastAPI service implementing REST endpoints, AI logic, database access.
• frontend: Vite + React + TypeScript + Tailwind UI for dashboards and pages.
• scraper: Standalone Python microservice that scrapes job portals and populates the database.
• .github/workflows: CI/CD (deploy.yml).
• docker-compose.yml: Orchestrates containers (frontend, backend, scraper, DB).
• Patterns:
• backend/app/routers ↔ HTTP API layer
• backend/app/services ↔ Business logic layer
• backend/app/llm & embeddings ↔ External AI integration
• scraper/db ↔ Data-access for Supabase/PostgreSQL
• Configuration & deployment:
• Dockerfiles in backend, frontend, scraper
• docs/DEPLOY.md
README
• Single-line description: “Análisis del mercado laboral ecuatoriano, enfocado en las habilidades más solicitadas de las TI.”
Putting it all together
• Connect:
• Browser → Frontend → Backend → Database
• Backend → External LLM API
• Scraper → Database
• Optionally, show GitHub Actions (under .github/workflows) triggering deploys or scraper runs.