How astrabert/praivatesearch works
Project type and purpose
• PrAIvateSearch is a full-stack application combining a Next.js frontend (“chatbot-ui”) and a Python FastAPI backend (“qwen-on-api”) to provide private, local AI-powered web search.
File structure
• Top‐level:
• .env.example, compose.yaml, conda_environment.yaml – environment and orchestration.
• chatbot-ui/ – Next.js frontend, Dockerfile, workspace config, public/static assets, styles, pages (layout.js, page.js).
• qwen-on-api/ – FastAPI app with main.py, utilities (chatHistoryUtils.py, textUtils.py, webSearch.py, qdrantRag.py), compose.yaml, static/.
• docs/, imgs/ – documentation and flowchart.
• Indicators of architecture: frontend/backend separation, containerization (Dockerfiles, compose.yaml), MVC‐style page structure in Next.js, service‐oriented design in Python module decomposition.
README for additional insights
• Confirms two service areas: Next.js UI on port 3000, FastAPI backend on port 8000.
• Describes dependencies: PostgreSQL, Crawl4AI crawler, Qwen model inference via Uvicorn, Docker Compose orchestration.
• Includes a flowchart (imgs/PrAIvateSearch_Flowchart.png) illustrating: user query → crawl & search pipeline → vector retrieval → language model inference → UI response.