About NASA Space Biology Knowledge Engine
An AI-powered search and analysis platform for NASA space biology research papers with advanced vector-based retrieval and on-demand content extraction.
This application connects to a FastAPI backend that provides comprehensive search and database management capabilities. The backend uses ChromaDB for vector storage and embedding-based search, with support for both pre-loaded and on-demand content extraction.
Papers are scraped from NASA's PubMed Central database, processed into chunks, and stored as embeddings for semantic search. The system supports real-time content extraction with image parsing and AI-powered answer generation.
Database Status
Papers currently available in the database.
Platform Features
Semantic Search
Vector-based search through research papers with AI-generated answers.
On-Demand Scraping
Real-time paper content extraction with image support.
Vector Database
ChromaDB-powered semantic similarity search.
AI Models
Support for various Google Gemini AI models.
Database Management
Complete CRUD operations for papers and embeddings.
Image Extraction
Automatic extraction of figures and images from papers.
Search Methods
Similarity Search
Vector-based semantic similarity matching using embeddings.
MMR Search
Maximum Marginal Relevance for diverse, non-redundant results.
Keyword Filtering
Filter results by specific keywords found in paper titles.
Configurable Results
Adjust number of results from 1-50 papers per search.
Search Operations
POST /searchSemantic search with optional AI answersPOST /search/on-demandReal-time scraping with imagesGET /papersList loaded papers
Database Management
POST /database/load-csvLoad papers metadataPOST /load-papersScrape and embed papersGET /database/statsDatabase statisticsPOST /reset-databaseReset all data