How abhi-vit/graph-net works
Graph-Net is a Django research tool: a user submits a topic, the result handler gathers search links and Wikipedia context, obtains a page-based AI summary, and renders a topic-centered knowledge graph with source links. Its main internal layers are request routing, retrieval/scraping, summarization, and graph construction. README capabilities are represented, but the excerpts do not show how the graph data consumed by the result handler is produced; that connection is therefore not asserted.
Main entry points: views.py Result handler · loluwu.py Page retrieval · urls.py URL routing · Crawler_final.py Search client
Components 14
Web application
- URL routing
projectIRT/entrypoint/urls.py - Result handler
projectIRT/entrypoint/views.py - Search page
projectIRT/entrypoint/views.py - Results view
projectIRT/entrypoint/views.py
Search and retrieval
- Search client
projectIRT/Crawler_final.py - Page retrieval
projectIRT/loluwu.py
Summarization
- Gemini summarizer
projectIRT/summarizer.py
Knowledge graph
- Pattern graph builder
projectIRT/make_graph.py - NetworkGraph
projectIRT/graph_irt.py
Other components
- Researcher
- SerpApi search
- Web pages
- Wikipedia API
- Gemini AI
Connections 13
- Researcher → URL routing: submits query
README.md - URL routing → Search page: routes request
projectIRT/entrypoint/urls.py - URL routing → Result handler: routes request
projectIRT/entrypoint/urls.py - Result handler → Search client: fetches links
projectIRT/entrypoint/views.py - Search client → SerpApi search: queries
projectIRT/Crawler_final.py - Result handler → Wikipedia API: fetches context
projectIRT/entrypoint/views.py - Result handler → Page retrieval: requests summary
projectIRT/entrypoint/views.py - Page retrieval → SerpApi search: queries
projectIRT/loluwu.py - Page retrieval → Web pages: fetches pages
projectIRT/loluwu.py - Page retrieval → Gemini summarizer: summarizes page text
projectIRT/loluwu.py - Gemini summarizer → Gemini AI: generates summary
projectIRT/summarizer.py - Result handler → Results view: renders results
projectIRT/entrypoint/views.py - Pattern graph builder → NetworkGraph: uses graph model
projectIRT/make_graph.py