How kamilwodnicki/hackaton2026 works
This repository combines a social-innovation assistant and discovery platform with semantic search, regional-needs analysis, and data-import pipelines. The main user-facing workflow accepts a chat message, applies crisis checks, gathers project context and county indicators, then requests and filters a language-model response. Separately, the documented search API retrieves and reranks indexed document passages. The frontend also contains account, mentoring, forum, idea, and needs-map surfaces; their runtime wiring is not shown in the sampled evidence, so those are represented without inferred API edges.
Main entry points: chat.py Chat API · app/main.py Search API · biblioteka/main.py Innovation library import · ioss/main.py IOSS import
Components 22
User experience
- Web surfaces
frontend - Accounts and roles
frontend/auth.js - Mentoring
frontend/mentor.js - Needs map
frontend/mapa-potrzeb.js - Ideas and forum
frontend/residents-ideas.html
Assistant services
- Chat API
app/chat.py - Safety checks
app/guardrails.py - County needs analysis
app/ioss_needs.py
Knowledge search
- Search API
app/main.py - Qdrant index
- Document ingestion
app/ingest.py
Data imports
- Innovation library import
scrapers/biblioteka/main.py - Library scraper
scrapers/biblioteka/scraper.py - IOSS import
scrapers/ioss/main.py - IOSS scraper
scrapers/ioss/scraper.py - IOSS persistence
scrapers/ioss/database.py - PostgreSQL IOSS data
Other components
- Platform user
- Ollama language model
- Embedding model
- Cross-encoder reranker
- Innovation library site
Connections 16
- Platform user → Chat API: sends message
- Chat API → Safety checks: checks crisis
app/chat.py - Chat API → County needs analysis: gets needs
app/chat.py - Chat API → Search API: requests context
app/chat.py - Chat API → Ollama language model: requests answer
app/chat.py - Search API → Embedding model: embeds query
app/main.py - Search API → Qdrant index: searches vectors
README.md - Search API → Cross-encoder reranker: reranks matches
app/main.py - Document ingestion → Qdrant index: writes vectors
app/ingest.py - Innovation library import → Library scraper: runs scraper
scrapers/biblioteka/main.py - Library scraper → Innovation library site: fetches content
scrapers/biblioteka/scraper.py - Innovation library import → Document ingestion: supplies index data
README.md - IOSS import → IOSS scraper: scrapes indicators
scrapers/ioss/main.py - IOSS import → IOSS persistence: saves results
scrapers/ioss/main.py - IOSS persistence → PostgreSQL IOSS data: writes table
scrapers/ioss/database.py - County needs analysis → PostgreSQL IOSS data: reads indicators
app/ioss_needs.py