How ttstive/challenge02_grupo10 works
This is a lightweight neurocognitive-speech RAG scaffold. In the documented example, a caller supplies a query and curated speech context to the RAG factory; retrieval builds a typed page index, searches it, and formats matching pages as context. A separate optional LangGraph entry point retrieves pages and assembles a simple answer from the question and context. The index uses term matching, not embeddings or a persistent vector store. Connections reflect the README example and sampled source; no unsampled document-ingestion-to-runtime path is established.
Main entry points: retrieval_chain.py RAG factory · page_index.py Page index · workflow.py LangGraph workflow · speech_context.py Speech context
Components 7
Domain content
- Speech context
src/challenge02_grupo10/data/speech_context.py
Retrieval
- RAG factory
src/challenge02_grupo10/rag/retrieval_chain.py - Page index
src/challenge02_grupo10/indexing/page_index.py - Retrieved context
src/challenge02_grupo10/rag/retrieval_chain.py
Graph workflow
- LangGraph workflow
src/challenge02_grupo10/langgraph/workflow.py - Formatted answer
src/challenge02_grupo10/langgraph/workflow.py
Other components
- RAG caller
Connections 7
- RAG caller → RAG factory: builds chain
README.md - RAG caller → Speech context: supplies context
README.md - RAG factory → Page index: builds index
src/challenge02_grupo10/rag/retrieval_chain.py - RAG factory → Page index: searches pages
src/challenge02_grupo10/rag/retrieval_chain.py - RAG factory → Retrieved context: formats matches
src/challenge02_grupo10/rag/retrieval_chain.py - LangGraph workflow → Page index: searches pages
src/challenge02_grupo10/langgraph/workflow.py - LangGraph workflow → Formatted answer: formats answer
src/challenge02_grupo10/langgraph/workflow.py