How no-country-simulation/g10-brasil-equipo-26 works
MentorIA Tech turns a technical PDF, Markdown, or text document into source-grounded study material personalized by learner profile, format, and depth. The Streamlit interface submits a request to the FastAPI generation endpoint; the backend extracts pages, retrieves relevant passages, generates and validates structured content, then persists results. The product also supports document suitability checks, quiz practice, and optional OCI storage. Edges follow README-described flow and SOURCE INDEX references; unsampled implementation details are omitted where relationships cannot be verified.
Main entry points: rag/main.py FastAPI service · app/main.py Streamlit app · llm_pipeline.py LLM generation · persistencia.py Persistence facade
Components 19
Learner experience
- Streamlit app
app/main.py - User identity
app/identidade.py - HTTP client
app/api_client.py - Quiz interface
app/quiz_ui.py - Content renderers
app/renderizadores.py
Generation API
- FastAPI service
rag/main.py - Document extraction
backend/app/ingestao.py
RAG and generation
- Data contracts
rag/pipeline/schemas.py - RAG retrieval
rag/pipeline/rag_retrieval.py - LLM generation
rag/pipeline/llm_pipeline.py - Document suitability
rag/pipeline/avaliacao_documento.py
Persistence and telemetry
- SQLite store
rag/database.py - Persistence facade
rag/persistencia.py - LLM telemetry
rag/pipeline/telemetria_llm.py
Cloud artifacts
- OCI bridge
app/oci_bridge.py - OCI artifact storage
services/oci_storage.py
Other components
- Learner
- Gemini and Groq
- OCI Object Storage
Connections 26
- Learner → Streamlit app: submits document
README.md - Streamlit app → User identity: identifies user
app/main.py - Streamlit app → HTTP client: requests generation
app/main.py - HTTP client → FastAPI service: posts generation
app/api_client.py - FastAPI service → Document extraction: extracts document
rag/main.py - FastAPI service → Data contracts: validates request
rag/main.py - FastAPI service → RAG retrieval: requests context
rag/main.py - RAG retrieval → Gemini and Groq: embeds and ranks
README.md - FastAPI service → LLM generation: generates content
rag/main.py - LLM generation → Gemini and Groq: calls with fallback
README.md - LLM generation → Data contracts: validates output
rag/pipeline/llm_pipeline.py - FastAPI service → Document suitability: assesses document
rag/main.py - Document suitability → LLM generation: uses model keys
rag/pipeline/avaliacao_documento.py - FastAPI service → Persistence facade: records generation
rag/main.py - Persistence facade → SQLite store: writes records
rag/persistencia.py - LLM generation → LLM telemetry: records calls
rag/pipeline/llm_pipeline.py - RAG retrieval → LLM telemetry: records calls
rag/pipeline/rag_retrieval.py - LLM telemetry → Persistence facade: registers metrics
rag/pipeline/telemetria_llm.py - Persistence facade → SQLite store: stores telemetry
rag/persistencia.py - FastAPI service → HTTP client: returns result
README.md - Streamlit app → Content renderers: displays material
app/main.py - Streamlit app → Quiz interface: supports quiz practice
app/main.py - Quiz interface → HTTP client: requests quiz data
app/quiz_ui.py - Content renderers → OCI bridge: requests storage
app/renderizadores.py - OCI bridge → OCI artifact storage: uploads artifact
app/oci_bridge.py - OCI artifact storage → OCI Object Storage: writes JSON
README.md