Logo Lanfrica
  • Home
  • Atlas
  • Insights
  • Docs
  • Sign in

© 2026 Lanfrica. All rights reserved. All copyrights of the resources shown on the Lanfrica website belong to the original copyright holders, unless explicitly stated otherwise.

AlwiaBabekir99/efiche-RAG-nurse-assistant

Domain:

healthcare

Record type:

software
Creator:
Alw
Host:
A clinical RAG pipeline for extracting, indexing, and retrieving evidence from Rwanda healthcare guidelines. # eFiche Clinical RAG — August Corpus and Retrieval Baseline Local-first, provenance-preserving clinical guideline ingestion and retrieval for eFiche. The August scope ends at retrieval and evaluation: it does **not** generate medical advice, ingest patient records, or implement an LLM/chatbot. ## Requirements - Python 3.11–3.13 (managed automatically by `uv`) - `uv` - Docker with Compose, for Qdrant - Tesseract only when the configurable fallback OCR path is enabled in a later phase ## Reproducible workflow ```powershell # Install the locked core/dev environment uv sync # Start local Qdrant docker compose up -d qdrant # Add approved PDFs without renaming or reorganizing them # Copy files into data/raw/pdfs/ (nested folders are allowed; PDFs remain immutable) # Phase 2: discover, hash, identify exact duplicates, and create the corpus register uv run clinical-rag inventory # Subsequent independently executable August stages uv run clinical-rag inspect uv run clinical-rag parse uv run clinical-rag chunk uv run clinical-rag validate uv run clinical-rag index uv run clinical-rag retrieve "example clinical question" --strategy hybrid --top-k 5 uv run clinical-rag evaluate uv run clinical-rag report # Quality checks uv run pytest uv run ruff check . uv run ruff format --check . ``` Commands beyond `inventory` are intentional CLI placeholders until their named implementation phase is completed; they report that status and do not pretend to process data. ## Data governance `data/raw/pdfs/` is an immutable input boundary and is ignored by Git except for `.gitkeep`. The pipeline recursively discovers case-insensitive `.pdf` extensions but never renames, moves, modifies, deletes, or uploads source files. Document organization and authority are metadata. Unknown metadata remains null and is marked `NEEDS_REVIEW`. Generated datasets are also ignored by default. The corpus register is written to `data/processed/manifests/corpus_register.csv`; it contains SHA-256 has …

Visit

github.com

Tasks

information retrieval

Similar

IsaakAlemu/amharic-rag-assistanthedi-mac/Law-Assistant-RAGmauke-231/acity-rag-assistantmohamedamine050/Tunisian-Legal-Assistant--RAGGeraldM-007/Kenya-Hotel-Discovery-and-Trip-Planning-RAG-AssistantBuilding a Conversational AI Assistant for African Travel Services with LLMs and RAG

IsaakAlemu/amharic-rag-assistant

# Conversational Amharic RAG Assistant An end-to-end **conversational Retrieval-Augmented Generatio

hedi-mac/Law-Assistant-RAG

⚖️ Law-Assistant-RAG ⚖️ is a dockerized legal chatbot 💬 leveraging Retrieval-Augmented Generation (R

mauke-231/acity-rag-assistant

RAG System to answer questions based on the 2025 Ghana Budget and Election Results dataset # ACity

mohamedamine050/Tunisian-Legal-Assistant--RAG

# 🏛️ Tunisian Legal Assistant - RAG Un assistant juridique intelligent utilisant la technologie RAG

GeraldM-007/Kenya-Hotel-Discovery-and-Trip-Planning-RAG-Assistant

A Retrieval Augmented Generation(RAG) assistant that helps people discover hotels, explore destinati

Building a Conversational AI Assistant for African Travel Services with LLMs and RAG