metadata
license: apache-2.0
task_categories:
- sentence-similarity
tags:
- medical
language:
- ar
Shifaa Medical RAG System
Advanced Retrieval-Augmented Generation (RAG) system for Arabic medical consultations.
๐ฏ Overview
The Shifaa Medical RAG system provides intelligent medical information retrieval through a four-stage pipeline:
Query โ Specialty Detection โ Topic Paths โ Consultation Retrieval โ Insight Extraction
Key Features
- Automatic Specialty Detection: Identifies relevant medical specialties from 23 categories
- Hierarchical Topic Navigation: Pinpoints specific medical topics from 585 diagnoses
- Semantic Search: Retrieves similar consultations from 84K+ medical cases
- Insight Extraction: Distills actionable medical information from retrieved consultations
- Multi-lingual Support: Primary support for Arabic with multilingual capabilities
- Auto-Download: Automatically manages vector database downloads
๐ Quick Start
Basic Usage
from shifaa.rag import MedicalRAGSystem
# Initialize the system (auto-downloads vector DB if needed)
rag = MedicalRAGSystem()
# Process a medical query
query = "ู
ุง ูู ุฃุนุฑุงุถ ุงุฑุชุฌุงุน ุงูู
ุฑูุกุ"
results = rag.process_query(query)
# Access results
print("Specialties:", [s.specialty for s in results.specialties])
print("Topics:", [t.path for t in results.topic_paths])
print("Insights:", [i.information for i in results.insights])
With Google API Key
from shifaa.rag import MedicalRAGSystem
rag = MedicalRAGSystem(
google_api_key="your-api-key-here"
)
results = rag.process_query("ู
ุง ุนูุงุฌ ุงูุตุฏุงุน ุงูู
ุฒู
ูุ")
๐ฆ Installation & Setup
Prerequisites
pip install shifaa