How kishorpatilaws-gith/clinical-assistance works
ClinicalAI is a clinical records and document knowledge application. Clinicians use the web UI to manage patients and documents, then ask questions over indexed material; the API coordinates records, document storage and processing, and retrieval-grounded answers. The Python service extracts and indexes documents and serves search, while the API adds PII redaction and Bedrock generation. The README is empty, so capabilities are inferred from the supplied tree and excerpts; dashed links mark likely runtime connections not evidenced by an excerpt.
Main entry points: Program.cs API runtime · app Clinical UI · main.py Processing service · DocumentController.cs Document endpoints
Components 22
Clinician UI
- Clinical UI
frontend/clinical-ai-ui/app - Authentication
frontend/clinical-ai-ui/lib/auth.ts
Clinical API
- API runtime
backend/ClinicalAI.Api/ClinicalAI.Api/Program.cs - Patient endpoints
backend/ClinicalAI.Api/ClinicalAI.Api/Controllers/PatientController.cs - Document endpoints
backend/ClinicalAI.Api/ClinicalAI.Api/Controllers/DocumentController.cs - RAG endpoint
backend/ClinicalAI.Api/ClinicalAI.Api/Controllers/RagController.cs - Clinical records
- Patient summaries
- Audit logging
Document processing
- Object storage
- Indexing queue
- Indexing worker
- Document indexer
- Processor client
- Processing service
document-processor/main.py - Text chunking
document-processor/chunking.py
RAG and AI
- Embeddings
document-processor/embedding.py - Vector search
document-processor/vectorstore.py - RAG orchestration
- PII redaction
- Bedrock generation
Other components
- Clinician
Connections 30
- Clinician → Clinical UI: uses
- Clinical UI → Authentication: authenticates
- Clinical UI → Patient endpoints: requests records
- Clinical UI → Document endpoints: uploads documents
- Clinical UI → RAG endpoint: asks questions
- API runtime → Clinical records: configures
backend/ClinicalAI.Api/ClinicalAI.Api/Program.cs - API runtime → Object storage: registers
backend/ClinicalAI.Api/ClinicalAI.Api/Program.cs - API runtime → Processor client: configures
backend/ClinicalAI.Api/ClinicalAI.Api/Program.cs - API runtime → Indexing queue: registers
backend/ClinicalAI.Api/ClinicalAI.Api/Program.cs - API runtime → Document indexer: registers
backend/ClinicalAI.Api/ClinicalAI.Api/Program.cs - API runtime → Indexing worker: registers
backend/ClinicalAI.Api/ClinicalAI.Api/Program.cs - API runtime → RAG orchestration: registers
- API runtime → Patient summaries: registers
backend/ClinicalAI.Api/ClinicalAI.Api/Program.cs - API runtime → Audit logging: registers
backend/ClinicalAI.Api/ClinicalAI.Api/Program.cs - Indexing worker → Clinical records: updates status
- Indexing worker → Document indexer: dispatches
- Document indexer → Clinical records: updates document
- Document indexer → Processor client: requests indexing
- Processor client → Processing service: calls over HTTP
document-processor/main.py - Processing service → Text chunking: chunks text
document-processor/main.py - Processing service → Embeddings: creates vectors
document-processor/main.py - Processing service → Vector search: indexes and searches
document-processor/main.py - Vector search → Embeddings: embeds queries
document-processor/vectorstore.py - RAG orchestration → PII redaction: redacts question
- RAG orchestration → Processor client: searches passages
- RAG orchestration → Bedrock generation: generates answer
- Patient endpoints → Clinical records: reads and writes
backend/ClinicalAI.Api/ClinicalAI.Api/Controllers/PatientController.cs - Document endpoints → Clinical records: reads and writes
backend/ClinicalAI.Api/ClinicalAI.Api/Controllers/DocumentController.cs - Document endpoints → Object storage: uses storage
- RAG endpoint → RAG orchestration: invokes