How 10-tux/murphy works
MURPHY is a FastAPI codebase assistant: developers provide a repository, which is parsed and indexed for retrieval-augmented natural-language answers. The graph follows the upload-to-query workflow and includes session, API, and model layers. README capabilities such as architecture exploration and documentation generation are product goals, not separately verified workflows here. Unsampled indexing internals are included where the source index identifies them; uncertain links are omitted.
Main entry points: main.py FastAPI app · repository.py Repository API · repository_indexing_service.py Repository indexer · rag_service.py RAG pipeline
Components 25
API and sessions
- FastAPI app
main.py - Repository API
app/api/repository.py - Query API
app/api/query.py - Files API
app/api/files.py - History API
app/api/history.py - Status API
app/api/status.py - Session registry
app/services/session_service.py
Repository intake
- Repository handling
app/services/repository_service.py - Temporary workspace
app/utils/temporary_workspace.py
Parsing and indexing
- Repository indexer
app/services/repository_indexing_service.py - Generic parser
app/parsers/generic_parser.py - Python parser
app/parsers/python_parser.py - Document indexing
app/services/indexing_service.py - Code chunking
app/services/chunking_service.py
Retrieval and answers
- FAISS vector store
app/services/vector_store_service.py - Embedding provider
app/services/embedding_service.py - RAG pipeline
app/services/rag_service.py - Prompt construction
app/services/prompt_service.py - LLM service
app/services/llm_service.py
Configuration and models
- Parsed file model
app/models/parsed_file.py - API schemas
app/models/api_models.py - Application settings
app/core/config.py
Other components
- Developer
- Google Gemini
- Sentence Transformers
Connections 36
- Developer → FastAPI app: uses API
README.md - FastAPI app → Repository API: mounts route
main.py - FastAPI app → Query API: mounts route
main.py - FastAPI app → Files API: mounts route
main.py - FastAPI app → History API: mounts route
main.py - FastAPI app → Status API: mounts route
main.py - Repository API → Repository handling: extracts and discovers
app/api/repository.py - Repository API → Repository indexer: builds index
app/api/repository.py - Repository API → Session registry: creates session
app/api/repository.py - Repository API → Temporary workspace: creates workspace
- Query API → Session registry: gets session
app/api/query.py - Query API → RAG pipeline: invokes query
app/api/query.py - Files API → Session registry: reads session
app/api/files.py - History API → Session registry: reads history
app/api/history.py - Status API → Session registry: checks session
app/api/status.py - Repository indexer → Generic parser: selects parser
app/services/repository_indexing_service.py - Repository indexer → Python parser: selects parser
app/services/repository_indexing_service.py - Repository indexer → Document indexing: indexes files
app/services/repository_indexing_service.py - Repository indexer → RAG pipeline: builds service
app/services/repository_indexing_service.py - Generic parser → Parsed file model: creates model
app/parsers/generic_parser.py - Python parser → Parsed file model: creates model
app/parsers/python_parser.py - Document indexing → Code chunking: chunks content
app/services/indexing_service.py - Document indexing → FAISS vector store: stores chunks
app/services/indexing_service.py - Code chunking → Parsed file model: reads structure
app/services/chunking_service.py - FAISS vector store → Embedding provider: creates vectors
app/services/vector_store_service.py - RAG pipeline → FAISS vector store: retrieves documents
app/services/rag_service.py - RAG pipeline → Prompt construction: builds prompt
app/services/rag_service.py - RAG pipeline → LLM service: generates answer
app/services/rag_service.py - Embedding provider → Google Gemini: creates provider
app/services/embedding_service.py - Embedding provider → Sentence Transformers: creates provider
app/services/embedding_service.py - LLM service → Google Gemini: configures model
app/services/llm_service.py - Embedding provider → Application settings: loads settings
app/services/embedding_service.py - LLM service → Application settings: loads settings
app/services/llm_service.py - FAISS vector store → Application settings: loads settings
app/services/vector_store_service.py - Query API → API schemas: validates response
app/api/query.py - Repository API → API schemas: returns upload schema
app/api/repository.py