How 27aran/jobby works
Jobby is an AI job-search assistant with two evidenced runtime strands: collecting job listings into SQLite, and preparing profile documents for vector search and job-profile matching. The README also advertises grounded cover-letter generation, but the supplied tree and excerpts do not show its implementation, so it is represented only as a documented capability. The RAG indexing path is shown in the store module’s executable block; no inspected source connects retrieval to job search or cover-letter generation.
Main entry points: store.py Qdrant index · retrieval.py Profile retrieval · models.py Job and profile records · cli.py Fetch-new-jobs flow
Components 14
Job collection
- Fetch-new-jobs flow
src/jobby/sources/cli.py - Listing search and details
src/jobby/sources/arbeitsagentur.py - Jobs SQLite store
src/jobby/storage/db.py - Arbeitsagentur API
Profile RAG
- Profile document loader
src/jobby/rag/ingest.py - Profile chunking
src/jobby/rag/chunk.py - Text embeddings
src/jobby/rag/embed.py - Qdrant index
src/jobby/rag/store.py
Shared domain
- Runtime settings
src/jobby/config.py - Job and profile records
src/jobby/models.py
Matching and applications
- Profile retrieval
src/jobby/rag/retrieval.py - Job-profile matching
src/jobby - Grounded cover letters
src/jobby
Other components
- Job seeker
Connections 21
- Job seeker → Fetch-new-jobs flow: starts collection
src/jobby/sources/cli.py - Fetch-new-jobs flow → Listing search and details: searches and fetches
src/jobby/sources/cli.py - Listing search and details → Arbeitsagentur API: requests listings
src/jobby/sources/arbeitsagentur.py - Fetch-new-jobs flow → Jobs SQLite store: checks and saves
src/jobby/sources/cli.py - Jobs SQLite store → Runtime settings: reads database path
src/jobby/storage/db.py - Listing search and details → Job and profile records: creates jobs
src/jobby/sources/arbeitsagentur.py - Profile document loader → Job and profile records: creates documents
src/jobby/rag/ingest.py - Profile chunking → Job and profile records: creates chunks
src/jobby/rag/chunk.py - Profile chunking → Profile document loader: loads documents in script
src/jobby/rag/chunk.py - Qdrant index → Profile document loader: loads documents in script
src/jobby/rag/store.py - Qdrant index → Profile chunking: chunks documents in script
src/jobby/rag/store.py - Qdrant index → Text embeddings: embeds chunks
src/jobby/rag/store.py - Text embeddings → Runtime settings: reads model setting
src/jobby/rag/embed.py - Qdrant index → Runtime settings: reads index path
src/jobby/rag/store.py - Profile retrieval → Text embeddings: embeds query
src/jobby/rag/retrieval.py - Profile retrieval → Qdrant index: queries index
src/jobby/rag/retrieval.py - Profile retrieval → Job and profile records: returns profile matches
src/jobby/rag/retrieval.py - Job seeker → Profile retrieval: searches profile
- Profile retrieval → Job-profile matching: supports matching
README.md - Job-profile matching → Grounded cover letters: grounds generation
README.md - Job seeker → Grounded cover letters: requests application help
README.md