How imnazirul/job-hunter-django works
JobHunter is a Django API that turns a reviewed candidate profile into personalized job-search results. The main workflow is CV upload and parsing, profile review, asynchronous searches across available boards, job normalization and deduplication, then rule-based scoring with optional LLM refinement. Users can poll search progress and browse scored jobs. The README also documents registration and authentication. The graph follows documented endpoints and the sampled task, parsing, source, and scoring relationships; some wiring is supported by the SOURCE INDEX rather than full-file inspection.
Main entry points: jobs/views.py Search and jobs API · jobs/tasks.py Search orchestration · profiles/tasks.py CV processing · profiles/views.py Profile and CV API
Components 26
Accounts and profiles
- Auth API
accounts/views.py - Account data
accounts/models.py - Profile and CV API
profiles/views.py - Profile and CV records
profiles/models.py - CV processing
profiles/tasks.py - CV text extraction
profiles/cv_text.py - Profile parsing
profiles/parsing.py - Skill vocabulary
profiles/skill_vocabulary.py
Job discovery
- Search and jobs API
jobs/views.py - Search orchestration
jobs/tasks.py - Source registry
jobs/sources/registry.py - Remotive source
jobs/sources/remotive.py - Source contract and HTTP
jobs/sources/base.py - Job normalization
jobs/normalize.py - Duplicate detection
jobs/dedup.py - Search and job records
jobs/models.py
Scoring and language models
- Scoring service
jobs/scoring/service.py - Rule scoring
jobs/scoring/rules.py - LLM re-scoring
jobs/scoring/llm_scorer.py - LLM client
llm/client.py - LLM output coercion
llm/coerce.py
Runtime and data
- Celery worker
jobhunter/celery.py - Uploaded CV storage
jobhunter/storage.py
Other components
- Candidate
- OpenRouter
- Remotive API
Connections 36
- Candidate → Auth API: registers or signs in
README.md - Auth API → Account data: uses accounts
accounts/views.py - Candidate → Profile and CV API: uploads CV; reviews profile
README.md - Profile and CV API → Profile and CV records: reads and updates
profiles/views.py - Profile and CV API → CV processing: queues parsing
profiles/views.py - CV processing → Profile and CV records: updates CV and profile
profiles/tasks.py - CV processing → CV text extraction: extracts text
profiles/tasks.py - CV processing → Profile parsing: parses extracted text
profiles/tasks.py - Profile parsing → Skill vocabulary: uses vocabulary
profiles/parsing.py - Profile parsing → LLM client: requests structured output
profiles/parsing.py - Profile parsing → LLM output coercion: validates output
profiles/parsing.py - Candidate → Search and jobs API: starts search; polls results
README.md - Search and jobs API → Profile and CV records: checks candidate profile
jobs/views.py - Search and jobs API → Search and job records: reads and creates search records
jobs/views.py - Search and jobs API → Source registry: lists available sources
jobs/views.py - Search and jobs API → Search orchestration: dispatches search
jobs/views.py - Search and jobs API → Celery worker: queues asynchronous work
README.md - Search orchestration → Search and job records: updates runs and jobs
jobs/tasks.py - Search orchestration → Source registry: selects source
jobs/tasks.py - Source registry → Remotive source: registers source
jobs/sources/registry.py - Search orchestration → Job normalization: normalizes fetched jobs
jobs/tasks.py - Remotive source → Source contract and HTTP: uses source HTTP helpers
jobs/sources/remotive.py - Remotive source → Remotive API: fetches listings
jobs/sources/remotive.py - Remotive source → Job normalization: normalizes source content
jobs/sources/remotive.py - Search orchestration → Duplicate detection: marks duplicates
jobs/tasks.py - Duplicate detection → Job normalization: uses normalized keys
jobs/dedup.py - Search orchestration → Scoring service: scores completed run
jobs/tasks.py - Scoring service → Search and job records: loads and updates scores
jobs/scoring/service.py - Scoring service → Rule scoring: scores every job
jobs/scoring/service.py - Rule scoring → Skill vocabulary: matches canonical skills
jobs/scoring/rules.py - Scoring service → LLM re-scoring: refines shortlist
jobs/scoring/service.py - LLM re-scoring → LLM client: requests validated scores
jobs/scoring/llm_scorer.py - LLM re-scoring → LLM output coercion: coerces response fields
jobs/scoring/llm_scorer.py - LLM client → OpenRouter: sends model requests
llm/client.py - Profile and CV API → Uploaded CV storage: stores uploaded CVs
README.md - CV processing → Celery worker: runs as background task
profiles/tasks.py