How prathameshdeshpande99/software-tool works
This repository is an AI-powered dataset-use compliance checker. Users submit rules and a codebase through the web API or CLI; the service parses rules, gathers source, runs a vanilla Gemini audit or an advanced retrieval-based audit, and returns a structured violation report. The asynchronous web workflow dispatches work to a Celery worker and exposes task status. The README also documents the Flask interface and legal-PDF extraction. Advanced RAG connections are grounded in the sampled audit source; unsampled UI internals and some web wiring are not inferred.
Main entry points: server.py Flask API · audit.py Advanced Audit · main.py Web API · analyzer.py Vanilla Analyzer
Components 21
Audit Intake
- Codebase Intake
main.py - Local/GitHub Loader
src/compliance_checker/codebase_loader.py
Rule Processing
- Rule Input
main.py - PDF Rule Extractor
src/compliance_checker/pdf_rule_extractor.py - YAML Rule Loader
src/compliance_checker/rules_loader.py - Rules Schema
src/compliance_checker/models.py
Compliance Analysis
- Vanilla Analyzer
src/compliance_checker/analyzer.py - Advanced Audit
src/audit.py - Semantic Chunker
src/semantic_chunker.py - Incremental Vector Store
src/compliance_checker/vector_store.py - Compliance Report
src/compliance_checker/models.py
Background Jobs
- Analysis Worker
worker.py - Celery App
celery_app.py
Other components
- User
- Qdrant
- Google Gemini
- Redis
Connections 31
- User → Web UI: submits audit
README.md - User → CLI: runs scan
README.md - Web UI → Web API: sends request
- User → Flask API: uses interface
README.md - CLI → YAML Rule Loader: loads rules
src/compliance_checker/cli.py - CLI → Local/GitHub Loader: loads codebase
src/compliance_checker/cli.py - CLI → Vanilla Analyzer: runs analysis
src/compliance_checker/cli.py - Web API → Codebase Intake: accepts input
main.py - Web API → Rule Input: parses rules
main.py - Web API → Analysis Worker: dispatches task
main.py - Web API → Celery App: reads task status
main.py - Flask API → Codebase Intake: accepts input
server.py - Flask API → Rule Input: parses rules
server.py - Flask API → Local/GitHub Loader: loads GitHub code
server.py - Flask API → Advanced Audit: runs advanced audit
server.py - Flask API → Vanilla Analyzer: runs vanilla audit
server.py - Rule Input → PDF Rule Extractor: extracts PDF rules
main.py - PDF Rule Extractor → Rules Schema: validates rules
src/compliance_checker/pdf_rule_extractor.py - YAML Rule Loader → Rules Schema: creates schema
src/compliance_checker/rules_loader.py - Codebase Intake → Local/GitHub Loader: fetches GitHub
main.py - Analysis Worker → Advanced Audit: runs advanced audit
worker.py - Analysis Worker → Vanilla Analyzer: runs vanilla audit
- Analysis Worker → Celery App: updates task state
worker.py - Celery App → Redis: stores task state
- Advanced Audit → Semantic Chunker: chunks Python
src/audit.py - Advanced Audit → Incremental Vector Store: syncs codebase
src/audit.py - Incremental Vector Store → Qdrant: indexes chunks
src/compliance_checker/vector_store.py - Advanced Audit → Google Gemini: queries violations
src/audit.py - Vanilla Analyzer → Google Gemini: requests analysis
src/compliance_checker/analyzer.py - Advanced Audit → Compliance Report: returns report
src/audit.py - Vanilla Analyzer → Compliance Report: returns report
src/compliance_checker/analyzer.py