How tanishkempegowda/veriq-bootstrap works
VERIQ is a Streamlit compliance-audit application: auditors extract requirements from policy PDFs, compare evidence against selected rules, log findings, and retain or replay cases through Hindsight. The dashboard summarizes structured audit records. The page implementations often call Groq and Hindsight directly; separate service modules also exist, but the sampled pages do not import them. The graph follows those observed calls and README capabilities; unsampled behavior is not inferred.
Main entry points: 2_Gap_Detection.py Gap Detection · db.py SQLite Access · 1_Document_Analysis.py Policy Analysis · app.py Compliance Dashboard
Components 11
Audit Workflow
- Policy Analysis
pages/1_Document_Analysis.py - PDF Text Extraction
pages/1_Document_Analysis.py - Gap Detection
pages/2_Gap_Detection.py - Evidence Extraction
pages/2_Gap_Detection.py - Memory Replay
pages/3_Memory_Replay.py
AI Services
- Groq LLM
- Hindsight Memory
Audit State
- SQLite Access
database/db.py - Audit Records
database/db.py
Application UI
- Compliance Dashboard
app.py
Other components
- Auditor
Connections 12
- Auditor → Compliance Dashboard: opens
README.md - Compliance Dashboard → SQLite Access: queries counts
app.py - Policy Analysis → PDF Text Extraction: extracts text
pages/1_Document_Analysis.py - Policy Analysis → Groq LLM: extracts rules
pages/1_Document_Analysis.py - Policy Analysis → SQLite Access: saves policy
pages/1_Document_Analysis.py - SQLite Access → Audit Records: reads and writes
database/db.py - Gap Detection → SQLite Access: loads policies
pages/2_Gap_Detection.py - Gap Detection → Evidence Extraction: extracts evidence
pages/2_Gap_Detection.py - Gap Detection → Groq LLM: assesses evidence
pages/2_Gap_Detection.py - Gap Detection → SQLite Access: logs findings
pages/2_Gap_Detection.py - Gap Detection → Hindsight Memory: retains finding
pages/2_Gap_Detection.py - Memory Replay → Hindsight Memory: recalls cases
pages/3_Memory_Replay.py