annampaul27/nyayasetu

How annampaul27/nyayasetu works

NyayaSetu’s represented product is a legal-query understanding agent: a citizen message is classified, normalized into a legal question and jurisdiction, enriched with domain-specific facts, then assessed for intake readiness and PII safety. The agent returns a structured, traceable result with follow-up or escalation guidance. The README describes this narrow module; source excerpts are partial, so internal stage wiring beyond the documented orchestration is shown only at the agent level.

Main entry points: query_understanding_agent.py Query Agent · llm_client.py LLM Client · models.py Query Result · query_understanding_prompts.py Prompt Suite

Components 15

Query Workflow

Query Agent src/query_understanding_agent.py
Domain Detection src/query_understanding_agent.py
Question Extraction src/query_understanding_agent.py
Jurisdiction Extraction src/query_understanding_agent.py
Fact Extraction src/query_understanding_agent.py
Readiness Check src/query_understanding_agent.py
PII Safety Check src/query_understanding_agent.py
Query Result src/models.py

Agent Support

Prompt Suite src/query_understanding_prompts.py
Domain Checklists src/domain_checklists.py
PII Redactor src/pii_redaction.py

LLM Provider

LLM Configuration src/config.py
LLM Client src/llm_client.py

Other components

Citizen
Google Gemini

Connections 18

Citizen → Query Agent: submits query
Query Agent → Domain Detection: runs stage
Query Agent → Question Extraction: runs stage
Query Agent → Jurisdiction Extraction: runs stage
Query Agent → Fact Extraction: runs stage
Query Agent → Readiness Check: runs stage
Query Agent → PII Safety Check: runs stage
Query Agent → Prompt Suite: uses prompts
Query Agent → Domain Checklists: checks requirements
Query Agent → Query Result: constructs result
Domain Detection → LLM Client: requests completion
Question Extraction → LLM Client: requests completion
Jurisdiction Extraction → LLM Client: requests completion
Fact Extraction → LLM Client: requests completion
Readiness Check → LLM Client: requests completion
PII Safety Check → LLM Client: requests completion
LLM Client → LLM Configuration: reads settings
LLM Client → Google Gemini: calls API

More Python diagrams: ictnlp/auto-rag · ictnlp/flexrag · mercadopago/sdk-python · oca/manufacture · piqueserver/piqueserver

Drawn by GitDiagram on September 28, 2026 from the repository's file tree, README and source files. Read as Markdown ·