Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.2.0] - 2026-08-04

Asynchronous RAG Service: Added buscar_async to CorpusLegal with asyncio.to_thread execution for non-blocking ChromaDB searches.
Asynchronous Pre-warming: Added lifespan event listener in main.py to pre-warm the legal corpus vectorstore on startup.
Automated Test Suite: Added E2E & Smoke test suite (test_e2e_smoke.py), Security & Edge Cases suite (test_edge_cases_security.py), and Performance/Concurrency suite (test_performance.py) reaching 44 passing tests (~92% line coverage).
XSS Protection: Added HTML sanitization (escapeHtml) in chrome-extension/popup.js and web/js/app.js.

Pydantic V2 Migration: Migrated backend/app/config.py from Pydantic V1 class Config: to SettingsConfigDict(env_file=".env", extra="ignore").
JSON Serialization: Updated /analizar route to dump Pydantic models with mode="json" before storing in SQLAlchemy JSON column.
Regex Performance: Pre-compiled regex patterns (re.compile) in RiskDetector and PDFParser on initialization.
ChromaDB Test Isolation: Isolated ChromaDB database path per corpus directory to prevent state collision between test runs.

Fixed TypeError on datetime serialization when saving analysis results in SQLite.
Fixed HTTP 500 error on invalid or corrupted PDF uploads by catching PDF parsing errors and returning HTTP 400.
Fixed risk level casing and accent mismatch (crítico vs critico) in RiskDetector.
Fixed missing dependencies aiosqlite, pypdf, and python-multipart in pyproject.toml and requirements.txt.