QA Lens Docs

QA Lens Roadmap

This roadmap reflects the current development direction. Priorities may shift based on community feedback.

---

v1.0 — Foundation

Goal: Reliable, local, explainable insights from Extent and Allure reports.

AreaStatus
Canonical Pydantic models✅ Done
Extent HTML parser✅ Done
Allure HTML/JSON parser✅ Done
Report type detector✅ Done
Failure signature engine✅ Done
Rule-based categorization✅ Done
Deterministic failure clustering✅ Done
Flaky scoring with run history✅ Done
Summary generators✅ Done
JSON output✅ Done
Markdown output✅ Done
Console/Rich output✅ Done
Typer CLI✅ Done
Python library API✅ Done
SQLite run history store✅ Done
Web UI/API server✅ Done
Run comparison engine✅ Done
LLM chat/ask integration✅ Done
Security hardening baseline✅ Done
Unit + fixture-based tests✅ Done
Ruff + mypy cleanup🔄 In progress
Pre-commit configuration🔄 In progress

---

v1.1 — Current Polish

AreaStatus
CI-friendly exit code semantics✅ Done
GitHub Actions example workflow✅ Done
Machine-readable analysis output✅ Done
Demo dataset for GitHub users✅ Done
Deterministic qalens ask answers for factual aggregate questions✅ Done
Frontend XSS sanitizer regression coverage✅ Done
Dependency/security CI checks🔄 In progress
Improved Extent v5 parser coverage🔄 Ongoing
Shell completion documentation⏳ Planned
Broader CLI JSON output coverage⏳ Planned

---

v1.2 — Historical Intelligence

AreaStatus
SQLite-backed run history✅ Done
New/recurring/recovered failure detection✅ Done
Historical flaky leaderboard✅ Done
Run comparison API and UI✅ Done
Trend analysis facts✅ Done
Trend charts in Markdown⏳ Planned
qalens history command✅ Done
Explicit-run qalens compare --run-id ... support✅ Done

---

v1.3 — Parser Breadth

  • pytest-html report parser
  • More JUnit XML fixture variants
  • More TestNG XML fixture variants
  • More Cypress / Playwright fixture variants
  • More real-world Extent and Allure fixture variants

---

v2.0 — ML Enrichment (Optional Layer)

  • Optional TF-IDF + cosine similarity fuzzy clustering
  • Optional sentence-transformer-based semantic grouping (local model, no cloud)
  • Ensemble confidence scoring (heuristic + ML combined)
  • Anomaly detection on timing patterns
  • All ML features opt-in, clearly labeled, and explainable

---

v2.1 — Richer Outputs

  • Standalone HTML insight report (self-contained, no server needed)
  • SARIF output (for GitHub Security tab and IDE integrations)
  • Slack/Teams webhook notification output (plugin)
  • PDF export option

---

Future Considerations

  • Plugin marketplace / entry-point auto-discovery
  • VS Code extension for inline insight annotations
  • Slack/Teams notification plugins
  • SARIF export

---

Non-Goals (Permanent)

  • QA Lens will never be a test reporting framework
  • QA Lens will never replace Extent Reports or Allure
  • QA Lens v1 will never require cloud connectivity
  • QA Lens will never silently discard data without an ExtractionWarning

---

Contributing to the Roadmap

If you have a use case that is not covered here, open a GitHub Discussion or a feature request issue. We especially welcome:

  • Report format parsers for tools used in your organization
  • Heuristic categorization signals from your real-world failure patterns
  • CI integration patterns and workflows