CLAUDE.md

CLAUDE.md

Global context for working in this monorepo with Claude Code.

Purpose

This repository combines:

• Python libraries and demos
• Pytest + Playwright automation
• Robot Framework suites
• Load-testing artifacts (JMeter/Postman)
• Shared utility modules

When making changes, prefer minimal, targeted edits and avoid broad refactors unless requested.

Environment

• OS used most often: Windows (PowerShell)
• Python: 3.11+ supported; tooling config targets Python 3.12
• Install deps:
  • pip install -r requirements.txt
• Install Playwright browsers when needed:
  • playwright install

Common Commands

Lint and format

Use Ruff (configured in pyproject.toml):

• python -m ruff check .
• python -m ruff check . --fix
• python -m ruff format .

Pytest

• Full run:
  • python -m pytest
• Marker-based run:
  • python -m pytest -m ui
  • python -m pytest -m api
• Single file:
  • python -m pytest pytest_demo/tests/unit/test_csv_reader.py -q

Notes:

• pytest.ini writes Allure output to temps/allure-results
• Default testpaths include:
  • pytest
  • robot

Robot Framework

• Run all Robot demos:
  • python -m robot --outputdir temps/robot_all robot_demo/
• Run calculator suite:
  • python -m robot --outputdir temps/robot_calculator robot_demo/calculator/
• Run Tangerine Playwright suite:
  • python -m robot --outputdir temps/robot_tangerine_playwright robot_demo/tangerine_playwright/
• Dry run (syntax/keyword wiring):
  • python -m robot --dryrun --outputdir temps/robot_tangerine_playwright_dryrun robot_demo/tangerine_playwright/

Allure report

• Generate results via pytest:
  • python -m pytest --alluredir=temps/allure-results --clean-alluredir
• Open report:
  • allure serve temps/allure-results

Repo Map

• pytest: pytest-based tests (ui/api/unit/ddt/playwright)
• robot: Robot Framework suites and keyword libraries
• self_healing/: locator healing logic for UI automation
• ai_gen/: AI-based test/code generation helpers
• utils/: shared helpers and integrations (config, analytics, qTest, DB)
• skill_spring/algorithms: algorithm and data-structure examples
• load_test: JMeter and Postman assets
• temps/: generated logs, reports, videos, and temporary artifacts

Working Conventions

• Keep changes scoped to the requested task.
• Do not edit generated outputs in temps/ unless explicitly asked.
• Prefer stable selectors and reusable helpers for UI tests.
• For Robot keyword/library changes, keep keyword names descriptive and consistent.
• Reuse shared helpers in utils/ before adding duplicate logic.
• Preserve existing style rules from pyproject.toml:
  • line length 100
  • double quotes
  • sorted imports via Ruff/isort profile

Test Strategy After Changes

Pick the smallest meaningful validation first, then broaden only if needed.

• Python utility/module change:
  • run targeted pytest file(s)
• Pytest UI/API change:
  • run impacted test module and marker subset
• Robot change:
  • run target suite, optionally dry run first
• Cross-cutting change:
  • run python -m pytest and at least one relevant Robot suite

Load Test Assets Guidance

• JMeter and Postman files in load_test are source artifacts.
• Avoid mass reformatting JSON/JMX files unless task requires it.
• If environments are modified, keep variable names backward compatible when possible.

Safety and Secrets

• Never hardcode secrets or tokens.
• Use environment variables for API keys and endpoints.
• Keep defaults in config/config.py and override through env vars.

PR/Change Notes

When summarizing work, include:

• files changed
• commands run
• result status (pass/fail/not run)
• follow-up actions if validation was partial