Add CLI and GUI integration tests for all parser plugins

[ndaelman-hu]

Purpose

Add integration testing for all 15 parser plugins via CLI and GUI to validate end-to-end user workflows.

Scope

Included

• CLI integration tests (20 test cases via nomad parse subprocess)
• Test matrix YAML defining expected outcomes for all parsers
• Documentation of framework and known parser bugs
• GUI integration tests with Playwright (framework only - see status below)
• GUI test fixtures and documentation

Out of scope

• Parser bug fixes (3 bugs identified for separate PRs)
• Deep field validation (covered by existing unit tests)

Context & Links

Integration tests complement existing unit tests by validating the full user workflow (parser registration, CLI/GUI interfaces, archive serialization). Unit tests verify parser logic correctness; integration tests verify parsers work end-to-end.

Test results:

• CLI: 17/17 passing ✅ (3 skipped pending parser fixes, yambo skipped - no test data)
• GUI: Framework complete, smoke tests pass (2/2), full workflow tests blocked (see status)

Parser bugs identified for separate PRs:

• exciting::C minimal - model_system not populated
• lammps::XYZ trajectory - parser timeout with large files
• phonopy::VASP phonopy - AttributeError with phonopy 2.35+ API

Reviewer Notes

CLI test validation:

• First-order checks: archive created, program identified, workflow type correct, required fields present
• Catches integration issues unit tests miss: file paths, plugin registration, subprocess behavior
• Uses YAML test matrix for maintainability

GUI test implementation:

• Uses Playwright for real browser automation
• Tests upload → processing → archive display workflow
• Validates UI elements, metadata extraction, workflow badges
• Separate [gui-test] dependency group (~300MB browser download)
• Can skip in CI with pytest -m "not gui"

GUI Testing Status & Options

Current Status:

• ✅ Infrastructure complete (Playwright, pytest-playwright, fixtures)
• ✅ Smoke tests passing (server detection, page loading)
• ❌ Upload workflow tests blocked: This repo has no GUI source code

Issue: This is a parser plugins repository. The NOMAD GUI source lives in the separate nomad-FAIR repository. The backend API server runs, but doesn't serve the React GUI in development mode.

Options for Resolution:

Option 1: Remove GUI Tests (Recommended)

CLI tests already provide comprehensive parser validation. GUI tests are redundant for parser plugin testing.

Rationale:

• Parser correctness validated by CLI tests (archive generation, field population)
• GUI testing belongs in nomad-FAIR repository (UI-focused)
• Reduces complexity and CI dependencies

Action: Remove tests/gui/ directory, keep [gui-test] dependencies for reference

Option 2: Document as Advanced/Optional

Keep GUI test framework as reference for developers with full NOMAD stack.

Requirements:

• Access to NOMAD GUI source (separate setup)
• React dev server running on port 3000, OR
• Built GUI served by backend

Action: Add "requires full NOMAD setup" notice to README, mark tests to skip by default

Option 3: Test Against External Instance

Point GUI tests to deployed NOMAD instance (e.g., nomad-lab.eu).

Downsides:

• External dependency (unreliable for CI)
• Requires authentication
• Can't control test data cleanup

Action: Update fixtures to use external URL, add auth fixtures

Implementation Progress

Phase 1: CLI Tests

• Create test framework (test_cli_parsing.py)
• Create test matrix YAML
• Add pytest markers to pyproject.toml
• Fix test matrix file paths (h5md, octopus)
• Fix expected workflow types (exciting, QE)
• Mark parser bugs as skipped with tracking comments
• Document known parser bugs in README
• Run and validate tests (17/17 passing)

Phase 2: GUI Tests

• Add Playwright dependencies to separate [gui-test] group
• Create GUI test framework (test_upload_workflow.py)
• Create test fixtures for server and browser management (conftest.py)
• Implement upload workflow test logic
• Document GUI test setup and usage (tests/gui/README.md)
• Install browsers and validate infrastructure (smoke tests 2/2 passing)
• ❌ Full workflow tests blocked (no GUI source in this repo)

Status

• Draft
• Ready for review (CLI tests complete, GUI decision pending)
• Blocked

Breaking Changes

• None

Dependencies / Blockers

• None for CLI tests (core functionality)
• Decision needed: GUI testing approach (see options above)

Testing / Validation

• CLI integration tests (17/17 passing in 174s)
• GUI smoke tests (2/2 passing - server detection, page load)
• GUI upload workflow tests (17 tests - blocked, see status)

CLI Test Results:

uv run pytest tests/cli/test_cli_parsing.py::test_cli_parsing -v
# ======================== 17 passed in 174.33s (0:02:54) ========================

GUI Smoke Test Results:

uv run pytest tests/gui/test_upload_workflow.py::test_gui_server_accessible tests/gui/test_upload_workflow.py::test_upload_page_loads -v
# ======================== 2 passed in 1.85s ========================

GUI Setup (for reference):

# Install dependencies
uv pip install -e ".[gui-test]"

# Install Chromium browser (~300MB)
uv run playwright install chromium

# Run smoke tests (working)
uv run pytest tests/gui/ -v -m gui -k "accessible or loads"

# Run full tests (blocked - requires GUI source)
uv run pytest tests/gui/ -v -m gui