Add CI Shepherd to main

[oxve]

Merge ci_shepherd.yml and script files to main branch for collation testing.

[github-actions]

🤖 Gemini Suggested Commit Message

---

ci: Implement CI Shepherd for report collation

Introduce the CI Shepherd workflow and supporting scripts to aggregate
results from disparate CI runs. This workflow triggers upon the
completion of platform-specific builds (Android, AOSP, Evergreen, and
Linux) to collect validation report fragments.

The collected data is collated into a single summary, formatted as a
Markdown table, and posted as a status update and pull request
comment. This provides a unified view of CI health across all
supported platforms, simplifying the verification process for
developers.

Issue: 867

---

💡 Pro Tips for a Better Commit Message:

1. Influence the Result: Want to change the output? You can write custom prompts or instructions directly in the Pull Request description. The model uses that text to generate the message.
2. Re-run the Generator: Post a comment with: /generate-commit-message

[gemini-code-assist]

Code Review

This pull request introduces two new Python scripts, collate_ci_reports.py and format_ci_report.py, to aggregate and format GitHub Actions CI workflow reports into a unified Markdown table. The review feedback highlights several robust improvements: handling potential command failures when downloading artifacts to prevent script crashes, clearing the temporary fragments directory to avoid stale data, defensively accessing dictionary keys to prevent KeyError exceptions, and providing fallback defaults for missing report fields to ensure clean Markdown output.

[gemini-code-assist]

If a workflow run fails before generating any artifacts (e.g., due to a compilation or setup failure), gh run download will fail with a non-zero exit code. Since run_cmd uses check=True, this will raise a CalledProcessError and crash the entire collation script, preventing other workflows from being reported. Wrapping this in a try-except block allows the script to handle failures gracefully and record the error.

    try:
      run_cmd([
          'gh',
          'run',
          'download',
          run_id,
          '--pattern',
          'validation-report-*',
          '--dir',
          'fragments',
      ])
    except subprocess.CalledProcessError:
      if conclusion != 'success':
        report_data.append({
            'workflow': wf_name,
            'status': 'completed',
            'error': f'Workflow failed ({conclusion})'
        })
      else:
        report_data.append({
            'workflow': wf_name,
            'status': 'completed',
            'error': 'Failed to download validation report'
        })

[gemini-code-assist]

If this script is run multiple times in a persistent environment or locally, stale JSON files in the fragments directory from previous runs will be read and merged into the new report. It is safer to clear the directory before downloading new fragments.

Suggested change

	os.makedirs('fragments', exist_ok=True)
	import shutil
	shutil.rmtree('fragments', ignore_errors=True)
	os.makedirs('fragments', exist_ok=True)

[gemini-code-assist]

Accessing d['failed'] directly can raise a KeyError if the key is missing from the JSON fragment. Using .get('failed') is safer and aligns with defensive programming practices.

Suggested change

	status_icon = '❌ FAIL' if d['failed'] else '✅ PASS'
	status_icon = '❌ FAIL' if d.get('failed') else '✅ PASS'

[gemini-code-assist]

If any of the expected keys (like platform, build_result, etc.) are missing from the JSON fragment, d.get(...) will return None, which will print as the string 'None' in the Markdown table. Providing fallback defaults (like 'Unknown' or '-') improves the visual presentation of the report.

Suggested change

	markdown_lines.append(
	f"| {d.get('platform')} | {d.get('build_result')} |"
	f" {d.get('on_host_test_result')} |"
	f" {d.get('on_device_test_result')} | {test_summary} |"
	f' {status_icon} |')
	markdown_lines.append(
	f"| {d.get('platform', 'Unknown')} | {d.get('build_result', '-')} |"
	f" {d.get('on_host_test_result', '-')} |"
	f" {d.get('on_device_test_result', '-')} | {test_summary} |"
	f' {status_icon} |')