Make --list-sources respect --output-type and --output-file

[rjeffman]

Refactor list_sources() to return structured data instead of printing directly to stdout. Route output through the output plugin system so that --output-type controls the format (json, human, prometheus) and --output-file controls the destination.

This allows users to get machine-readable JSON output of available sources and checks, making it easier to programmatically query what checks are available.

The human output format remains unchanged from the original plain text format. JSON produces an array of {source, checks} objects. Prometheus raises an error as source listing has no meaningful metric representation.

Assisted-by: Claude Sonnet 4.5 noreply@anthropic.com

Summary by Sourcery

Route source listing through the output plugin system so it respects configured output type and destination.

New Features:

• Support structured source listing output via JSON and human output plugins, controlled by --output-type and --output-file.

Enhancements:

• Refactor list_sources() to return structured source/check data instead of printing directly to stdout.
• Disallow using --list-sources with Prometheus output and surface a clear error message from both the CLI and Prometheus output plugin.

Tests:

• Add a CLI options test asserting that --list-sources with --output-type=prometheus raises a ValueError.

[sourcery-ai]

Hey - I've left some high level feedback:

• Calling output.render_source_list with the Prometheus output will now raise a RuntimeError that appears unhandled at the CLI level; consider either blocking --output-type=prometheus when --list-sources is used or catching this exception and turning it into a clean user-facing error message/exit code.
• The new list_sources now returns a list of dicts built from a dict and then iterated, which preserves insertion order but is an implicit dependency on dict ordering; if stable or sorted output is important (especially for JSON consumers) consider explicitly sorting by source and/or check name before returning.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- Calling `output.render_source_list` with the Prometheus output will now raise a `RuntimeError` that appears unhandled at the CLI level; consider either blocking `--output-type=prometheus` when `--list-sources` is used or catching this exception and turning it into a clean user-facing error message/exit code.
- The new `list_sources` now returns a list of dicts built from a `dict` and then iterated, which preserves insertion order but is an implicit dependency on dict ordering; if stable or sorted output is important (especially for JSON consumers) consider explicitly sorting by source and/or check name before returning.

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[sourcery-ai]

Hey - I've left some high level feedback:

• In Human.generate_source_list, the check lines now have three leading spaces instead of two (' %s' vs previously " ", name); if consumers rely on the exact human-readable layout, consider adjusting the indentation to truly preserve the legacy format as claimed.
• The Prometheus backend both raises at generate_source_list and is preemptively rejected in parse_options; consider picking a single enforcement layer (either only the option validation or only the backend RuntimeError) to avoid redundant checks and keep the responsibility boundary clearer.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- In `Human.generate_source_list`, the check lines now have three leading spaces instead of two (`'   %s'` vs previously `"  ", name`); if consumers rely on the exact human-readable layout, consider adjusting the indentation to truly preserve the legacy format as claimed.
- The Prometheus backend both raises at `generate_source_list` and is preemptively rejected in `parse_options`; consider picking a single enforcement layer (either only the option validation or only the backend RuntimeError) to avoid redundant checks and keep the responsibility boundary clearer.

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}