feat: exclude opted-out repos from the full pipeline (analysis + issue creation) with wildcard support

[Copilot]

Opted-out repositories are now excluded at both the metacheck analysis stage and the issue-creation stage, so no compute is wasted analyzing repos that will be skipped anyway.

Changes

• metacheck_wrapper.py: Added --opt-outs option (defaults to .opt-outs in the current directory). When the file exists and the input is a JSON file, a _filter_opt_out_repos() helper strips opted-out repos into a temp file before passing it to metacheck (temp file cleaned up after the run). If the default .opt-outs file is absent, filtering is silently skipped. Opt-outs entries support * as a wildcard (e.g. https://github.com/SoftwareUnderstanding/* excludes all repos in that org); URL characters such as . are always treated as literals so patterns are safe to write without escaping. Invalid patterns emit a warning on stderr.

• pipeline.py: Passes opt_outs_file as --opt-outs to the metacheck step, so a single file controls both analysis and issue-creation filtering.

• tests/test_metacheck_wrapper.py (new): Covers exact URL removal, trailing-slash normalization, org-level wildcard patterns, wildcard suffix patterns, literal-dot behaviour, extra JSON key preservation, empty opt-outs list, invalid format error, and missing default file.

• tests/test_pipeline.py: Updated expected metacheck args to include --opt-outs.

Example — standalone metacheck usage

sw-metadata-bot metacheck \
  --input assets/opt-ins.json \
  --opt-outs assets/opt-outs.json \
  --pitfalls-output pitfalls_outputs/

Example — .opt-outs file with wildcard patterns

{
  "repositories": [
    "https://github.com/SoftwareUnderstanding/*",
    "https://github.com/some-org/specific-repo"
  ]
}

Place a .opt-outs file in the working directory and it will be picked up automatically. The run-pipeline command passes the same file through to both stages automatically.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>handling white list / black list</issue_title>
<issue_description>the bot should handle a black list of repos that requested to unsubscribe or where we don't want to run the bot

this is a simple list of repos in a file that is checked when running the bot</issue_description>

<agent_instructions>open a PR to implement this new feature</agent_instructions>

Comments on the Issue (you are @copilot in this section)

• Fixes handling white list / black list #42

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.

[vuillaut]

@copilot

• rename --blacklist-file to --blacklist
• --blacklist should be set to True by default with a .blacklist file in the root repo as default. if no file is found, then consider blacklist is empty. user can specifiy a different file
• blacklist shuold handle regex so that we can specify all repos in an organisation with a single line, e.g. https://github.com/SoftwareUnderstanding/* would skip all repos in this orga

[Copilot]

@copilot

• rename --blacklist-file to --blacklist
• --blacklist should be set to True by default with a .blacklist file in the root repo as default. if no file is found, then cons...

Done in f4c4809:

• --blacklist-file → --blacklist: option renamed on metacheck_command and in pipeline.py.
• Default .blacklist file: the option defaults to .blacklist in the current directory. If the file doesn't exist (the common case when no blacklist is configured), filtering is silently skipped — no error.
• Wildcard/pattern support: each entry in the repositories list is treated as a pattern where * acts as a wildcard. The pattern is first passed through re.escape() so that URL characters like . are always treated as literals, then the escaped \* is replaced with .* to restore wildcard behaviour. This means https://github.com/SoftwareUnderstanding/* correctly excludes all repos in that org, without . in github.com accidentally matching other characters. Invalid patterns emit a warning on stderr rather than silently failing.

[vuillaut]

@copilot rename "blacklist" into "opt-outs"

[Copilot]

@copilot rename "blacklist" into "opt-outs"

Done in 7639cd6. All "blacklist" terminology has been replaced with "opt-outs":

• --blacklist → --opt-outs (CLI option)
• Default file .blacklist → .opt-outs
• _is_blacklisted() → _is_opted_out()
• _filter_blacklisted_repos() → _filter_opt_out_repos()
• Echo/warning messages updated accordingly
• Tests and pipeline args updated to match

[vuillaut]

Already implemented :D