ci: add hint validation checks

[cjdcordeiro]

Proposed changes

Following canonical/chisel#263, this PR introduces:

• validation of hints on every PR (according to the specification) - https://documentation.ubuntu.com/chisel/en/latest/reference/chisel-releases/slice-definitions/#slices-name-hint
• add a test-ci workflow to test our own CI scripts. This PR just initialized this workflow, testing only the validate_hints script being introduced, and should grow to cover the remaining CI scripts

Related issues/PRs

canonical/chisel#263

Checklist

• I have read the contributing guidelines
• I have tested my changes (see how)
• I have already submitted the CLA form

Additional Context

Example of a test-ci run: https://github.com/cjdcordeiro/chisel-releases/actions/runs/21833297147
Example of a PR with validation of hints: https://github.com/cjdcordeiro/chisel-releases/actions/runs/21853592833/job/63065485307?pr=35

[upils]

This is very nice, thanks!

[cjdcordeiro]

thanks @upils

applied two changes and answered all comments.

please have a look at the changes ;)

re-tested at https://github.com/cjdcordeiro/chisel-releases/actions/runs/21949859625/job/63397311896?pr=36 and https://github.com/cjdcordeiro/chisel-releases/actions/runs/21949954850/job/63397658295?pr=35

[upils]

Thanks!

[lczyk]

'request changes' but only due to volume of nitpicks / minor changes requested. overall happy with :)) ty!

[lczyk]

also

def not_too_long(text: str, max_length: int = 40) -> ErrorMessage | None:
    """Check that the text is not too long."""
    if len(text) > max_length:
        return f"must be at most {max_length} characters long (found {len(text)})"
    return None


def only_single_spaces(text: str) -> ErrorMessage | None:
    """Check that the text does not contain multiple consecutive spaces."""
    if "  " in text:
        return "multiple consecutive spaces are not allowed"
    return None


def no_leading_trailing_spaces(text: str) -> ErrorMessage | None:
    """Check that the text does not have leading or trailing spaces."""
    if text != text.strip():
        return "leading and trailing spaces are not allowed"
    return None

???

[lczyk]

also

    def test_all_caps_acronyms(self, tmp_path):
        f = tmp_path / "sdf.yaml"
        yaml_content = """slices:
  with-acronym:
    hint: Configure the CPU and GPU settings
  
  starting-with-acronym:
    hint: HTTP and HTTPS protocols
"""
        f.write_text(yaml_content, encoding="utf-8")
        errors = validate_hints.validate_hints(str(f))
        assert len(errors) == 0

this is currently passing, which is great, but id add it as a regression test / placeholder if someone adds something to do with acronyms.. ? 🤔

from https://documentation.ubuntu.com/chisel/en/latest/reference/chisel-releases/slice-definitions/:

Uppercase for acronyms (e.g. HTTP not Http).

[cjdcordeiro]

also

def not_too_long(text: str, max_length: int = 40) -> ErrorMessage | None:
    """Check that the text is not too long."""
    if len(text) > max_length:
        return f"must be at most {max_length} characters long (found {len(text)})"
    return None


def only_single_spaces(text: str) -> ErrorMessage | None:
    """Check that the text does not contain multiple consecutive spaces."""
    if "  " in text:
        return "multiple consecutive spaces are not allowed"
    return None


def no_leading_trailing_spaces(text: str) -> ErrorMessage | None:
    """Check that the text does not have leading or trailing spaces."""
    if text != text.strip():
        return "leading and trailing spaces are not allowed"
    return None

???

• Chisel already checks for the length of the hint - see above
• disallowing consecutive spaces, although not in the policy, is a good sanitization step IMO. See cd34fce
• Trailing spaces are already disallowed in the existing validators

also

    def test_all_caps_acronyms(self, tmp_path):
        f = tmp_path / "sdf.yaml"
        yaml_content = """slices:
  with-acronym:
    hint: Configure the CPU and GPU settings
  
  starting-with-acronym:
    hint: HTTP and HTTPS protocols
"""
        f.write_text(yaml_content, encoding="utf-8")
        errors = validate_hints.validate_hints(str(f))
        assert len(errors) == 0

this is currently passing, which is great, but id add it as a regression test / placeholder if someone adds something to do with acronyms.. ? 🤔

from https://documentation.ubuntu.com/chisel/en/latest/reference/chisel-releases/slice-definitions/:

Uppercase for acronyms (e.g. HTTP not Http).

This is a tricky one. It's the same for

Semicolons to separate multiple unrelated fragments of information.

Both require context, and although the NLP might be able to do a good job at it, it won't do it with a high level of confidence. I guess we gotta live with manual enforcement of these for now

[lczyk]

ty for the changes! :)) still marking as 'request changes` but only pending the resolution of #898 (comment) + 2 nits. no further comments

[lczyk]

thanks! big +1 from me. given lots of changes fro my review, @upils could you have a quick look over? the biggest functionality change is the new no_consecutive_spaces validator. other than that it's "backend" changes.

( adding the 'ready to merge' label but not merging yet )

[upils]

LGTM, thanks!