Normalize malformed DOI month fields in bibbot ingestion

[Copilot]

Bibbot failed while ingesting the new DOI entry (10.2514/6.2026-4026) due to malformed month values from DOI BibTeX payloads (e.g., 'june). This change hardens DOI parsing so malformed/bare month tokens no longer break issue_to_bibtex.py.

• Problem addressed

  • DOI-provided BibTeX can contain month fields in non-parseable forms (month = june,, month = 'june,), which caused bibtexparser.loads(...) to raise and stop ingestion.

• Code changes

  • In bin/issue_to_bibtex.py (DOI branch), added a pre-parse normalization step that rewrites month assignments to brace-wrapped BibTeX values before bibtexparser.loads(...).
  • Scope is intentionally narrow: only month = ... lines are normalized; no other fields or workflow behavior are changed.

• Normalization example

  bib = re.sub(
      r'(^\s*month\s*=\s*)(?:[\'"]?)([A-Za-z]+)(?:[\'"]?)\s*,',
      r'\1{\2},',
      bib,
      flags=re.MULTILINE,
  )

This keeps bibbot resilient to DOI metadata formatting variance while preserving existing entry generation flow.

[tlunet]

Wouldn't we want at least a small unit test to check a REGEX that is more and more complex (and maybe no one understand ...) ?