fix: guard against invalid orderRank data to prevent studio crashes

[ChrisNolan]

When an orderRank field contains invalid or corrupted data (non-string values, malformed lexorank strings), the LexoRank parser would throw an unhandled error and crash the studio.

Changes:

• New: Created parseOrderRank() helper in src/helpers/parseOrderRank.ts that safely parses LexoRank strings with try-catch error handling and
  configurable fallback values
• Updated: reorderDocuments.ts to use parseOrderRank() when accessing
  rank values from entities
• Updated: initialRank.ts to use parseOrderRank() when generating
  initial ranks
• Added: Console warnings when invalid orderRank data is encountered to
  help users identify corrupted documents
• Tests: Unit tests for parseOrderRank() covering valid parsing, non-string values, and parse failures. Regression tests for reorderDocuments and initialRank with bad data. All 14 tests passing.

Behavior:

• Invalid orderRank values now gracefully fall back to safe defaults
  (LexoRank.min() or LexoRank.max())
• Console warnings appear when bad data is encountered:
  [orderable-document-list] Invalid orderRank value...
• Studio remains stable and functional even with corrupted data
• Users can query *[_type == "yourType" && !isString(orderRank)] to find and fix corrupted documents

Closes #89

Description

What to review

Testing

[ChrisNolan]

Note there was code in the DocumentListWrapper to try and stop the display of everything when a string was found -- but it never got there because of 'when' in the load order the plugin was firing.

<code>{ORDER_FIELD_NAME}</code> field on Schema <code>{type}</code> must be{' '}

[stipsan]

Thank you! This is a great find 💖