feat: warn on misspelled language suffix in prompt filenames

[Serhan-Asad]

Summary

• Fixes default_language from .pddrc being ignored during language detection
• Adds fuzzy matching to detect misspelled language suffixes in prompt filenames (e.g., new_typscript.prompt) and warns the user before silently falling back to default_language
• Uses Levenshtein distance ≤ 2 with minimum token length of 4 to avoid false positives on short language names (r, d, go)
• Refactors _is_known_language to share the language set via _get_known_languages helper (no duplication)

Fixes #451

Test plan

• 23 new tests covering _get_known_languages, _levenshtein_distance, _closest_known_language, and end-to-end warning behavior
• All 90 construct_paths tests pass
• Full test suite passes (773 passed, 7 skipped, 1 pre-existing failure unrelated)

[Copilot]

Pull request overview

This PR enhances language detection in prompt filenames by adding fuzzy matching to detect and warn about misspelled language suffixes (e.g., "typscript" instead of "typescript"). Additionally, it fixes a bug where the default_language from .pddrc was being ignored during language detection.

Changes:

• Refactored language validation to use a shared _get_known_languages() helper to eliminate code duplication
• Added Levenshtein distance calculation and fuzzy matching to detect misspelled language suffixes with a threshold of ≤2 edits
• Implemented fallback to default_language from .pddrc when no language can be determined from other sources

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
pdd/construct_paths.py	Refactored _is_known_language to use _get_known_languages(), added _levenshtein_distance and _closest_known_language functions for fuzzy matching, and integrated misspelling warnings with default language fallback
tests/test_construct_paths.py	Added 23 new tests covering language detection helpers, fuzzy matching behavior, and end-to-end warning scenarios; patched _find_pddrc_file in existing tests to prevent configuration interference

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The new languages 'lisp', 'scheme', and 'ada' were added on the same line as 'prisma', 'lean', and 'agda', breaking the logical grouping pattern. Based on the removed comment, these appear to be different categories. Consider moving them to a separate line or adding a comment to clarify their grouping.

Suggested change

	'prisma', 'lean', 'agda', 'lisp', 'scheme', 'ada',
	'prisma', 'lean', 'agda',
	'lisp', 'scheme', 'ada',

[Copilot]

This test assumes that exact matches return the language itself, but the comment suggests this case should be handled by _is_known_language instead. Consider testing the actual expected behavior where _closest_known_language is only called for non-exact matches, or explicitly document that returning exact matches is intentional fallback behavior.