Add wildcard ～ support for Japanese grammar patterns

[RadiantSol]

Only drawbacks to this implementation is the addition of potentially a lot more lookups and the inability to scan for X～Y～Z entries.

AI Summary

• Added insertWildcard text processor for Japanese that generates wildcard variants of grammar patterns
• Supports X～Y patterns (e.g., "いくら騒いでも" → "いくら～でも")
• Generates up to 51 variants per input string to avoid excessive processing
• Integrated new processor into Japanese language descriptor pipeline
• Added comprehensive test suite with 91 test cases covering various pattern scenarios
• Updated TypeScript type definitions

[codspeed-hq]

Merging this PR will degrade performance by 62.53%

❌ 1 regressed benchmark
✅ 4 untouched benchmarks
⏩ 2 skipped benchmarks¹

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
❌	Translator.prototype.findTerms - (n=47)	123.5 ms	329.6 ms	-62.53%

---

_{Comparing RadiantSol:claude/wizardly-meitner (2c66da9) with master (530d165)}

Footnotes

1. 2 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[keanu-thakalath]

maybe we could avoid the performance hit by doing this in the deinflection layer instead? a new wildcardInflection transform with empty conditionsIn and conditionsOut only emits candidates for known patterns instead of 51x before transforming.

the only issue is we'd have to start maintaining a hardcoded list of known patterns, or create a new system for dynamically creating rules from imported dictionaries.

are you using a dictionary that already has these wildcard entries? i don't see any in jitendex