feat(audio): add chunking_strategy field to AudioRequest

[alnah]

Summary

Adds chunking_strategy field to AudioRequest and full speaker diarization support:

• ChunkingStrategy *TranscriptionChunkingStrategy field in AudioRequest
• AudioChunkingStrategyAuto and AudioChunkingStrategyServerVAD constants
• AudioResponseFormatDiarizedJSON response format
• CreateDiarizedTranscription method with DiarizedAudioResponse return type
• DiarizedSegment and AudioUsage types for response parsing

Closes #1101

Usage

resp, err := client.CreateDiarizedTranscription(ctx, openai.AudioRequest{
    Model:    "gpt-4o-transcribe-diarize",
    FilePath: "audio.mp3",
    Format:   openai.AudioResponseFormatDiarizedJSON,
    ChunkingStrategy: &openai.TranscriptionChunkingStrategy{
        Type: openai.AudioChunkingStrategyAuto,
    },
})

Test plan

• go test -race -cover ./... passes (99.5% coverage)
• go vet ./... passes
• staticcheck ./... passes
• golangci-lint run passes (only pre-existing issues in unrelated files)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.62%. Comparing base (5d7a276) to head (f98e487).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1102      +/-   ##
==========================================
+ Coverage   99.59%   99.62%   +0.03%     
==========================================
  Files          34       34              
  Lines        2209     1876     -333     
==========================================
- Hits         2200     1869     -331     
+ Misses          6        4       -2     
  Partials        3        3              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.