[Copilot]

This PR adds MakHamTTS integration to PyThaiTTS, using the Paradee Thai INT8 ONNX model with FastThaiG2P phonemization. It also packages the required model artifacts so pretrained="makhamtts" works out of the box.

• New pretrained backend: makhamtts

  • Added pythaitts/pretrained/makham_tts.py with ONNXRuntime inference flow aligned with MakHamTTS run.py (G2P -> ipa_to_kokoro -> vocab ids -> ONNX).
  • Supports return_type="waveform" and file output.
  • Adds explicit supported voice handling (paradee_thai) with compatibility mapping from legacy defaults.

• Core TTS integration

  • Extended model loading in pythaitts/__init__.py to support:
    • TTS(pretrained="makhamtts")
    • TTS(pretrained="MakHamTTS")
  • Routed tts() calls to the new backend with correct speaker normalization.

• Packaged model artifacts

  • Included MakHamTTS assets under pythaitts/pretrained/makhamtts/:
    • paradee_thai_int8.onnx
    • config.json
  • Updated setup.py package data so artifacts are shipped with the package.

• Docs and focused coverage

  • Updated README model usage examples to include MakHamTTS.
  • Added targeted integration tests in tests/test_makhamtts.py.

from pythaitts import TTS

tts = TTS(pretrained="makhamtts")
wav = tts.tts("สวัสดีค่ะ ยินดีต้อนรับสู่บริการของเรา",
              speaker_idx="paradee_thai",
              return_type="waveform")

• Fixes Add MakHamTTS #33