Gradio UI

The Gradio entry point is src/quran_muaalem/gradio_app.py and is exposed as a CLI script quran-muaalem-ui in pyproject.toml.

What the UI does

Key functions in src/quran_muaalem/gradio_app.py:

• process_audio(...)
  • Loads audio via librosa.load.
  • Builds a reference phonetic script using quran_phonetizer.
  • Runs Muaalem inference and returns HTML output.
• update_uthmani_ref(...)
  • Retrieves the Uthmani reference text via quran_transcript.Aya.
• create_gradio_input_for_field(...)
  • Builds UI inputs from MoshafAttributes.model_fields.

Typical workflow (UI)

1. Choose Sura and Ayah.
2. Select the start word index and number of words.
3. Upload or record audio (16 kHz preferred).
4. Click “Analyze” to get phoneme + sifat comparison.

If the word span cuts a Uthmani word, the UI warns with PartOfUthmaniWord and asks you to adjust the span.

How the UI is launched

app.launch(server_name="0.0.0.0", share=True)

• To disable public share links, set share=False in main().
• To bind a different port or interface, edit the app.launch(...) call.

Run the UI locally

pip install "quran-muaalem[ui]"
quran-muaalem-ui

The CLI points to quran_muaalem.gradio_app:main (see pyproject.toml).

Known limitations

• The UI performs non‑streaming inference (full audio at once).
• Performance depends on GPU availability and audio length.
• If you need batch processing, use the Python API instead.