139_spanish-voice-trainer/anki_exporter.py

# anki_exporter.py
import os
import tempfile
import subprocess
import shutil
import genanki

def compile_anki_package(records, output_path, deck_name):
    """
    Compiles database records into an .apkg package using native macOS TTS.
    Uses 'Monica' for Spanish targets and the default premium system voice for English.
    """
    # Create a unique random Model ID and Deck ID for genanki
    model_id = 1684329011
    deck_id = 1684329012

    # Define the Anki Card Layout structure with audio fields
    anki_model = genanki.Model(
        model_id,
        'Spanish Voice Trainer Model',
        fields=[
            {'name': 'EnglishText'},
            {'name': 'SpanishText'},
            {'name': 'Notes'},
            {'name': 'EnglishAudio'},
            {'name': 'SpanishAudio'}
        ],
        templates=[
            {
                'name': 'Card 1',
                'qfmt': '<div style="font-family: Arial; font-size: 20px; text-align: center; color: #34495E;">'
                        'Translate to Spanish:<br><br><b>{{EnglishText}}</b></div><div style="display:none;">{{EnglishAudio}}</div>',
                'afmt': '{{FrontSide}}<hr id="answer">'
                        '<div style="font-family: Arial; font-size: 28px; text-align: center; color: #2980B9; font-weight: bold;">'
                        '{{SpanishText}}</div><br>'
                        '<div style="font-family: Arial; font-size: 14px; text-align: center; color: #7F8C8D; font-style: italic;">'
                        '{{Notes}}</div><br>'
                        '<div style="text-align: center;">{{SpanishAudio}}</div>',
            },
        ]
    )

    deck = genanki.Deck(deck_id, deck_name)
    media_files = []

    # Process all records inside a secure temporary directory workspace
    with tempfile.TemporaryDirectory() as tmpdir:
        for idx, record in enumerate(records):
            en_text = record["en_text"]
            es_text = record["es_text"]
            
            # Extract and parse tags for native Anki tag support (split comma-separated text)
            raw_tags = record.get("tags") or ""
            note_tags = [t.strip().replace(" ", "_") for t in raw_tags.split(",") if t.strip()]
            
            # Build text for visual notes area, appending tags if they exist
            notes_parts = []
            if record.get('source_context'):
                notes_parts.append(f"Context: {record['source_context']}")
            if record.get('notes'):
                notes_parts.append(f"Notes: {record['notes']}")
            if raw_tags:
                notes_parts.append(f"Tags: {raw_tags}")
                
            notes_display_text = " | ".join(notes_parts)

            # Generate unique filenames for the media assets
            en_audio_filename = f"en_audio_{idx}.mp3"
            es_audio_filename = f"es_audio_{idx}.mp3"
            
            en_audio_path = os.path.join(tmpdir, en_audio_filename)
            es_audio_path = os.path.join(tmpdir, es_audio_filename)

            try:
                # 1. Render English Audio using native macOS text-to-speech engine
                subprocess.run(
                    ["say", "-o", en_audio_path, "--data-format=Iface", en_text],
                    check=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL
                )
                media_files.append(en_audio_path)
                en_audio_field = f"[sound:{en_audio_filename}]"
            except Exception:
                en_audio_field = ""

            try:
                # 2. Render Spanish Audio explicitly targeting the Monica voice profile
                subprocess.run(
                    ["say", "-v", "Monica", "-o", es_audio_path, "--data-format=Iface", es_text],
                    check=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL
                )
                media_files.append(es_audio_path)
                es_audio_field = f"[sound:{es_audio_filename}]"
            except Exception:
                es_audio_field = ""

            # Build the card note stack, natively injecting the parsed Anki tags list
            note = genanki.Note(
                model=anki_model,
                fields=[en_text, es_text, notes_display_text, en_audio_field, es_audio_field],
                tags=note_tags
            )
            deck.add_note(note)

        # Build package collection mapping archive pipelines
        package = genanki.Package(deck)
        package.media_files = media_files
        package.write_to_file(output_path)