139_spanish-voice-trainer/anki_exporter.py

# anki_exporter.py
import os
import tempfile
import asyncio
import time
import genanki
import edge_tts
import database
from tts_utils import parse_text_for_edgetts, get_configured_tts_rate

async def generate_edge_audio(text, voice, output_path, rate_modifier="+0%"):
    """Asynchronously streams data packages via the Microsoft Edge API pipeline."""
    try:
        communicate = edge_tts.Communicate(text=text, voice=voice, rate=rate_modifier)
        await communicate.save(output_path)
        return True
    except Exception as e:
        print(f"Edge-TTS synthesis anomaly: {e}")
        return False

def compile_anki_package(records, output_path, deck_name):
    """
    Compiles database records into a bidirectional card payload package.
    Resolves voice models dynamically by gender selection parameters and applies
    global speed coefficient rates from the active configurations.
    Injects sort_order as field 0 for structured card sequencing.
    """
    # Generate deterministic positive 32-bit integers from deck name and model name
    # to avoid collisions across different decks while keeping imports stable
    deck_id = abs(hash(deck_name)) % (2**31)
    model_id = abs(hash("Spanish Bidirectional Multi-Note HTML Model")) % (2**31)

    # Unique timestamp prefix for media files to prevent overwriting prior exports in Anki
    run_prefix = int(time.time())

    # Global Configuration Pace Resolver Mapping
    settings = database.load_all_settings() or {}
    rate_string = get_configured_tts_rate(settings)

    anki_model = genanki.Model(
        model_id,
        'Spanish Bidirectional Multi-Note HTML Model',
        fields=[
            {'name': 'SortOrder'},     # Field 0: Anki primary sort field
            {'name': 'EnglishText'},
            {'name': 'SpanishText'},
            {'name': 'AnkiNotes'},
            {'name': 'EnglishAudio'},
            {'name': 'SpanishAudio'}
        ],
        templates=[
            {
                'name': 'Card 1: English ➔ Spanish',
                'qfmt': '<div style="font-family: Arial; font-size: 13px; font-weight: bold; color: #BDC3C7; text-align: center; letter-spacing: 1px;">TRANSLATE TO SPANISH:</div><br>'
                        '<div style="font-family: Arial; font-size: 20px; text-align: center; color: #34495E;"><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>'
                        '{{#AnkiNotes}}<div style="font-family: Arial; font-size: 13px; text-align: center; color: #8E44AD; border-top: 1px dashed #E5E7E9; padding-top: 6px; margin-top: 6px;"><b>Anki Meta:</b> {{{AnkiNotes}}}</div>{{/AnkiNotes}}<br>'
                        '<div style="text-align: center;">{{SpanishAudio}}</div>',
            },
            {
                'name': 'Card 2: Spanish ➔ English',
                'qfmt': '<div style="font-family: Arial; font-size: 13px; font-weight: bold; color: #E67E22; text-align: center; letter-spacing: 1px;">TRANSLATE TO ENGLISH:</div><br>'
                        '<div style="font-family: Arial; font-size: 26px; text-align: center; color: #2980B9; font-weight: bold;"><b>{{SpanishText}}</b></div>'
                        '<div style="display:none;">{{SpanishAudio}}</div>',
                'afmt': '{{FrontSide}}<hr id="answer">'
                        '<div style="font-family: Arial; font-size: 22px; text-align: center; color: #2C3E50; font-weight: 500;">{{EnglishText}}</div><br>'
                        '{{#AnkiNotes}}<div style="font-family: Arial; font-size: 13px; text-align: center; color: #8E44AD; border-top: 1px dashed #E5E7E9; padding-top: 6px; margin-top: 6px;"><b>Anki Meta:</b> {{{AnkiNotes}}}</div>{{/AnkiNotes}}<br>'
                        '<div style="text-align: center;">{{EnglishAudio}}</div>',
            }
        ]
    )

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

    try:
        loop = asyncio.get_running_loop()
    except RuntimeError:
        loop = asyncio.new_event_loop()
        asyncio.set_event_loop(loop)

    with tempfile.TemporaryDirectory() as tmpdir:
        for idx, record in enumerate(records):
            en_raw = record["en_text"]
            es_raw = record["es_text"]
            gender_flag = record.get("gender", "Female")
            
            # Read sort_order from database record (fallback to loop index if unset)
            raw_order = record.get("sort_order")
            order_val = int(raw_order) if raw_order is not None else (idx + 1)
            sort_order_str = f"{order_val:04d}"  # Zero-padded integer string for exact sorting
            
            # Sanitize text payloads for TTS engine
            en_tts_text = parse_text_for_edgetts(en_raw)
            es_tts_text = parse_text_for_edgetts(es_raw)
            
            # Map native neural voice files matching gender settings
            spanish_voice = "es-ES-AlvaroNeural" if gender_flag == "Male" else "es-ES-ElviraNeural"
            english_voice = "en-US-EmmaNeural"

            raw_tags = record.get("tags") or ""
            note_tags = [t.strip().replace(" ", "_") for t in raw_tags.split(",") if t.strip()]
            
            # Safely isolate the raw text/HTML data string down to Anki notes field
            anki_notes_html = record['anki_notes'].strip() if record.get('anki_notes') else ""

            # Unique Media Filenames combining execution timestamp and index
            en_audio_filename = f"edge_en_{run_prefix}_{idx}.mp3"
            es_audio_filename = f"edge_es_{run_prefix}_{idx}.mp3"
            
            en_audio_path = os.path.join(tmpdir, en_audio_filename)
            es_audio_path = os.path.join(tmpdir, es_audio_filename)

            # English Audio Synthesis
            if en_tts_text.strip() and loop.run_until_complete(generate_edge_audio(en_tts_text, english_voice, en_audio_path, rate_string)):
                media_files_to_pack.append(en_audio_path)
                en_audio_field = f"[sound:{en_audio_filename}]"
            else:
                en_audio_field = ""

            # Spanish Audio Synthesis
            if es_tts_text.strip() and loop.run_until_complete(generate_edge_audio(es_tts_text, spanish_voice, es_audio_path, rate_string)):
                media_files_to_pack.append(es_audio_path)
                es_audio_field = f"[sound:{es_audio_filename}]"
            else:
                es_audio_field = ""

            # Clean sequential matching fields array matching schema mapping above
            note = genanki.Note(
                model=anki_model,
                fields=[sort_order_str, en_raw, es_raw, anki_notes_html, en_audio_field, es_audio_field],
                tags=note_tags
            )
            deck.add_note(note)

        # Build Package while media assets are guaranteed contextually active inside tmpdir
        package = genanki.Package(deck)
        package.media_files = media_files_to_pack
        package.write_to_file(output_path)