139_spanish-voice-trainer/anki_exporter.py
2026-06-29 22:08:39 +10:00

141 lines
No EOL
6.9 KiB
Python

# anki_exporter.py
import os
import tempfile
import asyncio
import genanki
import edge_tts
import shutil
import database
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.
"""
# Incrementing both forces a completely clean slate for both layout and deck container
model_id = 1684329015
deck_id = 1684329015
# Global Configuration Pace Resolver Mapping
settings = database.load_all_settings() or {}
config_speed = settings.get("tts_playback_speed", "1.0")
# Transform numeric string floats (e.g., 1.2) into Edge-TTS percentage strings (e.g., +20%)
try:
pct = int((float(config_speed) - 1.0) * 100)
rate_string = f"{'+' if pct >= 0 else ''}{pct}%"
except Exception:
rate_string = "+0%"
anki_model = genanki.Model(
model_id,
'Spanish Bidirectional Multi-Note HTML Model',
fields=[
{'name': 'EnglishText'},
{'name': 'SpanishText'},
{'name': 'Notes'},
{'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>'
'<div style="font-family: Arial; font-size: 14px; text-align: center; color: #34495E;">{{Notes}}</div>'
'{{#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>'
'<div style="font-family: Arial; font-size: 14px; text-align: center; color: #34495E;">{{Notes}}</div>'
'{{#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_text = record["en_text"]
es_text = record["es_text"]
gender_flag = record.get("gender", "Female")
# 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()]
# Dynamic HTML Notes Mapping Builder
notes_html_parts = []
if record.get('notes') and record['notes'].strip():
notes_html_parts.append(f"<div style='margin-bottom: 4px;'>{record['notes'].strip()}</div>")
if record.get('source_context') and record['source_context'].strip():
notes_html_parts.append(f"<div style='font-size: 12px; color: #95A5A6; font-style: italic;'>Context: {record['source_context'].strip()}</div>")
notes_html = "".join(notes_html_parts)
anki_notes_html = f"<div>{record['anki_notes'].strip()}</div>" if record.get('anki_notes') else ""
# Standard Unique Media Filenames
en_audio_filename = f"edge_en_{idx}_{model_id}.mp3"
es_audio_filename = f"edge_es_{idx}_{model_id}.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 loop.run_until_complete(generate_edge_audio(en_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 loop.run_until_complete(generate_edge_audio(es_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 = ""
note = genanki.Note(
model=anki_model,
fields=[en_text, es_text, notes_html, 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)