File size: 1,840 Bytes
de01f6a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 | """Write dataset manifests: metadata.csv, dataset.jsonl, tts_metadata.csv."""
from __future__ import annotations
import csv
from pathlib import Path
from typing import Iterable
from video_vec2wav2_tokenizer.utils.io import DatasetEntry, ensure_dir, write_jsonl
from video_vec2wav2_tokenizer.utils.logging import get_logger
logger = get_logger(__name__)
def write_metadata_csv(path: str | Path, entries: Iterable[DatasetEntry]) -> int:
"""Write ``metadata.csv`` in ``filename|text`` format (pipe-separated)."""
ensure_dir(Path(path).parent)
count = 0
with Path(path).open("w", encoding="utf-8", newline="") as fh:
writer = csv.writer(fh, delimiter="|", quoting=csv.QUOTE_MINIMAL)
for entry in entries:
filename = Path(entry.audio).name
writer.writerow([filename, entry.text])
count += 1
logger.info("Wrote %d rows to %s", count, path)
return count
def write_dataset_jsonl(path: str | Path, entries: Iterable[DatasetEntry]) -> int:
"""Write ``dataset.jsonl`` with audio/text/duration rows."""
n = write_jsonl(path, (e.to_jsonl_dict() for e in entries))
logger.info("Wrote %d rows to %s", n, path)
return n
def write_tts_metadata_csv(path: str | Path, entries: Iterable[DatasetEntry]) -> int:
"""Write ``tts_metadata.csv`` in ``filename|speaker|text`` format."""
ensure_dir(Path(path).parent)
count = 0
with Path(path).open("w", encoding="utf-8", newline="") as fh:
writer = csv.writer(fh, delimiter="|", quoting=csv.QUOTE_MINIMAL)
for entry in entries:
filename = Path(entry.audio).name
speaker = entry.speaker or "speaker_001"
writer.writerow([filename, speaker, entry.text])
count += 1
logger.info("Wrote %d rows to %s", count, path)
return count
|