download_audio.py

import json
import os
import tarfile
import zipfile
import gzip
import subprocess
from os.path import join as p_join
from tqdm import tqdm
from multiprocessing import Pool
from typing import Optional

import pandas as pd

# dataset config
url_metadata_dict = {
    "enA-jaA": "https://dl.fbaipublicfiles.com/seamless/data/seamless_align_nov2023_extension/seamless.dataset.metadata.public.enA-jaA.tsv.gz",
    "enA-jpn": "https://dl.fbaipublicfiles.com/seamless/data/seamless.dataset.metadata.public.enA-jpn.withduration.tsv.gz"
}
direction = os.getenv("DIRECTION", "enA-jaA")
sides = set(direction.split("-"))
# processor config
n_pool = int(os.getenv("N_POOL", 8))
wget_max_retry = os.getenv("MAX_RETRY", "1")
wget_timeout = os.getenv("TIMEOUT", "20")
line_no_start = int(os.getenv("LINE_NO_START", 0))
line_no_end = int(os.getenv("LINE_NO_END", 500000))


def wget(url: str, cache_dir: str, filename: Optional[str] = None):
    os.makedirs(cache_dir, exist_ok=True)
    filename = os.path.basename(url) if not filename else filename
    output_file = p_join(cache_dir, filename)
    subprocess.run(["wget", url, "-O", output_file, "--tries", wget_max_retry, "--timeout", wget_timeout])
    if not os.path.exists(output_file):
        return False
    if output_file.endswith('.tar.gz') or output_file.endswith('.tgz') or output_file.endswith('.tar'):
        if output_file.endswith('.tar'):
            tar = tarfile.open(output_file)
        else:
            tar = tarfile.open(output_file, "r:gz")
        tar.extractall(cache_dir)
        tar.close()
        os.remove(output_file)
    elif output_file.endswith('.gz'):
        with gzip.open(output_file, 'rb') as f:
            with open(output_file.replace('.gz', ''), 'wb') as f_write:
                f_write.write(f.read())
        os.remove(output_file)
    elif output_file.endswith('.zip'):
        with zipfile.ZipFile(output_file, 'r') as zip_ref:
            zip_ref.extractall(cache_dir)
        os.remove(output_file)
    return True


def get_metadata():
    url_metadata = url_metadata_dict[direction]
    meta_data_filename = os.path.basename(url_metadata).replace(".gz", "")
    cache_dir_metadata = p_join("download", "meta")
    if not os.path.exists(p_join(cache_dir_metadata, meta_data_filename)):
        assert wget(url_metadata, cache_dir=cache_dir_metadata)
    df = pd.read_csv(p_join(cache_dir_metadata, meta_data_filename), sep=r'[\t\s]', header=None)[[0, 2, 6, 9, 10, 11, 12]]
    df.columns = ["id", "url", "text_lid_score", "laser_score", "direction", "side", "line_no"]
    assert len(df["direction"].unique()) == 1
    df.pop("direction")
    return df.sort_values(by=["line_no", "side"])


def get_audio(dataframe: pd.DataFrame):
    cache_dir_audio = p_join("download", "audio", direction)
    cache_dir_feature = p_join("download", "feature", direction)
    features = {"line_no": dataframe.pop('line_no').values[0]}
    for side, df in dataframe.groupby("side"):
        df.pop("side")
        features.update({f"{side}.{k}": v for k, v in df.iloc[0].to_dict().items()})
        features[f"{side}.path"] = p_join(cache_dir_audio, os.path.basename(features[f"{side}.url"]))
        if not os.path.exists(features[f"{side}.path"]):
            if not wget(features[f"{side}.url"], filename=features[f"{side}.path"], cache_dir=cache_dir_audio):
                return False
    with open(cache_dir_feature, "w") as f:
        json.dump(features, f)
    return True


def process_dataset():
    df_metadata = get_metadata()
    print(f"metadata: {len(df_metadata)}")
    inputs = [g for line_no, g in df_metadata.groupby("line_no") if line_no_start <= line_no < line_no_end]
    print(f"filtered unique lines: {len(inputs)}")
    inputs = [g for g in inputs if len(g) == 2]
    print(f"removed != 2: {len(inputs)}")
    inputs = [g for g in inputs if len(g["side"].unique()) == 2 and set(g["side"].unique()) == sides]
    print(f"removed side != 2: {len(inputs)}")
    if n_pool == 1:
        for g in tqdm(inputs, total=len(inputs)):
            if not get_audio(g):
                print(f"failed:\n{g['url']}")
    else:
        with Pool(n_pool) as pool:
            pool.starmap(get_audio, tqdm(inputs, total=len(inputs)))


if __name__ == '__main__':
    process_dataset()