init

.gitignore

@@ -3,4 +3,4 @@ build
 preprocess
 download
 __pycache__
-.DS_Store
+.DS_Store

download_audio.py

@@ -6,13 +6,12 @@ import gzip
 import subprocess
 from os.path import join as p_join
 
-import torch
 from tqdm import tqdm
 from multiprocessing import Pool
 from typing import Optional, Dict
 from glob import glob
+import librosa
 
-import julius
 import pandas as pd
 import soundfile as sf
 from datasets import Dataset, Audio, DatasetDict
@@ -129,11 +128,8 @@ def get_audio(dataframe: pd.DataFrame):
                         wav = wav[:, 0]
                     wav = wav[:int(end/sampling_rate * sr) + sr]
                     if sr != sampling_rate:
-                        if sr not in resampler:
-                            resampler[sr] = julius.ResampleFrac(sr, sampling_rate)
-                            print(f"resampler added: {len(resampler)}")
                         print(f"RESAMPLING: {wav.shape} length audio")
-                        wav = resampler[sr](torch.as_tensor(wav)).numpy()
+                        wav = librosa.resample(wav, orig_sr=sr, target_sr=sr)
                     if start < end <= len(wav):
                         sf.write(features[f"{side}.path"], wav[start:end], sampling_rate)
                     else:

main.sh

@@ -20,6 +20,15 @@ python -c 'n=10; import os; from glob import glob; tmp = [int(os.path.basename(i
 python -c 'n=10; import os; from glob import glob; tmp = [int(os.path.basename(i).split(".")[0]) for i in glob("download/audio/enA-jaA/jaA/*")]; print(len([x for x in tmp if (n-1) * 2500 <= x < n * 2500]))'
 
 python -c 'file_name="tmp.mp3"; from datasets import Audio; a=Audio(); wav=a.decode_example({"path": file_name, "bytes": None}); print(wav)'
+# test
+export N_POOL=1
+export DATASET_ID=1
+export DIRECTION="enA-jaA"
+export LINE_NO_START=$(((DATASET_ID-1) * 2500))
+export LINE_NO_END=$((DATASET_ID * 2500))
+echo ${LINE_NO_START}
+python download_audio.py
+
 ####################
 # enA-jaA: 718_606 #
 ####################