init

download_audio.py

@@ -96,14 +96,18 @@ def get_audio(dataframe: pd.DataFrame):
 def process_dataset():
     df_metadata = get_metadata()
     print(f"metadata: {len(df_metadata)}, {line_no_start} --> {line_no_end}")
-    inputs = [g for line_no, g in df_metadata.groupby("line_no") if line_no_start <= line_no < line_no_end]
+    inputs = [
+        g for line_no, g in df_metadata.groupby("line_no")
+        if line_no_start <= line_no < line_no_end and not os.path.exists(p_join(cache_dir_feature, f'{line_no}.json'))
+    ]
     print(f"filtered unique lines: {len(inputs)}")
     if direction == "enA-jaA":
         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):
+            flag = get_audio(g)
+            if not flag:
                 print(f"failed:\n{g['url']}")
     else:
         with Pool(n_pool) as pool:

download_text.py

@@ -0,0 +1,32 @@
+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
+
+chunk_size = 10
+url = "https://dl.fbaipublicfiles.com/seamless/data/seamless.dataset.metadata.public.enA-jpn.withduration.tsv.gz"
+filename = os.path.basename(url)
+subprocess.run(["wget", url, "-O", filename])
+df = pd.read_csv(filename, sep='\t', header=None, dtype=str)
+df.columns = ["cc_warc", "cc_sha", "cc_document_url", "cc_lineno", "paragraph_digest", "sentence_digest", "text_lid_score", "laser_score", "direction", "side", "line_no"]
+df = df[df.side == "jpn"]
+df["cc_lineno"] = df["cc_lineno"].astype(int)
+df.sort_values(by=["cc_warc", "cc_sha", "cc_document_url", "cc_lineno"], inplace=True)
+batch_size = int(len(df)/chunk_size)
+start = 0
+end = batch_size
+index = 1
+while start != end:
+    df.iloc[start:end].to_csv(f"seamless.dataset.metadata.public.enA-jpn.withduration.reordered.batch_{index}.tsv", sep="\t", index=False, header=False)
+    index += 1
+    start = end
+    end += batch_size
+    end = min(len(df), end)

main.sh

@@ -2,7 +2,7 @@
 # enA-jaA: 718_606 #
 ####################
 export DIRECTION="enA-jaA"
-
+export MAX_RETRY=10
 export LINE_NO_START=0
 export LINE_NO_END=50000
 python download_audio.py
@@ -23,43 +23,33 @@ export LINE_NO_START=300000
 export LINE_NO_END=360000
 python download_audio.py
 
+######################
+# enA-jpn: 1_468_292 #
+######################
+# AUDIO
+export DIRECTION="enA-jpn"
 
-
-# server 2
-export LINE_NO_START=150000
-export LINE_NO_END=180000
-python download_audio.py
-
-export LINE_NO_START=180000
-export LINE_NO_END=210000
+export LINE_NO_START=0
+export LINE_NO_END=50000
 python download_audio.py
 
-export LINE_NO_START=210000
-export LINE_NO_END=240000
+export LINE_NO_START=50000
+export LINE_NO_END=100000
 python download_audio.py
 
-export LINE_NO_START=240000
-export LINE_NO_END=270000
+export LINE_NO_START=100000
+export LINE_NO_END=150000
 python download_audio.py
 
-export LINE_NO_START=270000
+export LINE_NO_START=150000
 export LINE_NO_END=300000
 python download_audio.py
 
-# server 3
 export LINE_NO_START=300000
-export LINE_NO_END=330000
-python download_audio.py
-
-export LINE_NO_START=330000
 export LINE_NO_END=360000
 python download_audio.py
 
-
-####################
-# enA-jpn: 718_606 #
-####################
-
+# TEXT
 git clone https://github.com/kpu/preprocess
 cd preprocess
 git checkout wet
@@ -68,115 +58,27 @@ mkdir build
 cd build
 cmake ..
 make -j4
-alias wet_lines="${PWD}/preprocess/build/bin/wet_lines"
+alias wet_lines="${PWD}/build/bin/wet_lines"
+cd ../
 wget https://dl.fbaipublicfiles.com/seamless/data/seamless.dataset.metadata.public.enA-jpn.withduration.tsv.gz
-zcat seamless.dataset.metadata.public.enA-jpn.withduration.tsv.gz | egrep ^crawl-data | tr '\t' ' ' | wet_lines
-
-
-export DIRECTION="enA-jpn"
-
-export LINE_NO_START=0
-export LINE_NO_END=30000
-python download_audio.py
-
-export LINE_NO_START=30000
-export LINE_NO_END=60000
-python download_audio.py
-
-export LINE_NO_START=60000
-export LINE_NO_END=90000
-python download_audio.py
-
-export LINE_NO_START=90000
-export LINE_NO_END=120000
-python download_audio.py
-
-export LINE_NO_START=120000
-export LINE_NO_END=150000
-python download_audio.py
-
-# server 2
-export LINE_NO_START=150000
-export LINE_NO_END=180000
-python download_audio.py
-
-export LINE_NO_START=180000
-export LINE_NO_END=210000
-python download_audio.py
-
-export LINE_NO_START=210000
-export LINE_NO_END=240000
-python download_audio.py
-
-export LINE_NO_START=240000
-export LINE_NO_END=270000
-python download_audio.py
-
-export LINE_NO_START=270000
-export LINE_NO_END=300000
-python download_audio.py
-
-# server 3
-export LINE_NO_START=300000
-export LINE_NO_END=330000
-python download_audio.py
-
-export LINE_NO_START=330000
-export LINE_NO_END=360000
-python download_audio.py
-
-
-
-
-export LINE_NO_START=360000
-export LINE_NO_END=390000
-python download_audio.py
-
-export LINE_NO_START=390000
-export LINE_NO_END=420000
-python download_audio.py
-
-export LINE_NO_START=420000
-export LINE_NO_END=450000
-python download_audio.py
-
-# server 4
-export LINE_NO_START=450000
-export LINE_NO_END=480000
-python download_audio.py
-
-export LINE_NO_START=480000
-export LINE_NO_END=510000
-python download_audio.py
-
-export LINE_NO_START=510000
-export LINE_NO_END=540000
-python download_audio.py
-
-export LINE_NO_START=540000
-export LINE_NO_END=570000
-python download_audio.py
-
-export LINE_NO_START=570000
-export LINE_NO_END=600000
-python download_audio.py
-
-
-# server 5
-export LINE_NO_START=600000
-export LINE_NO_END=630000
-python download_audio.py
-
-export LINE_NO_START=630000
-export LINE_NO_END=660000
-python download_audio.py
-
-export LINE_NO_START=660000
-export LINE_NO_END=690000
-python download_audio.py
-
-export LINE_NO_START=690000
-export LINE_NO_END=720000
-python download_audio.py
-
+cp ../download_text.py ./
+python download_text.py
+cat seamless.dataset.metadata.public.enA-jpn.withduration.reordered.batch_1.tsv | egrep ^crawl-data | tr '\t' ' ' | wet_lines | tee seamless.dataset.metadata.public.jpn.batch_1.tsv
+cat seamless.dataset.metadata.public.enA-jpn.withduration.reordered.batch_2.tsv | egrep ^crawl-data | tr '\t' ' ' | wet_lines | tee seamless.dataset.metadata.public.jpn.batch_2.tsv
+cat seamless.dataset.metadata.public.enA-jpn.withduration.reordered.batch_3.tsv | egrep ^crawl-data | tr '\t' ' ' | wet_lines | tee seamless.dataset.metadata.public.jpn.batch_3.tsv
+cat seamless.dataset.metadata.public.enA-jpn.withduration.reordered.batch_4.tsv | egrep ^crawl-data | tr '\t' ' ' | wet_lines | tee seamless.dataset.metadata.public.jpn.batch_4.tsv
+cat seamless.dataset.metadata.public.enA-jpn.withduration.reordered.batch_5.tsv | egrep ^crawl-data | tr '\t' ' ' | wet_lines | tee seamless.dataset.metadata.public.jpn.batch_5.tsv
+cat seamless.dataset.metadata.public.enA-jpn.withduration.reordered.batch_6.tsv | egrep ^crawl-data | tr '\t' ' ' | wet_lines | tee seamless.dataset.metadata.public.jpn.batch_6.tsv
+cat seamless.dataset.metadata.public.enA-jpn.withduration.reordered.batch_7.tsv | egrep ^crawl-data | tr '\t' ' ' | wet_lines | tee seamless.dataset.metadata.public.jpn.batch_7.tsv
+cat seamless.dataset.metadata.public.enA-jpn.withduration.reordered.batch_8.tsv | egrep ^crawl-data | tr '\t' ' ' | wet_lines | tee seamless.dataset.metadata.public.jpn.batch_8.tsv
+cat seamless.dataset.metadata.public.enA-jpn.withduration.reordered.batch_9.tsv | egrep ^crawl-data | tr '\t' ' ' | wet_lines | tee seamless.dataset.metadata.public.jpn.batch_9.tsv
+cat seamless.dataset.metadata.public.enA-jpn.withduration.reordered.batch_10.tsv | egrep ^crawl-data | tr '\t' ' ' | wet_lines | tee seamless.dataset.metadata.public.jpn.batch_10.tsv
+cat seamless.dataset.metadata.public.enA-jpn.withduration.reordered.batch_11.tsv | egrep ^crawl-data | tr '\t' ' ' | wet_lines | tee seamless.dataset.metadata.public.jpn.batch_11.tsv
+
+
+########
+# NLLB #
+########
+# https://www.kecl.ntt.co.jp/icl/lirg/jparacrawl/
+python -c "from datasets import load_dataset; load_dataset('allenai/nllb', 'eng_Latn-jpn_Jpan')"