webvid / main.py
Ligeng-Zhu's picture
Upload main.py with huggingface_hub
aeb3819 verified
raw
history blame
3.44 kB
import json
import sys, os, os.path as osp
import asyncio
import requests
import fire
import pandas as pd
from random import random
from concurrent.futures import ProcessPoolExecutor
def ytb_download(uid, url, json_info, output_dir="ytb_videos/"):
os.makedirs(output_dir, exist_ok=True)
# uid = url.split("?v=")[-1]
video_path = osp.join(output_dir, f"{uid}.mp4")
meta_path = osp.join(output_dir, f"{uid}.json")
if osp.exists(video_path) and osp.exists(meta_path):
print(f"{uid} already downloaded.")
return 0
if osp.exists(video_path):
print(f"[video] {uid} already downloaded.")
else:
with requests.get(url) as resp:
if resp.status_code != 200:
print(f"{uid} failed to fetch.")
return -1
print(f"downloading {uid}: {url} to {output_dir}")
open(video_path, "wb").write(resp.content)
if not osp.exists(meta_path):
with open(osp.join(output_dir, f"{uid}.json"), "w") as fp:
json.dump(json_info, fp, indent=2)
return 0
async def main(csv_path, max_workers=2, shards=0, total=-1, limit=False):
PPE = ProcessPoolExecutor(max_workers=max_workers)
loop = asyncio.get_event_loop()
df = pd.read_csv(csv_path)
output_dir = csv_path.split(".")[0]
tasks = []
data_list = df.iterrows()
if total > 0:
data_list = list(df.iterrows())
chunk = len(data_list) // total
begin_idx = shards * chunk
end_idx = (shards + 1) * chunk
if shards == total - 1:
end_idx = len(data_list)
data_list = data_list[begin_idx:end_idx]
print(f"Downloading total {len(data_list)} videos")
for idx, (index, row) in enumerate(data_list):
uid = row["videoid"]
url = row["contentUrl"]
# name,page_idx,page_dir,duration
try:
json_info = {
"name": row["name"],
"url": url,
# "page_idx": row["page_idx"],
# "page_dir": row["page_dir"],
"duration": row["duration"],
}
except KeyError as e:
print(row)
print(row.keys())
print(e)
exit(0)
tasks.append(
loop.run_in_executor(PPE, ytb_download, uid, url, json_info, output_dir)
)
if idx >= 20 and limit:
break
res = await asyncio.gather(*tasks)
print(f"[{sum(res)} / {len(res)}]")
output_log = f"log_{output_dir}.json"
key = "total"
if total is not None:
key = f"{shards}-of-{total}"
json_output = {}
if osp.exists(output_log):
with open(output_log, "r") as fp:
json_output = json.load(fp)
json_output[key] = f"[{sum(res)} / {len(res)}]"
with open(output_log, "w") as fp:
json.dump(json_output, fp, indent=2)
def entry(csv="panda70m_testing.csv", shards=0, total=-1, limit=False):
asyncio.run(main(csv, shards=shards, total=total, limit=limit))
if __name__ == "__main__":
fire.Fire(entry)
# url = "https://ak.picdn.net/shutterstock/videos/1053841541/preview/stock-footage-travel-blogger-shoot-a-story-on-top-of-mountains-young-man-holds-camera-in-forest.mp4"
# uid = "1053841541"
# json_info = {}
# res = ytb_download(uid, url, json_info, output_dir="test/")
# print(res)