from datasets import load_dataset
import pandas as pd
import re
# # Load the WebVid dataset
# dataset = load_dataset('webvid', 'webvid-10m', split='train')
# from datasets import load_dataset

ds = load_dataset("TempoFunk/webvid-10M", cache_dir="/mnt/data/lizhikai/webvid/")
v = ds['validation']['name']
# 定义字符串长度的合理范围
MIN_LENGTH = 30
MAX_LENGTH = 300
pattern = re.compile(r'^[a-zA-Z\s]+$')

# 过滤掉空字符串和特别长特别短的字符串
v = [s for s in v if len(s) >= MIN_LENGTH and len(s) <= MAX_LENGTH and pattern.match(s)]

# 指定保存文件的路径
file_path = 'webvid_prompt.txt'

# 打开文件，以写入模式
with open(file_path, 'w', encoding='utf-8') as file:
    # 遍历列表中的每个字符串并写入文件
    for item in v:
        if '\n' in item:
            continue
        else:
            file.write(item + '\n')

print("字符串列表已成功保存到文件中。")