import json

# 1. 读取原始数据
with open("movies.json", "r", encoding="utf-8") as f:
    movies = json.load(f)

# 2. 过滤掉quote为空的条目
filtered_data = [
    {
        "id": movie["rank"],
        "title": movie["title"],
        "quote": movie["quote"]
    }
    for movie in movies
    if movie["quote"].strip() != ""
]

print(f"过滤前：{len(movies)} 条")
print(f"过滤后：{len(filtered_data)} 条")

# 3. 保存为Label Studio可导入的格式
with open("filtered_quotes.json", "w", encoding="utf-8") as f:
    json.dump(filtered_data, f, ensure_ascii=False, indent=4)

print("✅ 已生成 filtered_quotes.json，可直接导入Label Studio")