import pandas as pd


def extract_features(file_path):
    # 读取csv文件
    df = pd.read_csv(file_path, delimiter=',')

    # 按第2列数值降序排序
    df["count"] = pd.to_numeric(df["count"], errors='coerce')
    df_sorted = df.sort_values(by='count', ascending=True)

    # 筛选出第2列值大于10000的行，并提取第1列内容
    features = df_sorted[df_sorted['count'] <0]

    return features

if __name__ == '__main__':

    # 使用函数，传入csv文件路径
    features = extract_features('./out/3gram.csv')
    print(features)