ApolloCorpus / train /question_dup.py
nchen909's picture
Update train/question_dup.py
8fde075 verified
import json
def extract_unique_strings(input_file_path, output_file_path):
with open(input_file_path, 'r', encoding='utf-8') as file:
items = json.load(file)
# 提取item[1],这里假设每个item都至少有两个元素
strings = [item[1] for item in items if len(item) > 1]
# 去重
unique_strings = list(set(strings))
# 保存为JSON文件
with open(output_file_path, 'w', encoding='utf-8') as file:
json.dump(unique_strings, file, ensure_ascii=False, indent=4)
# 打印去重后的列表长度
print(f"Number of unique strings: {len(unique_strings)}")
# 替换以下路径为你的输入文件和输出文件路径
input_file_path = 'train/sft/medicalExam_es_dup.json'
output_file_path = 'train/sft/medicalExam_es_dup_question.json'
extract_unique_strings(input_file_path, output_file_path)