import os
import re
import json


def extract_qa_from_txt(txt_path):
    with open(txt_path, 'r', encoding='utf-8') as file:
        text = file.read()

    # 正则表达式匹配问答对
    pattern = re.compile(r'(问[:：](.*?))\s*答[:：](.*?)(?=(问[:：]|$))', re.S)
    qa_pairs = []

    for match in re.findall(pattern, text):
        question = match[1].strip()
        answer = match[2].strip()
        qa_pairs.append({'问': question, '答': answer})

    return qa_pairs


def save_as_json(data, json_path):
    with open(json_path, 'w', encoding='utf-8') as f:
        json.dump(data, f, ensure_ascii=False, indent=4)


def process_folder(input_folder_path, output_folder_path):
    if not os.path.exists(output_folder_path):
        os.makedirs(output_folder_path)

    for filename in os.listdir(input_folder_path):
        if filename.endswith('.txt'):
            txt_path = os.path.join(input_folder_path, filename)
            json_filename = filename.replace('.txt', '.json')
            json_path = os.path.join(output_folder_path, json_filename)

            qa_pairs = extract_qa_from_txt(txt_path)
            save_as_json(qa_pairs, json_path)

            print(f"问答对已成功提取并保存到 {json_path}")


# 设置你的输入和输出文件夹路径
input_folder_path = r'E:\Project\PaddleOCR\bilu_txt_dispose\OCR\OCR_txt'
output_folder_path = r'E:\Project\PaddleOCR\bilu_txt_dispose\OCR\txt_json'

process_folder(input_folder_path, output_folder_path)