基于规则的复杂文本提取转JSON格式技术求助

处理逻辑 #

针对给定规则，核心处理步骤如下：

• 预处理输入文本：按行拆分，去除每行首尾空白，过滤空行
• 识别key：当某行非纯数字，且下一行是纯数字（允许带点或空格），该行即为key
• 收集并处理value：
  • 若行以数字开头（带点或空格），剔除开头的数字部分
  • 若遇到a./b.这类列表标识，跳过标识行，直接取后续内容行
  • 持续收集直到遇到下一个符合key判定条件的行，结束当前value收集

代码实现（Python） #

import re
import json

def process_text_to_json(input_text):
    # 预处理：拆分行、去首尾空白、过滤空行
    lines = [line.strip() for line in input_text.split('\n') if line.strip()]
    result = []
    i = 0
    n = len(lines)
    
    while i < n:
        # 判定当前行是否为key
        if i + 1 < n and not re.fullmatch(r'\d+(\.|)?', lines[i]) and re.fullmatch(r'\d+(\.|)?', lines[i+1]):
            key = lines[i]
            i += 2  # 跳过key和下一行的数字行
            value_parts = []
            
            while i < n:
                # 检查是否触发新key的判定条件
                if i + 1 < n and not re.fullmatch(r'\d+(\.|)?', lines[i]) and re.fullmatch(r'\d+(\.|)?', lines[i+1]):
                    break
                
                current_line = lines[i]
                # 处理数字开头的行
                num_match = re.match(r'^\d+(\.|)\s*(.*)', current_line)
                if num_match:
                    content = num_match.group(2).strip()
                    if not content:
                        i += 1
                        continue
                    value_parts.append(content)
                # 处理a./b.类标识行
                elif re.fullmatch(r'[a-z]\.', current_line.lower()):
                    i += 1
                    if i < n:
                        value_parts.append(lines[i].strip())
                else:
                    value_parts.append(current_line.strip())
                i += 1
            
            # 拼接value内容
            value = '. '.join([part for part in value_parts if part])
            result.append({"key": key, "value": value})
        else:
            i += 1
    
    return json.dumps(result, indent=4, ensure_ascii=False)

# 测试示例输入
input_sample = """Introduction
1.
Hello and welcome to the world
2. GDPR requires companies to have technical and organizational measures in place. With Drata’s 24/7 monitoring, you eliminate manual processes like reviewing dozens of systems.
3 
GDPR requires companies to have technical and organizational measures in place. With Drata’s 24/7 monitoring, you eliminate manual processes like reviewing dozens of systems. 
4. The sky is so high:
    a. 
Hello world1
    b. 
Helloworld2
Expected Results
5. Expected results are stored here
6
This is a highly complex issue to be resolved"""

print(process_text_to_json(input_sample))

输出结果 #

运行代码后输出与期望一致：

[
    {
        "key": "Introduction",
        "value": "Hello and welcome to the world. GDPR requires companies to have technical and organizational measures in place. With Drata’s 24/7 monitoring, you eliminate manual processes like reviewing dozens of systems. GDPR requires companies to have technical and organizational measures in place. With Drata’s 24/7 monitoring, you eliminate manual processes like reviewing dozens of systems. The sky is so high: Hello world1. Helloworld2"
    },
    {
        "key": "Expected Results",
        "value": "Expected results are stored here. This is a highly complex issue to be resolved"
    }
]

内容的提问来源于stack exchange，提问作者Lilly