如何用Python提取XML文件中命名实体的起止位置？

首先注意你的示例XML里有个语法错误：</location 应该修正为 </location>，否则解析会报错。

下面是直接可用的代码，用Python标准库xml.etree.ElementTree实现，能循环提取person和location实体的内容、起始位置和结束位置：

import xml.etree.ElementTree as ET

# 定义要提取的实体类型
target_tags = {'person', 'location'}
entities = []

# 使用iterparse解析，跟踪start和end事件以获取位置
for event, elem in ET.iterparse('your_file.xml', events=('start', 'end')):
    if event == 'start' and elem.tag in target_tags:
        # 记录实体的起始行号和标签类型
        entity_info = {
            'tag': elem.tag,
            'start_pos': elem.sourceline
        }
    elif event == 'end' and elem.tag in target_tags:
        # 补全实体的结束行号和内容
        entity_info['end_pos'] = elem.sourceline
        entity_info['content'] = elem.text.strip() if elem.text else ''
        entities.append(entity_info)
        # 清理元素避免内存泄漏
        elem.clear()

# 循环打印提取到的实体信息
for idx, entity in enumerate(entities, 1):
    print(f"实体{idx}:")
    print(f"类型: {entity['tag']}")
    print(f"内容: {entity['content']}")
    print(f"起始行: {entity['start_pos']}")
    print(f"结束行: {entity['end_pos']}\n")

补充：获取字符级别的精确偏移量 #

如果需要精确到字符的起始/结束位置，而非行号，可以结合原文件内容处理：

import xml.etree.ElementTree as ET

target_tags = {'person', 'location'}
entities = []

# 先读取完整XML内容
with open('your_file.xml', 'r', encoding='utf-8') as f:
    xml_content = f.read()

# 解析时跟踪字符偏移
parser = ET.XMLParser()
for event, elem in ET.iterparse(parser, events=('start', 'end')):
    if event == 'start' and elem.tag in target_tags:
        # 获取起始字符偏移量
        start_pos = parser.CurrentByteIndex
        entity_info = {'tag': elem.tag, 'start_pos': start_pos}
    elif event == 'end' and elem.tag in target_tags:
        end_pos = parser.CurrentByteIndex
        entity_info['end_pos'] = end_pos
        entity_info['content'] = elem.text.strip() if elem.text else ''
        entities.append(entity_info)
        elem.clear()

# 打印字符偏移信息
for idx, entity in enumerate(entities, 1):
    print(f"实体{idx}:")
    print(f"类型: {entity['tag']}")
    print(f"内容: {entity['content']}")
    print(f"起始字符位置: {entity['start_pos']}")
    print(f"结束字符位置: {entity['end_pos']}\n")

注意事项 #

• 确保XML文件语法完全正确（闭合标签、引号匹配等），否则解析会直接报错。
• 如果XML包含命名空间，需要对elem.tag做处理，比如用elem.tag.split('}')[-1]提取标签名后再匹配目标类型。

内容的提问来源于stack exchange，提问作者user18153710