def extract(self,
html,
title_xpath='',
author_xpath='',
publish_time_xpath='',
host='',
noise_node_list=None,
with_body_html=False):
element = pre_parse(html)
remove_noise_node(element, noise_node_list)
content = ContentExtractor().extract(element, host, with_body_html)
title = TitleExtractor().extract(element, title_xpath=title_xpath)
publish_time = TimeExtractor().extractor(
element, publish_time_xpath=publish_time_xpath)
author = AuthorExtractor().extractor(element,
author_xpath=author_xpath)
result = {
'title': title,
'author': author,
'publish_time': publish_time,
'content': content[0][1]['text'],
'images': content[0][1]['images']
}
if with_body_html or config.get('with_body_html', False):
result['body_html'] = content[0][1]['body_html']
return result
def extract(self,
html,
title_xpath='',
author_xpath='',
publish_time_xpath='',
host='',
body_xpath='',
noise_node_list=None,
with_body_html=False):
# 对 HTML 进行预处理可能会破坏 HTML 原有的结构，导致根据原始 HTML 编写的 XPath 不可用
# 因此，如果指定了 title_xpath/author_xpath/publish_time_xpath，那么需要先提取再进行
# 预处理
normal_html = normalize_text(html)
element = html2element(normal_html)
title = TitleExtractor().extract(element, title_xpath=title_xpath)
publish_time = TimeExtractor().extractor(element, publish_time_xpath=publish_time_xpath)
author = AuthorExtractor().extractor(element, author_xpath=author_xpath)
element = pre_parse(element)
remove_noise_node(element, noise_node_list)
content = ContentExtractor().extract(element,
host=host,
with_body_html=with_body_html,
body_xpath=body_xpath)
result = {'title': title,
'author': author,
'publish_time': publish_time,
'content': content[0][1]['text'],
'images': content[0][1]['images']
}
if with_body_html or config.get('with_body_html', False):
result['body_html'] = content[0][1]['body_html']
return result
def extract(
self,
html, # 目标网站的源代码
title_xpath='', # 新闻标题的 XPath，用于定向提取标题
author_xpath='', # 文章作者的 XPath，用于定向提取文章作者
publish_time_xpath='', # 文章发布时间的 XPath，用于定向提取文章发布时间
host='', # 图片所在的域名，例如 https://www.kingname.info, 那么，当GNE 从新闻网站提取到图片的相对连接``/images/123.png``时，会把 host 拼接上去，变成``https://www.kingname.info/images/123.png``
noise_node_list=None, # 移除会导致干扰的标签。(XPath 的列表,列表中的 XPath 对应的标签，会在预处理时被直接删除掉，从而避免他们影响新闻正文的提取)
with_body_html=False
): # 为 True时，返回的结果会包含字段 body_html，内容是新闻正文所在标签的 HTML 源代码，默认为False
# 预解析html(剔除换行和无用节点)
element = pre_parse(html)
# 剔除用户自定义的干扰节点
remove_noise_node(element, noise_node_list)
# 正文抽取
content = ContentExtractor().extract(element, host, with_body_html)
# 标题抽取
title = TitleExtractor().extract(element, title_xpath=title_xpath)
# 发表时间抽取
publish_time = TimeExtractor().extractor(
element, publish_time_xpath=publish_time_xpath)
# 作者抽取
author = AuthorExtractor().extractor(element,
author_xpath=author_xpath)
# 汇总结果字典
result = {
'title': title,
'author': author,
'publish_time': publish_time,
'content': content[0][1]['text'],
'images': content[0][1]['images']
}
# 判断结果字典是否加入源码
if with_body_html or config.get('with_body_html', False):
result['body_html'] = content[0][1]['body_html']
return result
def __init__(self):
self.content_extractor = ContentExtractor()
self.title_extractor = TitleExtractor()
self.author_extractor = AuthorExtractor()
self.time_extractor = TimeExtractor()