Home Explore Help
Sign In
data_spider
/
match_spider
3
0
Fork 0
Files Issues 0 Pull Requests 0 Wiki
Browse Source

Merge remote-tracking branch 'origin/master'

dongzhaorui 9 months ago
parent
9d20bc91d2 0b26cf8eaf
commit
ddc8ed7109
2 changed files with 199 additions and 7 deletions
Split View Show Diff Stats
  1. 50 7
      zbytb/crawler/clean_html.py
  2. 149 0
      zbytb/crawler/clean_html_old.py

+ 50 - 7
zbytb/crawler/clean_html.py
View File 

@@ -1,3 +1,4 @@
 
+# -*- coding: utf-8 -*-
 
 import re
 
 
 __all__ = ['cleaner', 'clean_js']
 
@@ -14,21 +15,26 @@ INDEPENDENT_TAGS = {
 
     '<style[^<>]*>[\s\S]*?</style>': '',  # 样式
 
     '<script[^<>]*>[\s\S]*?</script>': '',  # JavaScript
 
     '<input>': '',  # 输入框
 
+    '</input>': '',  # 输入框
 
     '<img[^>]*>': '<br>',  # 图片
 
 }
 
 # 行内元素
 
 INLINE_TAGS = {
 
     '<a>|<a [^>]*>|</a>': '',  # 超链接
 
+    '<link>|<link [^>]*>|</link>': '',  # 超链接
 
     '<span>|<span [^>]*>|</span>': '',  # span
 
     '<label>|<label [^>]*>|</label>': '<br>',  # label
 
     '<font>|<font [^>]*>|</font>': '',  # font
 
+    'data:image(.*?) ': '',            # 图片base64
 
 }
 
 # 块级元素
 
 BLOCK_TAGS = {
 
-    '<h[1-6][^>]*>[\s\S]*?</h[1-6]>': '',  # 标题
 
-    # '<h[1-6][^>]*>|</h[1-6]>': '',  # 标题
 
-    '<p>|<p [^>]*>|</p>': '<br>',  # 段落
 
-    '<div>|<div [^>]*>|</div>': '<br>',  # 分割 division
 
+    '<div>\s*?</div>':'',
 
+    '<h[1-6][^>]*>|</h[1-6]>': '',  # 标题
 
+    '<p>|<p [^>]*>': '<br>',  # 段落
 
+    '</p>': '',  # 段落
 
+    '<div>|<div [^>]*>': '<br>',  # 分割 division
 
+    '</div>': '',  # 分割 division
 
     '<o:p>|<o:p [^>]*>|</o:p>': ''  # OFFICE微软WORD段落
 
 }
 
 # 其他
 
@@ -53,6 +59,7 @@ CSS_STYLE = {
 
     'class="[\s\S]*?"|class ="[\s\S]*?"': '',
 
     'align="[\s\S]*?"|align ="[\s\S]*?"': '',
 
     'cellpadding="(\d+)"|cellspacing="(\d+)"': '',
 
+
 
 }
 
 # 空白符
 
 BLANKS = {
 
@@ -65,6 +72,10 @@ BLANKS = {
 
 TAGS = {'table', 'tr', 'td', 'div', 'span', 'p'}
 
 # css属性集合
 
 ATTRS = {'id', 'class', 'style', 'width'}
 
+# 特殊样式+指定样式的标签
 
+SPECIAL_TAGS = {
 
+    re.compile('(?i)<[^>]+style="display: none".*[^>]+>'): "<br>",
 
+}
 
 
 
 def _repair_tag():
 
@@ -84,14 +95,36 @@ def _escape_character(html):
 
     html = html.replace('&gt;', '>')
 
     html = html.replace('&quot;', '"')
 
     html = html.replace('&amp;', '&')
 
+    # 不显示输入框边框
 
+    html = html.replace('<input', '<input style="border-color: transparent;"')
 
+    return html
 
+
 
+
 
+def _clean_input(html):
 
+    """提取value值，替换input标签"""
 
+    inputTxt = re.compile(r'<input .*?>', re.S)
 
+    valueTxt = re.compile(r'value=["|\'](.*?)["|\']')
 
+
 
+    input_list = re.findall(inputTxt, html) or []
 
+    for ipt in input_list:
 
+        val = re.findall(valueTxt, ipt)
 
+        if val and "hidden" not in ipt and "hide" not in ipt and "display: none" not in ipt:
 
+            html = html.replace(ipt,val[0])
 
     return html
 
 
 
 def _lowercase_tag(html):
 
-    """标签归一化处理（全部小写）"""
 
+    """标签归一化处理（全部小写 + 标签修复）"""
 
     tags = re.findall("<[^>]+>", html)
 
-    for tag in tags:
 
-        html = html.replace(tag, str(tag).lower())
 
+    tag_sets = set(tags)
 
+
 
+    if len(tag_sets) > 10000:
 
+        from bs4 import BeautifulSoup
 
+        soup = BeautifulSoup(html, "lxml")
 
+        html = str(soup.body.next_element)
 
+    else:
 
+        for tag in tag_sets:
 
+            html = html.replace(tag, str(tag).lower())
 
 
     repair_tags = _repair_tag()
 
     for err, right in repair_tags.items():
 
@@ -100,6 +133,13 @@ def _lowercase_tag(html):
 
     return html
 
 
 
+def _del_tag(html):
 
+    """删除特殊+指定样式的标签"""
 
+    for tag, repl in SPECIAL_TAGS.items():
 
+        html = tag.sub(repl, html)
 
+    return html
 
+
 
+
 
 def cleaner(html, special=None, completely=False):
 
     """
 
     数据清洗
 
@@ -121,6 +161,7 @@ def cleaner(html, special=None, completely=False):
 
         **BLANKS,
 
     }
 
     html = _lowercase_tag(html)
 
+    # html = _del_tag(html)  # 优先处理
 
     for tag, repl in remove_tags.items():
 
         html = re.sub(tag, repl, html)
 
 
@@ -133,6 +174,7 @@ def cleaner(html, special=None, completely=False):
 
         html = re.sub(r'([,|.|。|，|；|;|?|&|$|#|@|!|！|%|*|\'|"|‘|’|“|￥|？| ]*?)$', "", html.strip())  # 清除文本末尾符号
 
 
     html = _escape_character(html)
 
+    html = _clean_input(html)  # 处理 input
 
     return html
 
 
 
@@ -147,3 +189,4 @@ def clean_js(html):
 
     for tag, repl in remove_tags.items():
 
         html = re.sub(tag, repl, html)
 
     return html
 
+

+ 149 - 0
zbytb/crawler/clean_html_old.py
View File 

@@ -0,0 +1,149 @@
 
+import re
 
+
 
+__all__ = ['cleaner', 'clean_js']
 
+
 
+# 独立元素
 
+INDEPENDENT_TAGS = {
 
+    '<head>[\s\S]*?</head>': '',
 
+    '<html>|<html [^>]*>|</html>': '',
 
+    '<body>|<body [^>]*>|</body>': '',
 
+    '<meta[^<>]*>|<meta [^<>]*>|<meta[^<>]*>[\s\S]*?</meta>|</meta>': '',  # 元数据
 
+    '&(nbsp|e[mn]sp|thinsp|zwn?j|#13);': '',  # 空格
 
+    '\\xa0|\\u3000': '',  # 空格
 
+    '<!--[\s\S]*?-->': '',  # 注释
 
+    '<style[^<>]*>[\s\S]*?</style>': '',  # 样式
 
+    '<script[^<>]*>[\s\S]*?</script>': '',  # JavaScript
 
+    '<input>': '',  # 输入框
 
+    '<img[^>]*>': '<br>',  # 图片
 
+}
 
+# 行内元素
 
+INLINE_TAGS = {
 
+    '<a>|<a [^>]*>|</a>': '',  # 超链接
 
+    '<span>|<span [^>]*>|</span>': '',  # span
 
+    '<label>|<label [^>]*>|</label>': '<br>',  # label
 
+    '<font>|<font [^>]*>|</font>': '',  # font
 
+}
 
+# 块级元素
 
+BLOCK_TAGS = {
 
+    '<h[1-6][^>]*>[\s\S]*?</h[1-6]>': '',  # 标题
 
+    # '<h[1-6][^>]*>|</h[1-6]>': '',  # 标题
 
+    '<p>|<p [^>]*>|</p>': '<br>',  # 段落
 
+    '<div>|<div [^>]*>|</div>': '<br>',  # 分割 division
 
+    '<o:p>|<o:p [^>]*>|</o:p>': ''  # OFFICE微软WORD段落
 
+}
 
+# 其他
 
+OTHER = {
 
+    '<?xml[^>]*>|<?xml [^>]*>|<?xml:.*?>': '',
 
+    '<epointform>': '',
 
+    '<!doctype html>|<!doctype html [^>]*>': '',
 
+    '【关闭】|关闭': '',
 
+    '【打印】|打印本页': '',
 
+    '【字体：[\s\S]*】': '',
 
+    '文章来源：[\u4e00-\u9fa5]+': '',
 
+    '浏览次数：.*[<]+': '',
 
+    '（责任编辑：.*?）': '',
 
+    '分享到[：]': '',
 
+    '阅读数[:：]\d+': '',
 
+}
 
+# 样式
 
+CSS_STYLE = {
 
+    'style="[\s\S]*?"|style ="[\s\S]*?"': '',
 
+    'bgcolor="[\s\S]*?"|bgcolor ="[\s\S]*?"': '',
 
+    'bordercolor="[\s\S]*?"|bordercolor ="[\s\S]*?"': '',
 
+    'class="[\s\S]*?"|class ="[\s\S]*?"': '',
 
+    'align="[\s\S]*?"|align ="[\s\S]*?"': '',
 
+    'cellpadding="(\d+)"|cellspacing="(\d+)"': '',
 
+}
 
+# 空白符
 
+BLANKS = {
 
+    '\n\s*\n': '\n',
 
+    '\s*\n\s*': '\n',
 
+    '[^\S\n]': ' ',
 
+    '\s+': ' ',
 
+}
 
+# css标签集合
 
+TAGS = {'table', 'tr', 'td', 'div', 'span', 'p'}
 
+# css属性集合
 
+ATTRS = {'id', 'class', 'style', 'width'}
 
+
 
+
 
+def _repair_tag():
 
+    """异常的标签组合,用来替换非标准页面的标签"""
 
+    _repairs = {}
 
+    for tag in TAGS:
 
+        for attr in ATTRS:
 
+            key = '{}{}'.format(tag, attr)
 
+            val = '{} {}'.format(tag, attr)
 
+            _repairs[key] = val
 
+    return _repairs
 
+
 
+
 
+def _escape_character(html):
 
+    """转义字符"""
 
+    html = html.replace('&lt;', '<')
 
+    html = html.replace('&gt;', '>')
 
+    html = html.replace('&quot;', '"')
 
+    html = html.replace('&amp;', '&')
 
+    return html
 
+
 
+
 
+def _lowercase_tag(html):
 
+    """标签归一化处理（全部小写）"""
 
+    tags = re.findall("<[^>]+>", html)
 
+    for tag in tags:
 
+        html = html.replace(tag, str(tag).lower())
 
+
 
+    repair_tags = _repair_tag()
 
+    for err, right in repair_tags.items():
 
+        html = html.replace(err, right)
 
+
 
+    return html
 
+
 
+
 
+def cleaner(html, special=None, completely=False):
 
+    """
 
+    数据清洗
 
+
 
+    :param html: 清洗的页面
 
+    :param special: 额外指定页面清洗规则
 
+    :param completely: 是否完全清洗页面
 
+    :return: 清洗后的页面源码
 
+    """
 
+    if special is None:
 
+        special = {}
 
+    OTHER.update(special)
 
+    remove_tags = {
 
+        **INDEPENDENT_TAGS,
 
+        **INLINE_TAGS,
 
+        **BLOCK_TAGS,
 
+        **OTHER,
 
+        **CSS_STYLE,
 
+        **BLANKS,
 
+    }
 
+    html = _lowercase_tag(html)
 
+    for tag, repl in remove_tags.items():
 
+        html = re.sub(tag, repl, html)
 
+
 
+    if completely:
 
+        html = re.sub(r'<canvas[^<>]*>[\s\S]*?</canvas>', '', html)  # 画布
 
+        html = re.sub(r'<iframe[^<>]*>[\s\S]*?</iframe>', '', html)  # 内框架
 
+        html = re.sub('<([^<>\u4e00-\u9fa5]|微软雅黑|宋体|仿宋)+>', '', html)
 
+
 
+    if html:
 
+        html = re.sub(r'([,|.|。|，|；|;|?|&|$|#|@|!|！|%|*|\'|"|‘|’|“|￥|？| ]*?)$', "", html.strip())  # 清除文本末尾符号
 
+
 
+    html = _escape_character(html)
 
+    return html
 
+
 
+
 
+def clean_js(html):
 
+    remove_tags = {
 
+        '&(nbsp|e[mn]sp|thinsp|zwn?j|#13);': '',  # 空格
 
+        '\\xa0|\\u3000': '',  # 空格
 
+        '<script[^<>]*>[\s\S]*?</script>': '',  # JavaScript
 
+        **BLANKS,
 
+    }
 
+    html = _lowercase_tag(html)
 
+    for tag, repl in remove_tags.items():
 
+        html = re.sub(tag, repl, html)
 
+    return html