import re import requests from parsel import Selector class DaZongDianPing: def __init__(self): self.headers = { 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/86.0.4240.183 Safari/537.36', 'Host': 'www.dianping.com', } self.main() def main(self): html = self.get_index() css_url, class_name = self.get_url_and_tag(html) di = self.get_css_and_svg(css_url, class_name) self.parse_index(html, di) def get_index(self): url = 'http://www.dianping.com/shop/G8svaNSPiUlDoeYK/review_all' resp = requests.get(url, headers=self.headers) if resp.status_code == 200: return resp.text def get_url_and_tag(self, html): '''获取css_url和网页中的加密字体标签的class名''' css_url = re.findall(r'href="(.*?svgtextcss.*?)"', html) if css_url: css_url = 'http:' + css_url[0] # print(css_url) # 加密字体的class名 class_name = re.findall(r'<svgmtsi class="(.*?)">', html) return css_url, class_name def get_css_and_svg(self, css_url, class_name): ''' 获取css属性和svg地址，根据css属性查找真实数据，构建替换字典 svg地址有3个 cc[class^="wgx"] 电话 bb[class^="wnu"] 地址 svgmtsi[class^="kvg"] 评论 ''' css_resp = requests.get(css_url).text.replace("\n", "").replace(" ", "") # print(css_resp) # 获取评论的svg地址 svg_url = re.findall(r'svgmtsi.*?url\((.*?)\);', css_resp) if svg_url: svg_url = 'http:' + svg_url[0] # print(svg_url) svg_resp = requests.get(svg_url).text # 获取css属性值 对应的坐标值 d = {} for name in class_name: coord = re.findall(r"%s{background:-(.*?)px-(.*?)px;}" % name, css_resp) x, y = coord[0] css_x, css_y = int(float(x)), int(float(y)) # 获取svg标签对应的y值，规则是svg_y>=css_y svg_data = Selector(svg_resp) tests = svg_data.xpath('//text') # 3.如何选择svg_y？比较y坐标，选择大于等于css_y的最接近的svg_y svg_y = [i.attrib.get('y') for i in tests if css_y <= int(i.attrib.get('y'))][0] # 根据svg_y确定具体的text的标签 svg_text = svg_data.xpath(f'//text[@y="{svg_y}"]/text()').extract_first() # 4、确认SVG中的文字大小 font_size = re.findall(r'font-size:(\d+)px', svg_resp)[0] # 5、得到css样式vhkbvu属性映射svg的位置 # css_x // 字体大小 的值就是数值的下标 position = css_x // int(font_size) s = svg_text[position] d[name] = s # 加密字体整个标签与真实值之间的字典 di = {f'<svgmtsi class="{k}"></svgmtsi>': v for k, v in d.items()} return di def parse_index(self, html, di): '''解析网页数据''' for key, value in di.items(): if key in html: html = html.replace(key, value) # print(html) selector = Selector(html) # 评论摘要 desc_li = selector.xpath('//div[@class="review-truncated-words"]/text()').extract() for desc in desc_li: desc = desc.replace('\t', '').replace('\n', '').replace(' ', '') print(desc) if __name__ == '__main__': a = DaZongDianPing()
