newsevent.py 文件源码

def latest_content(url):
    '''
        ??????????
    Parameter
    --------
        url:????

    Return
    --------
        string:?????????
    '''
    try:
        html = lxml.html.parse(url)
        res = html.xpath('//div[@id=\"artibody\"]/p')
        if ct.PY3:
            sarr = [etree.tostring(node).decode('utf-8') for node in res]
        else:
            sarr = [etree.tostring(node) for node in res]
        sarr = ''.join(sarr).replace('　', '')#.replace('\n\n', '\n').
        html_content = lxml.html.fromstring(sarr)
        content = html_content.text_content()
        return content
    except Exception as er:
        print(str(er))