source: products/quintagroup.seoptimizer/trunk/quintagroup/seoptimizer/browser/keywords.py @ 1680

import urllib, re, os, commands
from xml.dom import minidom, Node

from zope.interface import implements
from Products.Five.browser import BrowserView

from Products.CMFCore.utils import getToolByName

from interfaces import IValidateSEOKeywordsView

class ValidateSEOKeywordsView(BrowserView):

    implements(IValidateSEOKeywordsView)

    def validateKeywords(self, text):
        """ see interface """
        ts = getToolByName(self.context, 'translation_service')
        # extract keywords from text
        if text.lower().strip():
            keywords = map(lambda x: x.strip(), text.lower().strip().split('\n'))
        else:
            return ts.utranslate(domain='quintagroup.seoptimizer', msgid=u'Keywords list is empty!', context=self.context)
        # request html page of context object
        url = '%s?without_metatag_keywords=1' % self.context.absolute_url()

        # extract words from url page using lynx browser (test page by 'url' randered without metatag keywords)
        page_text = commands.getoutput('lynx --dump --nolist %s' % url).lower()
        if page_text and page_text != 'sh: lynx: command not found':
            page_text = page_text.decode('utf8')
        else:
            return ts.utranslate(domain='quintagroup.seoptimizer', msgid=u'Could not find lynx browser!', context=self.context)

        # check every keyword on appearing in body of html page
        missing = []
        finding = []
        added = {}
        finded = {}
        for keyword in keywords:
            keyword = keyword.decode('utf8')
            if keyword:
                keyword_on_page =  len(re.findall(u'\\b%s\\b' % keyword, page_text, re.I|re.U))
                if keyword not in added.keys() and not keyword_on_page:
                    missing.append(keyword+u' - 0')
                    added[keyword] = 1
                if keyword not in finded.keys() and keyword_on_page:
                    finding.append(keyword+u' - '+repr(keyword_on_page))
                    finded[keyword] = 1
        # return list of missing and fount keywords
        if missing or finding:
            msg = ts.utranslate(domain='quintagroup.seoptimizer', msgid=u'number_keywords',
                                default=u'Number of keywords at page:\n${found}\n${missing}',
                                mapping={'missing':'\n'.join(missing), 'found': '\n'.join(finding)},
                                context=self.context)
        else:
            msg = ''
        return msg

    def walkTextNodes(self, parent, page_words=[]):
        for node in parent.childNodes:
            if node.nodeType == Node.ELEMENT_NODE:
                self.walkTextNodes(node, page_words)
            elif node.nodeType == Node.TEXT_NODE:
                value = node.nodeValue
                if value is not None:
                    page_words.extend(map(lambda x: x.lower(), value.split()))

    def strip_tags(self, in_text):
        s_list = list(in_text)
        i,j = 0,0

        while i < len(s_list):
            if s_list[i] == '<':
                while s_list[i] != '>':
                    # pop everything from the the left-angle bracket until the right-angle bracket
                    s_list.pop(i)

                # pops the right-angle bracket, too
                s_list.pop(i)
            else:
                i=i+1

        # convert the list back into text
        join_char=''
        return join_char.join(s_list)