iwla/plugins/display/referers.py

import time
import re
import xml.sax.saxutils as saxutils

from iwla import IWLA
from iplugin import IPlugin
from display import *

import awstats_data

class IWLADisplayReferers(IPlugin):
    def __init__(self, iwla):
        super(IWLADisplayReferers, self).__init__(iwla)
        self.API_VERSION = 1

    def _getSearchEngine(self, hashid):
        #print 'Look for %s' % engine
        for (k, e) in self.search_engines.items():
            for (h,h_re) in e['hashid']:
                if hashid == h:
                    return k
        #print 'Not found %s' % (hashid)
        return None

    def load(self):
        domain_name = self.iwla.getConfValue('domain_name', '')
        
        if not domain_name:
            print 'domain_name required in conf'
            return False

        self.own_domain_re = re.compile(r'.*%s.*' % (domain_name))
        self.search_engines = {}

        for (hashid, name) in awstats_data.search_engines_hashid.items():
            hashid_re = re.compile(r'.*%s.*' % (hashid))
            if not name in self.search_engines.keys():
                self.search_engines[name] = {
                    'hashid' : [(hashid, hashid_re)]
                    }
            else:
                self.search_engines[name]['hashid'].append((hashid, hashid_re))
            #print 'Hashid %s => %s' % (name, hashid)

        for (name, known_url) in awstats_data.search_engines_knwown_url.items():
            self.search_engines[name]['known_url'] = re.compile(known_url + '(?P<key_phrase>.+)')

        for (engine, not_engine) in awstats_data.not_search_engines_keys.items():
            not_engine_re = re.compile(r'.*%s.*' % (not_engine))
            key = self._getSearchEngine(engine)
            if key:
                self.search_engines[key]['not_search_engine'] = not_engine_re

        #self.html_parser = html.parser.HTMLParser()

        return True

    def _extractKeyPhrase(self, key_phrase_re, parameters, key_phrases):
        if not parameters or not key_phrase_re: return

        for p in parameters.split('&'):
            groups = key_phrase_re.match(p)
            if groups:
                key_phrase = groups.groupdict()['key_phrase']
                key_phrase = key_phrase.replace('+', ' ').lower()
                key_phrase = saxutils.unescape(key_phrase)
                if not key_phrase in key_phrases.keys():
                    key_phrases[key_phrase] = 1
                else:
                    key_phrases[key_phrase] += 1
                break

    def hook(self, iwla):
        stats = iwla.getCurrentVisists()
        referers = {}
        robots_referers = {}
        search_engine_referers = {}
        key_phrases = {}

        for (k, super_hit) in stats.items():
            for r in super_hit['requests']:
                if not r['http_referer']: continue
                
                uri = r['extract_referer']['extract_uri']
                is_search_engine = False

                if self.own_domain_re.match(uri): continue

                for (name, engine) in self.search_engines.items():
                    for (hashid, hashid_re) in engine['hashid']:
                        if not hashid_re.match(uri): continue

                        not_engine = engine.get('not_search_engine', None)
                        # Try not engine
                        if not_engine and not_engine.match(uri): break
                        is_search_engine = True
                        uri = name

                        parameters = r['extract_referer'].get('extract_parameters', None)
                        key_phrase_re = engine.get('known_url', None)

                        self._extractKeyPhrase(key_phrase_re, parameters, key_phrases)
                        break

                if is_search_engine:
                    dictionary = search_engine_referers
                elif super_hit['robot']:
                    dictionary = robots_referers
                    # print '%s => %s' % (uri, super_hit['remote_ip'])
                else:
                    dictionary = referers
                if r['is_page']:
                    key = 'pages'
                else:
                    key = 'hits'
                if not uri in dictionary: dictionary[uri] = {'pages':0, 'hits':0}
                dictionary[uri][key] += 1

        top_referers = [(k, referers[k]['pages']) for k in referers.keys()]
        top_referers = sorted(top_referers, key=lambda t: t[1], reverse=True)

        top_robots_referers = [(k, robots_referers[k]['pages']) for k in robots_referers.keys()]
        top_robots_referers = sorted(top_robots_referers, key=lambda t: t[1], reverse=True)

        top_search_engine_referers = [(k, search_engine_referers[k]['pages']) for k in search_engine_referers.keys()]
        top_search_engine_referers = sorted(top_search_engine_referers, key=lambda t: t[1], reverse=True)

        top_key_phrases = key_phrases.items()
        top_key_phrases = sorted(top_key_phrases, key=lambda t: t[1], reverse=True)

        # Top referers in index
        index = self.iwla.getDisplayIndex()

        table = DisplayHTMLBlockTable('Connexion from', ['Origin', 'Pages', 'Hits'])
        table.appendRow(['<b>Search Engine</b>', '', ''])
        for r,_ in top_search_engine_referers[:10]:
            row = [r, search_engine_referers[r]['pages'], search_engine_referers[r]['hits']]
            table.appendRow(row)

        table.appendRow(['<b>External URL</b>', '', ''])
        for r,_ in top_referers[:10]:
            row = [r, referers[r]['pages'], referers[r]['hits']]
            table.appendRow(row)

        table.appendRow(['<b>External URL (robot)</b>', '', ''])
        for r,_ in top_robots_referers[:10]:
            row = [r, robots_referers[r]['pages'], robots_referers[r]['hits']]
            table.appendRow(row)

        index.appendBlock(table)

        # All referers in a file
        cur_time = self.iwla.getCurTime()
        title = time.strftime('Connexion from - %B %Y', cur_time)

        filename = 'referers_%d.html' % (cur_time.tm_mon)
        path = '%d/%s' % (cur_time.tm_year, filename)

        page = DisplayHTMLPage(title, path)
        table = DisplayHTMLBlockTable('Connexion from', ['Origin', 'Pages', 'Hits'])
                        
        table.appendRow(['<b>Search Engine</b>', '', ''])
        for r,_ in top_search_engine_referers:
            row = [r, search_engine_referers[r]['pages'], search_engine_referers[r]['hits']]
            table.appendRow(row)

        table.appendRow(['<b>External URL</b>', '', ''])
        for r,_ in top_referers:
            row = [r, referers[r]['pages'], referers[r]['hits']]
            table.appendRow(row)

        table.appendRow(['<b>External URL (robot)</b>', '', ''])
        for r,_ in top_robots_referers:
            row = [r, robots_referers[r]['pages'], robots_referers[r]['hits']]
            table.appendRow(row)

        page.appendBlock(table)

        display = self.iwla.getDisplay()
        display.addPage(page)

        block = DisplayHTMLRawBlock()
        block.setRawHTML('<a href=\'%s\'>All referers</a>' % (filename))
        index.appendBlock(block)

        # Top key phrases in index
        table = DisplayHTMLBlockTable('Top key phrases', ['Key phrase', 'Search'])
        for phrase in top_key_phrases[:10]:
            table.appendRow([phrase[0], phrase[1]])
        index.appendBlock(table)

        # All key phrases in a file
        cur_time = self.iwla.getCurTime()
        title = time.strftime('Key Phrases - %B %Y', cur_time)

        filename = 'key_phrases_%d.html' % (cur_time.tm_mon)
        path = '%d/%s' % (cur_time.tm_year, filename)

        page = DisplayHTMLPage(title, path)
        table = DisplayHTMLBlockTable('Top key phrases', ['Key phrase', 'Search'])
        for phrase in top_key_phrases:
            table.appendRow([phrase[0], phrase[1]])
        page.appendBlock(table)
                        
        display.addPage(page)

        block = DisplayHTMLRawBlock()
        block.setRawHTML('<a href=\'%s\'>All key phrases</a>' % (filename))
        index.appendBlock(block)