import re
import urllib

import annotate

all_engines = []

class Engine(object):

    """
    A container for engine information.  All instances go into the
    global variable ``all_engines``.
    """

    def __init__(self, name, regexes, url):
        self.name = name
        self.url = url
        self.regexes = []
        if not isinstance(regexes, (list, tuple)):
            regexes = [regexes]
        for regex in regexes:
            if isinstance(regex, (str, unicode)):
                regex = re.compile(regex)
            self.regexes.append(regex)

def engine(*args, **kw):
    """Make an engine"""
    all_engines.append(Engine(*args, **kw))

def quick(name, url):
    """
    Make an engine using an abbreviated definition, given a name
    and a URL.
    """
    if not url.startswith('http://'):
        url = 'http://' + url
    real_url = url.split('?')[0].split('/')[0]
    search_url = '^' + url.split('?')[0]
    if '?' in url:
        param_name = url.split('?')[1]
        search_url = search_url + '.*' + param_name + '=([^&]*)'
    else:
        search_url = search_url + '(.*)'
    all_engines.append(Engine(
        name,
        [search_url],
        real_url))

quick('Google', 'www.google.[a-z.]+/?q')
quick('Google Images', 'images.google.com/images?q')
engine('Google Image View',
       [r'images.google.[a-z.]+/images.*&q=([^&]*)',
        r'(?:images|www).google.[a-z.]+/imgres.*%3Fq%3D(.*?)%26',
        r'(?:images|www).google.[a-z.]+/imgres.*()'],
       'http://images.google.com')
engine('Yahoo',
       [r'search.yahoo.com/(?:bin/)?search.*[?&]p=([^&]*)'],
       'http://search.yahoo.com')
engine('Yahoo Images',
       [r'[a-z.]*search.yahoo.[a-z]+/(?:search/)?images/view.*%26(?:va|p)%3D(.*?)%26',
        r'[a-z.]*search.yahoo.[a-z]+/(?:search/)?images/view.*&p=(.*?)&',
        r'kr.imagesearch.yahoo.com/imgsrch/av/av_detail()'],
       'http://images.search.yahoo.com')
engine('Yahoo RD',
       [r'rds.yahoo.com.*/K=([^/]*)'],
       'http://rds.yahoo.com')
quick('Metacrawler', 'http://www.metacrawler.com/info.metac/search/web/')
quick('Dogpile', 'http://www.dogpile.com/info.dogpl/search/web/')
quick('Altavista', 'http://www.altavista.com/web/results?q')
quick('Altavista Images', 'http://www.altavista.com/image/results?q')
quick('Excite', 'http://msxml.excite.com/info.xcite/search/web/')
quick('Freshmeat', 'http://freshmeat.net/search?q')
quick('Snap', 'http://www.snap.com/search.php?query')
engine('AOL Search',
       [r'aolsearch.aol.com/aol/search.*[?&]query=([^&]*)',
        r'sucheaol.aol.de/suche/bilder/search.*[?&]q=([^&]*)'],
       'http://aolsearch.aol.com')
engine('AOL Images',
       [r'aolsearch.aol.com/aol/imageDetails.*[?&]query=([^&]*)',
        r'aolimages.aol.fr/image.*[?&]query=([^&]*)',
        r'aolsearch.aol.co.uk/image_browse.*[?&]query=([^&]*)'],
       'http://aolsearch.aol.com/aol/imagehome')
engine('MSN',
       [r'http://search.msn.com/.*results.asp.*[?&]q=([^&]*)',
        r'http://sea.search.msn.com/(?:preview|spresults).asp.*[?&]q=([^&]*)',
        r'http://search.sympatico.msn.ca/results.asp.*[?&]q=([^&]*)',
        r'http://search.ninemsn.com.au/results.aspx.*[?&]q=([^&]*)'],
       'http://search.msn.com')
quick('Mamma', 'http://mamma.com/Mamma?query')
quick('Voila', 'http://search.ke.voila.fr/S/voila?kw')
quick('Compuserve', 'http://websearch.cs.com/cs/search?query')
quick('Acoon', 'http://www.acoon.de/cgi-bin/search.exe?begriff')
quick('Ask Jeeves', 'http://web.ask.com/web?q')
engine('MyWay',
       [r'mysearch.myway.com/jsp/GGmain.jsp.*[?&]searchfor=([^&]*)'],
       'http://mysearch.myway.com')
quick('Search.com', 'http://www.search.com/search?q')
quick('Lycos', 'http://search.lycos.com/default.asp?query')
quick('AllTheWeb', 'http://www.alltheweb.com/search?q')
quick('Netscape.com', 'http://search.netscape.com/ns/?query')
quick('ICQsearch', 'http://google.icq.com/search/results.php?q')
quick('Searchalot', 'http://www.searchalot.com/texis/open/search?q')
quick('Go Hip!', 'http://search.gohip.com/nph-general_search.cgi?sc')
quick('Sensis.com.au', 'http://www.sensis.com.au/search.do?find')
quick('Profusion', 'http://www.profusion.com/results?queryterm')
quick('WebTV', 'http://search.bay.webtv.net/?q')
quick('Teoma', 'http://s.teoma.com/search?q')
quick('Earthlink', 'http://search.earthlink.net/search?q')
quick('HotBot', 'http://www.hotbot.com/?query')
engine('Vivisimo',
       [r'http://[a-z.]*vivisimo.com/search.*[&?]query=([^&]*)'],
       'http://vivisimo.com')
quick('iWon', 'http://search.iwon.com/?searchfor')
quick('Seeker Bar', 'http://www.websearch.com/?qkw')
quick('Delv UK', 'http://www.delv.co.uk/results.asp?qry')
quick('Marsfind', 'http://www.marsfind.com/search.html?Keywords')
quick('Viewpoint', 'http://search.viewpoint.com/?k')
quick('TheSearchEngine', 'http://www.thesearchengine.org/?src')

all_regexes = []
for engine in all_engines:
    for regex in engine.regexes:
        all_regexes.append((regex, engine))

class AnnotateSearchEngine(annotate.Annotate):

    """
    Adds an attribute ``search_engine``, which will be the search
    engine description if the request was referred by a search engine
    (or None otherwise).  If a search engine, adds ``search_query``,
    which is url-quoted string query.
    """

    require = ('referer',)
    provides = ('search_engine', 'search_query',)

    def __call__(self, log):
        referer = log.referer
        for regex, engine in all_regexes:
            match = regex.search(referer)
            if match:
                query = match.group(1)
                log.search_engine = engine
                log.search_query = query
                return log
        log.search_engine = None
        return log

class AnnotateKeywords(annotate.Annotate):

    """
    Adds ``search_keywords``, which is a sorted (alphabetically) list
    of the words in the ``search_query`` (if any query exists, of
    course).
    """

    require = ('search_engine', 'search_query')
    provides = ('search_keywords',)

    _nonletter_re = re.compile('[^a-zA-Z0-9: ]')

    def __call__(self, log):
        if not log.search_engine:
            return log
        query = log.search_query
        query = urllib.unquote(query)
        query = query.replace('+', ' ')
        if '%' in query:
            # @@ Some queries are double-quoted
            query = urllib.unquote(query)
        query = self._nonletter_re.sub('', query)
        query = query.lower()
        words = filter(None, query.split())
        words_orig = words[:]
        words.sort()
        words = tuple(words)
        log.search_keywords = words
        log.search_keywords_orig = words_orig
        return log