def makeSearchString(q, index, limits, handler=None): # sort,
"""translates search parameters into Solr query syntax."""
q = q.replace("_", " ")
# strip out stopwords here... this is necessary due to some
# unexpected side-effects of the general keyword search
# TODO: do NOT strip stopwords if a quoted search, eg. "the the" or "to be or not to be"
qWords = q.split(" ")
nonStopWords = []
for wordOn in qWords:
if wordOn not in STOPWORDS:
nonStopWords.append( wordOn )
q = " ".join(nonStopWords)
for orig,replacement in SEARCH_CHARACTER_REPLACEMENTS.iteritems():
q = q.replace(orig,replacement)
q = urllib.quote( q ).strip()
if not handler or (handler is "standard"):
if index in allFacetCodes:
# then treat as an exact match search; it is a facet, not free text entered by user
ret = '%s:"%s"' % (index, q)
else:
# then it is a search index, not a facet -- NOT exact match
ret = '%s:%s' % (index, q)
else:
# if you specify a handler you can't also specify an index...
ret = '%s' % q
for limitOn in limits:
# csdebug: how are these forbidden characters making it this far?
_limOn = urllib.unquote( limitOn )
limitSplit = _limOn.split(":")
logger.error("limitSplit is %s\n\n\n\n" % limitSplit) # csdebug
index = limitSplit[0] # todo error handling
term = ":".join( limitSplit[1:] )
for orig,replacement in SEARCH_CHARACTER_REPLACEMENTS.iteritems():
term = term.replace(orig,replacement)
logger.error("\n\nterm is now %s\n\n" % term)
logger.error("\n\nxxterm is now %s\n\n" % term)
term = term.replace("_", " ")
#.replace('"', "")
term = term.strip()
logger.error("\n\nyyterm is now %s\n\n" % term)
# we are going to put exact quotes around the whole thing, so we don't want to double-quote
term = urllib.quote(term)
logger.error("\n\n!!!term is now %s\n\n" % term)
ret = """%s AND %s:"%s\"""" % ( ret, index, term)
# get rid of any double spaces.
ret = ret.replace(" ", " ")
ret = ret.replace(" ", "%20")
logger.debug( "search string is %s" % ret )
return ret.strip()
def spellCheck( phraseToCheck ):
"""takes the phrase to check and returns a list of potential suggestions."""
ret = []
if not config.USE_YAHOO_SPELLING_WEB_SERVICE:
pass
else:
# TODO: will it handle utf8?
# check cache first.
cacheKey = "spellcheck~~%s~~" % phraseToCheck
suggestionsFromCache = cache.get( cacheKey )
if not suggestionsFromCache:
query = urllib.quote( phraseToCheck )
urlToGet = config.YAHOO_WEB_SERVICE_URL % dict( YAHOO_APPID = config.YAHOO_APPID, query = query )
logger.debug("fetching URL %s" % urlToGet )
data = urllib.urlopen( urlToGet ).read()
try:
respObject = simplejson.loads( data )
# note we do this to prevent caching bad data
cache.set( cacheKey, data, config.YAHOO_SPELLCHECK_CACHE_TIME)
# TODO: figure out if it will *ever* return multiple suggestions...
except:
logger.error( "exception doing spellCheck service" )
else:
logger.debug( "got spelling suggestion from cache")
respObject = simplejson.loads( suggestionsFromCache )
#print "respObject is %s, type is %s" % (respObject, type(respObject) ) # csdebug
if respObject.has_key("ResultSet") and type(respObject['ResultSet']) == types.DictType and respObject['ResultSet'].has_key("Result"):
ret.append( respObject['ResultSet']['Result'] )
logger.debug("returning %s" % ret)
return ret
#return ['pants', 'pantaloons'] # csdebug