from mod_python import apache, util
from PyLucene import QueryParser, IndexSearcher, StandardAnalyzer, FSDirectory
replace = ["_"] # these chars will be replaced by a space
escape = ["\\", "+", "-", "&&", "||", "!", "(", ")", "{", "}", "[", "]", "^", "\"", "~", "*", "?", ":"] # these chars will be escaped
lower = ["AND", "OR", "NOT"] # these words will be forced to lowercase
fncts = ["\\", "/", "_", "-", "+", ":", ".", ",", "|"] # characters in filename we should convert to spaces
maxhits = 100 # max amount we'll ever return
maxdurationdiff = 250000 # max difference in duration
metasearch = None
def handler(req):
global metasearch
if not metasearch:
metasearch = MetaSearch()
search = createquery(util.FieldStorage(req))
if search == "":
req.content_type = "text/html"
req.write("")
req.write("")
req.write("")
req.write("Error")
req.write("")
req.write("")
req.write("Hi there, apparently you're new to this :)
")
req.write("You'll have to specify the parameters, the valid ones are:
")
req.write("")
req.write("- limit (amount of tracks returned, 1-100, default 50)
")
req.write("- durationdiff (max duration difference given in milliseconds, 0-100000, default 10000)
")
req.write("- artistid
")
req.write("- releaseid
")
req.write("- trackid
")
req.write("- artist
")
req.write("- release
")
req.write("- track
")
req.write("- tracknum
")
req.write("- duration (milliseconds)
")
req.write("- filename
")
req.write("
")
req.write("Brief example:
")
req.write("http://mb.samfundet.no/?artist=europe&track=countdown
")
req.write("")
req.write("")
else:
req.content_type = "text/xml"
req.write("\n")
req.write("")
req.write(metasearch.asXML(metasearch.search(search), limit))
req.write("")
return apache.OK
def createquery(fs):
if fs.getfirst("limit", [""])[0].isdigit():
limit = int(fs.getfirst("limit", [50])[0])
else:
limit = 50
if fs.getfirst("durationdiff", [""])[0].isdigit():
durationdiff = int(fs.getfirst("durationdiff", [10000])[0])
else:
durationdiff = 10000
if durationdiff < 1000:
durationdiff = 1000
if durationdiff > maxdurationdiff:
durationdiff = maxdurationdiff
if limit < 1:
limit = 1
if limit > maxhits:
limit = maxhits
search = ""
if fs.getfirst("artistid"):
search += "artistid:" + fixstring(fs.getfirst("artistid")) + "^4 "
if fs.getfirst("releaseid"):
search += "releaseid:" + fixstring(fs.getfirst("releaseid")) + "^4 "
if fs.getfirst("trackid"):
search += "trackid:" + fixstring(fs.getfirst("trackid")) + "^4 "
if fs.getfirst("artist"):
search += fixstring(fs.getfirst("artist")) + " "
if fs.getfirst("release"):
search += fixstring(fs.getfirst("release")) + " "
if fs.getfirst("track"):
search += fixstring(fs.getfirst("track")) + " "
if fs.getfirst("filename"):
search += fixstring(fs.getfirst("filename")) + " "
if fs.getfirst("tracknum") and fs.getfirst("tracknum").isdigit():
search += "tracknum:" + str(int(fs.getfirst("tracknum"))) + "^2 " + str(int(fs.getfirst("tracknum"))) + " "
if fs.getfirst("duration") and fs.getfirst("duration").isdigit():
minvalue = (int(fs.getfirst("duration")) - durationdiff) / 1000
if minvalue < 0:
minvalue = 0
maxvalue = (int(fs.getfirst("duration")) + durationdiff) / 1000
# why can't we use range search [x TO y]?
# because lucene thinks 4000-4200, 40000-42000 and so on is between 400-420
# most likely because lucene handles text, not numeric values
search += "+tracklengthsec:("
for a in xrange(minvalue, maxvalue + 1):
search += str(a) + " "
search += "0)"
search = search.strip()
return search
def fixstring(string):
for a in xrange(len(replace)):
string = string.replace(replace[a], " ")
for a in xrange(len(escape)):
string = string.replace(escape[a], "\\" + escape[a])
for a in xrange(len(lower)):
string = string.replace(lower[a], lower[a].lower())
return string
class MetaSearch(object):
def __init__(self):
self.searcher = IndexSearcher(FSDirectory.getDirectory("/export/mb/index", False))
self.querypar = QueryParser("content", StandardAnalyzer([]))
def search(self, search):
query = self.querypar.parse(search)
return self.searcher.search(query)
def asXML(self, hits, maxhits):
'''
Output result from metasearch as XML
'''
out = "\t\n"
for i in xrange(min(hits.length(), maxhits)):
doc = hits.doc(i)
out += "\t\t\n"
out += "\t"
return out
def escape(self, text):
'''
Escape XML/HTML entities and convert output to utf-8
'''
return (text.replace(u'&', u'&').replace(u'<', u'<').replace(u'>', u'>')).encode('utf-8', 'replace')