script for testing facet queries using a compressed bitset

bitset_facet_query.py

import sys 
import urllib2
from invenio import intbitset
from optparse import OptionParser
import simplejson
import mimetools

if __name__ == '__main__':

    op = OptionParser()
    op.set_usage("usage: bitset_facet_query.py [options] ")
    op.add_option('--query', dest='query', action='store',
        help='solr query', type=str, default="galaxy cluster")
    op.add_option('--solr_url', dest='solr_url', action='store',
        help='solr url', type=str, default="http://localhost:8983/solr")
    opts, args = op.parse_args()

    facet_query_url = "%s/invenio_facets" % opts.solr_url
    invenio_query_url = "%s/select?qt=invenio_query&q=%s" % (opts.solr_url, urllib2.quote(opts.query))
    print invenio_query_url

    # query to get a bitset
    bitset = intbitset.intbitset()
    u = urllib2.urlopen(invenio_query_url)
    data = u.read()
    bitset.fastload(data)
    print "query result: " + str(bitset)
    print "results length: %d" % len(bitset.tolist())

    # now use the bitset to fetch the facet data
    r = urllib2.Request(facet_query_url)
    data = bitset.fastdump()
    boundary = mimetools.choose_boundary()

    # fool solr into thinking we're uploading a file so it will read our data as a stream
    data_size = len(data)
    contents = '--%s\r\n' % boundary
    contents += 'Content-Disposition: form-data; name="bitset"; filename="bitset"\r\n'
    contents += 'Content-Type: application/octet-stream\r\n'
    contents += '\r\n' + data + '\r\n'
    contents += '--%s--\r\n\r\n' % boundary
    r.add_data(contents)

    contenttype = 'multipart/form-data; boundary=%s' % boundary
    r.add_unredirected_header('Content-Type', contenttype)

    # post the request and get back the facets as json
    u = urllib2.urlopen(r)
    print simplejson.load(u)