def test_queries(self):
query_fh = cStringIO.StringIO(self.test_queries)
queries = qu.Queries(query_fh, self.test_num_features)
query = queries['1']
query_fh.close()
self.assertEqual(4, query.get_document_count())
self.assertEqual(4, len(query.get_feature_vectors()))
self.assertEqual([0, 1, 2, 3], [d.docid for d in query.get_docids()])
# TODO: do "labels" have to be np array? not a list?
self.assertEqual([4, 1, 0, 0], query.get_labels().tolist())
# self.assertEqual(1, query.get_label(1)) TODO: FIX
self.assertEqual(None, query.get_predictions())
self.assertEqual(None, query.get_comments())
self.assertEqual(None, query.get_comment(0))
def setUp(self):
# initialize query
self.test_num_features = 6
test_query = """
4 qid:1 1:2.6 2:1 3:2.1 4:0 5:2 6:1.4 # highly relevant
1 qid:1 1:1.2 2:1 3:2.9 4:0 5:2 6:1.9 # bad
0 qid:1 1:0.5 2:1 3:2.3 4:0 5:2 6:5.6 # not relevant
0 qid:1 1:0.5 2:1 3:2.3 4:0 5:2 6:5.6 # not relevant
"""
self.query_fh = cStringIO.StringIO(test_query)
self.queries = query.Queries(self.query_fh, self.test_num_features)
self.query = self.queries['1']
# initialize pairwise learner
self.learner = PairwiseLearningSystem(self.test_num_features,
"--init_weights 0,0,1,0,0,0 --epsilon 0.0 --eta 0.001 --ranker "
"ranker.DeterministicRankingFunction --ranker_tie first")
def test_query_with_comments(self):
query_fh = cStringIO.StringIO(self.test_query)
queries = qu.Queries(query_fh, self.test_num_features, True)
query = queries['1']
query_fh.close()
self.assertEqual(4, query.get_document_count())
self.assertEqual(4, len(query.get_feature_vectors()))
self.assertEqual([0, 1, 2, 3], query.get_docids())
# TODO: do "labels" have to be np array? not a list?
self.assertEqual([4, 1, 0, 0], query.get_labels().tolist())
self.assertEqual(1, query.get_label(1))
self.assertEqual(None, query.get_predictions())
self.assertEqual(
["# highly relevant", "# bad", "# not relevant", "# not relevant"],
query.get_comments())
self.assertEqual("# highly relevant", query.get_comment(0))
def setUp(self):
# initialize query
self.test_num_features = 6
test_query = """
4 qid:1 1:2.6 2:1 3:2.1 4:0 5:2 6:1.4 # highly relevant
1 qid:1 1:1.2 2:1 3:2.9 4:0 5:2 6:1.9 # bad
0 qid:1 1:0.5 2:1 3:2.3 4:0 5:2 6:5.6 # not relevant
0 qid:1 1:0.5 2:1 3:2.3 4:0 5:2 6:5.6 # not relevant
"""
self.query_fh = cStringIO.StringIO(test_query)
self.queries = query.Queries(self.query_fh, self.test_num_features)
self.query = self.queries['1']
# initialize listwise learner
self.learner = ListwiseLearningSystem(self.test_num_features,
"--init_weights 0,0,1,0,0,0 --delta 1.0 --alpha 0.01 --ranker "
"ranker.ProbabilisticRankingFunction --ranker_args 3 --ranker_tie "
"first --comparison comparison.ProbabilisticInterleaveWithHistory"
" --comparison_args \"--history_length 10 --biased true\"")
def setUp(self):
self.test_num_features = 6
test_query = """
4 qid:1 1:2.6 2:1 3:2.1 4:0 5:2 6:1.4 # highly relevant
1 qid:1 1:1.2 2:1 3:2.9 4:0 5:2 6:1.9 # bad
0 qid:1 1:0.5 2:1 3:2.3 4:0 5:2 6:5.6 # not relevant
0 qid:1 1:0.5 2:1 3:2.3 4:0 5:2 6:5.6 # not relevant
"""
self.query_fh = cStringIO.StringIO(test_query)
self.queries = query.Queries(self.query_fh, self.test_num_features)
self.query = self.queries['1']
zero_weight_str = "0 0 0 0 0 0"
self.zero_weights = np.asarray(
[float(x) for x in zero_weight_str.split()])
weight_str = "0 0 1 0 0 0"
self.weights = np.asarray([float(x) for x in weight_str.split()])
def setUp(self):
self.test_num_features = 6
test_query = """
1 qid:1 1:2.6 2:1 3:2.1 4:0 5:2 6:1.4 # relevant
1 qid:1 1:1.2 2:1 3:2.9 4:0 5:2 6:1.9 # relevant
0 qid:1 1:0.5 2:1 3:2.3 4:0 5:2 6:5.6 # not relevant
0 qid:1 1:0.5 2:1 3:2.3 4:0 5:2.1 6:5.6 # not relevant
"""
self.query_fh = cStringIO.StringIO(test_query)
self.queries = qu.Queries(self.query_fh, self.test_num_features)
self.query = self.queries['1']
zero_weight_str = "0 0 0 0 0 0"
self.zero_weights = np.asarray(
[float(x) for x in zero_weight_str.split()])
# results in ranking: 1, 3, 2, 0
weight_str_1 = "0 0 1 0 1 0"
self.weights_1 = np.asarray([float(x) for x in weight_str_1.split()])
weight_str_2 = "1 0 0 0 1 0"
self.weights_2 = np.asarray([float(x) for x in weight_str_2.split()])
import os
import cStringIO
sys.path.insert(0, os.path.abspath('..'))
import query
from numpy import array
from PairwiseLearningSystem import PairwiseLearningSystem
# initialize query
test_num_features = 64
f = open('train.txt', 'r')
test_query = f.read()
f.close()
query_fh = cStringIO.StringIO(test_query)
queries = query.Queries(query_fh, test_num_features)
# query = queries['1'] #?
def addRanker(rankers, path):
f = open(path, 'r')
str = f.read().lstrip('final_weights: [')
str = str.rstrip('] \n')
rankers.append(str)
f.close()
return rankers
rankers = []
rankers = addRanker(rankers, 'features64/ranker-00.txt')
# print rankers[0]
def test_queries(self):
query_fh = cStringIO.StringIO(self.test_queries)
queries = qu.Queries(query_fh, self.test_num_features)
query_fh.close()
self.assertEqual(1, queries.get_size())
