Spaces:
Runtime error
Runtime error
File size: 6,528 Bytes
d6585f5 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 |
#
# Pyserini: Reproducible IR research with sparse and dense representations
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
"""Integration tests for TCT-ColBERTv2 models using on-the-fly query encoding."""
import os
import socket
import unittest
from integrations.utils import clean_files, run_command, parse_score
from pyserini.search import QueryEncoder
from pyserini.search import get_topics
class TestSearchIntegration(unittest.TestCase):
def setUp(self):
self.temp_files = []
self.threads = 16
self.batch_size = 256
# Hard-code larger values for internal servers
if socket.gethostname().startswith('damiano') or socket.gethostname().startswith('orca'):
self.threads = 36
self.batch_size = 144
def test_msmarco_passage_tct_colbert_v2_bf_otf(self):
output_file = 'test_run.msmarco-passage.tct_colbert-v2.bf-otf.tsv'
self.temp_files.append(output_file)
cmd1 = f'python -m pyserini.search.faiss --topics msmarco-passage-dev-subset \
--index msmarco-passage-tct_colbert-v2-bf \
--encoder castorini/tct_colbert-v2-msmarco \
--batch-size {self.batch_size} \
--threads {self.threads} \
--output {output_file} \
--output-format msmarco'
cmd2 = f'python -m pyserini.eval.msmarco_passage_eval msmarco-passage-dev-subset {output_file}'
status = os.system(cmd1)
stdout, stderr = run_command(cmd2)
score = parse_score(stdout, "MRR @10")
self.assertEqual(status, 0)
self.assertAlmostEqual(score, 0.3440, delta=0.0001)
def test_msmarco_passage_tct_colbert_v2_hn_otf(self):
output_file = 'test_run.msmarco-passage.tct_colbert-v2-hn.bf-otf.tsv'
self.temp_files.append(output_file)
cmd1 = f'python -m pyserini.search.faiss --topics msmarco-passage-dev-subset \
--index msmarco-passage-tct_colbert-v2-hn-bf \
--encoder castorini/tct_colbert-v2-hn-msmarco \
--batch-size {self.batch_size} \
--threads {self.threads} \
--output {output_file} \
--output-format msmarco'
cmd2 = f'python -m pyserini.eval.msmarco_passage_eval msmarco-passage-dev-subset {output_file}'
status = os.system(cmd1)
stdout, stderr = run_command(cmd2)
score = parse_score(stdout, "MRR @10")
self.assertEqual(status, 0)
self.assertAlmostEqual(score, 0.3543, delta=0.0001)
def test_msmarco_passage_tct_colbert_v2_hnp_bf_bm25_hybrid_otf(self):
output_file = 'test_run.msmarco-passage.tct_colbert-v2-hnp.bf-otf.bm25.tsv'
self.temp_files.append(output_file)
cmd1 = f'python -m pyserini.search.hybrid dense --index msmarco-passage-tct_colbert-v2-hnp-bf \
--encoder castorini/tct_colbert-v2-hnp-msmarco \
sparse --index msmarco-passage \
fusion --alpha 0.06 \
run --topics msmarco-passage-dev-subset \
--output {output_file} \
--batch-size {self.batch_size} --threads {self.threads} \
--output-format msmarco'
cmd2 = f'python -m pyserini.eval.msmarco_passage_eval msmarco-passage-dev-subset {output_file}'
status = os.system(cmd1)
stdout, stderr = run_command(cmd2)
score = parse_score(stdout, "MRR @10")
self.assertEqual(status, 0)
self.assertAlmostEqual(score, 0.3682, delta=0.0001)
def test_msmarco_passage_tct_colbert_v2_hnp_bf_d2q_hybrid_otf(self):
output_file = 'test_run.msmarco-passage.tct_colbert-v2-hnp.bf-otf.doc2queryT5.tsv'
self.temp_files.append(output_file)
cmd1 = f'python -m pyserini.search.hybrid dense --index msmarco-passage-tct_colbert-v2-hnp-bf \
--encoder castorini/tct_colbert-v2-hnp-msmarco \
sparse --index msmarco-passage-expanded \
fusion --alpha 0.1 \
run --topics msmarco-passage-dev-subset \
--output {output_file} \
--batch-size {self.batch_size} --threads {self.threads} \
--output-format msmarco'
cmd2 = f'python -m pyserini.eval.msmarco_passage_eval msmarco-passage-dev-subset {output_file}'
status = os.system(cmd1)
stdout, stderr = run_command(cmd2)
score = parse_score(stdout, "MRR @10")
self.assertEqual(status, 0)
self.assertAlmostEqual(score, 0.3731, delta=0.0001)
def test_msmarco_passage_tct_colbert_v2_encoded_queries(self):
encoded = QueryEncoder.load_encoded_queries('tct_colbert-v2-msmarco-passage-dev-subset')
topics = get_topics('msmarco-passage-dev-subset')
for t in topics:
self.assertTrue(topics[t]['title'] in encoded.embedding)
def test_msmarco_passage_tct_colbert_v2_hn_encoded_queries(self):
encoded = QueryEncoder.load_encoded_queries('tct_colbert-v2-hn-msmarco-passage-dev-subset')
topics = get_topics('msmarco-passage-dev-subset')
for t in topics:
self.assertTrue(topics[t]['title'] in encoded.embedding)
def test_msmarco_passage_tct_colbert_v2_hnp_encoded_queries(self):
encoded = QueryEncoder.load_encoded_queries('tct_colbert-v2-hnp-msmarco-passage-dev-subset')
topics = get_topics('msmarco-passage-dev-subset')
for t in topics:
self.assertTrue(topics[t]['title'] in encoded.embedding)
def tearDown(self):
clean_files(self.temp_files)
if __name__ == '__main__':
unittest.main()
|