File size: 6,627 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
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
#
# 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.
#
import argparse
from typing import List
from tqdm import tqdm
from transformers import AutoTokenizer
from pyserini.search import get_topics
from pyserini.search.lucene.irst import LuceneIrstSearcher


def normalize(scores: List[float]):
    low = min(scores)
    high = max(scores)
    width = high - low
    if width != 0:
        return [(s-low)/width for s in scores]
    return scores


def query_loader(topic: str):
    queries = {}
    bert_tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
    topics_dic = get_topics(topic)
    line_num = 0
    for topic_id in topics_dic:
        line_num += 1
        query_text = topics_dic[topic_id]['title']
        text_bert_tok = bert_tokenizer.tokenize(query_text.lower())
        if len(text_bert_tok) >= 0:
            query = {"raw": query_text,
                "contents": ' '.join(text_bert_tok)}
            queries[topic_id] = query
        if line_num % 10000 == 0:
            print(f"Processed {line_num} queries")
    print(f"Processed {line_num} queries")
    return queries


def baseline_loader(base_path: str):
    result_dic = {}
    with open(base_path, 'r') as f:
        for line in f:
            tokens = line.strip().split()
            topic = tokens[0]
            doc_id = tokens[2]
            score = float(tokens[-2])
            if topic in result_dic.keys():
                result_dic[topic][0].append(doc_id)
                result_dic[topic][1].append(score)
            else:
                result_dic[topic] = [[doc_id], [score]]

    return result_dic


def generate_maxP(preds: List[float], docs: List[str]):
    scores = {}
    for index, (score, doc_id) in enumerate(zip(preds, docs)):
        docid = doc_id.split('#')[0]
        if (docid not in scores or score > scores[docid]):
            scores[docid] = score
    docid_scores = sorted(scores.items(), key=lambda kv: kv[1], reverse=True)
    return docid_scores


def sort_dual_list(pred: List[float], docs: List[str]):
    zipped_lists = zip(pred, docs)
    sorted_pairs = sorted(zipped_lists)

    tuples = zip(*sorted_pairs)
    pred, docs = [list(tuple) for tuple in tuples]

    pred.reverse()
    docs.reverse()
    return pred, docs


if __name__ == "__main__":
    parser = argparse.ArgumentParser(
        description='use ibm model 1 feature to rerank the base run file')
    parser.add_argument('--tag', type=str, default="ibm",
                        metavar="tag_name", help='tag name for resulting Qrun')
    parser.add_argument('--base-path', type=str, required=False,
                        metavar="path_to_base_run", help='path to base run')
    parser.add_argument('--topics', type=str, required=True,
                        help='existing topics name or path to query topics')
    parser.add_argument('--index', type=str, required=True,
                        metavar="path_to_lucene_index", help='path to lucene index folder')
    parser.add_argument('--output', type=str, required=True,
                        metavar="path_to_reranked_run", help='the path to store reranked run file')
    parser.add_argument('--alpha', type=float, default="0.3",
                        metavar="type of field", help='interpolation weight')
    parser.add_argument('--num-threads', type=int, default="24",
                        metavar="num_of_threads", help='number of threads to use')
    parser.add_argument('--max-sim', default=False, action="store_true",
                        help='whether we use max sim operator or avg instead')
    parser.add_argument('--segments', default=False, action="store_true",
                        help='whether we use segmented index or not')
    parser.add_argument('--k1', type=float, default="0.81",
                        metavar="bm25_k1_parameter", help='k1 parameter for bm25 search')
    parser.add_argument('--b', type=float, default="0.68",
                        metavar="bm25_b_parameter", help='b parameter for bm25 search')
    parser.add_argument('--hits', type=int, metavar='number of hits generated in runfile',
                        required=False, default=1000, help="Number of hits.")
    args = parser.parse_args()

    print('Using max sim operator or not:', args.max_sim)

    f = open(args.output, 'w')

    reranker = LuceneIrstSearcher(args.index, args.k1, args.b, args.num_threads)
    queries = query_loader(args.topics)
    query_text_lst = [queries[topic]['raw'] for topic in queries.keys()]
    qid_lst = [str(topic) for topic in queries.keys()]
    i = 0
    for topic in queries:
        if i % 100 == 0:
            print(f'Reranking {i} topic')
        query_text_field = queries[topic]['contents']
        query_text = queries[topic]['raw']
        bm25_results = reranker.bm25search.search(query_text, args.hits)
        if args.base_path:
            baseline_dic = baseline_loader(args.base_path)
            docids, rank_scores, base_scores = reranker.rerank(
                query_text, query_text_field, baseline_dic[topic], args.max_sim, bm25_results)
        else:
            docids, rank_scores, base_scores = reranker.search(
                query_text, query_text_field, args.max_sim, bm25_results)
        ibm_scores = normalize([p for p in rank_scores])
        base_scores = normalize([p for p in base_scores])

        interpolated_scores = [
            a * args.alpha + b * (1-args.alpha) for a, b in zip(base_scores, ibm_scores)]

        preds, docs = sort_dual_list(interpolated_scores, docids)
        i = i+1
        if args.segments:
            docid_scores = generate_maxP(preds, docs)
            rank = 1
            for doc_id, score in docid_scores:
                if rank > 1000:
                    break
                f.write(f'{topic} Q0 {doc_id} {rank} {score} {args.tag}\n')
                rank = rank + 1
        else:
            for index, (score, doc_id) in enumerate(zip(preds, docs)):
                rank = index + 1
                f.write(f'{topic} Q0 {doc_id} {rank} {score} {args.tag}\n')
    f.close()