Conv2Query/evaluate_ranking.py at main · ChuanMeng/Conv2Query · GitHub

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
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
import argparse
import os
import pytrec_eval
import json
from npdcg import calculate_npdcg

mapping = {"ndcg_cut_3": "ndcg@3",
           "ndcg_cut_5": "ndcg@5",
           "ndcg_cut_10": "ndcg@10",
           "ndcg_cut_20": "ndcg@20",
           "ndcg_cut_100": "ndcg@100",
           "ndcg_cut_1000": "ndcg@1000",
           "mrr_5": "mrr@5",
           "mrr_10": "mrr@10",
           "mrr_20": "mrr@20",
           "mrr_100": "mrr@100",
           "map_cut_10": "map@10",
           "map_cut_100": "map@100",
           "map_cut_1000": "map@1000",
           "recall_5": "recall@5",
           "recall_20": "recall@20",
           "recall_100": "recall@100",
           "recall_1000": 'recall@1000',
           "P_1": "precision@1",
           "P_3": "precision@3",
           "P_5": "precision@5",
           "P_10": "precision@10",
           "P_100": "precision@100",
           }

if __name__ == "__main__":
    parser = argparse.ArgumentParser()

    parser.add_argument('--run_dir', type=str, required=True)
    parser.add_argument('--qrels_dir', type=str, required=True)
    parser.add_argument('--output_path', type=str)
    parser.add_argument('--rel_scale', type=int, required=True)

    args = parser.parse_args()


    with open(args.run_dir, 'r') as r:
        run = pytrec_eval.parse_run(r)

    with open(args.qrels_dir, 'r') as r:
        qrel = pytrec_eval.parse_qrel(r)

    print("len(list(run))", len(list(run)))
    print("len(list(qrel))", len(list(qrel)))


    # metric averaged on queries

    avg = {}

    # judged@10
    q2judge_10 = {}
    q2judge_20 = {}
    q2judge_100 = {}

    for qid, did_score in run.items():

        if qid not in qrel:
            continue

        sorted_did = [did for did, score in sorted(did_score.items(), key=lambda item: item[1], reverse=True)]
        judge_list = []

        for docid in sorted_did:
            if docid in qrel[qid]:
                judge_list.append(1)
            else:
                judge_list.append(0)

        q2judge_10[qid]=sum(judge_list[0:10])/10
        q2judge_20[qid]=sum(judge_list[0:20])/20
        q2judge_100[qid]=sum(judge_list[0:100])/100

    print('{}: {:.4f}'.format("judge_10", sum(q2judge_10.values())/len(q2judge_10)))
    print('{}: {:.4f}'.format("judge_20", sum(q2judge_20.values()) /len(q2judge_20)))
    print('{}: {:.4f}'.format("judge_100", sum(q2judge_100.values()) /len(q2judge_100)))

    avg[f"judge@{10}"] = sum(q2judge_10.values())/len(q2judge_10)
    avg[f"judge@{20}"] = sum(q2judge_20.values()) /len(q2judge_20)
    avg[f"judge@{100}"] = sum(q2judge_100.values()) /len(q2judge_100)

    run_5 = {}
    run_10 = {}
    run_20 = {}
    run_100 = {}

    for qid, did_score in run.items():
        sorted_did_score = [(did, score) for did, score in
                            sorted(did_score.items(), key=lambda item: item[1], reverse=True)]
        run_5[qid] = dict(sorted_did_score[0:5])
        run_10[qid] = dict(sorted_did_score[0:10])
        run_20[qid] = dict(sorted_did_score[0:20])
        run_100[qid] = dict(sorted_did_score[0:100])

    evaluator_ndcg = pytrec_eval.RelevanceEvaluator(qrel, {'ndcg_cut_3', 'ndcg_cut_5', 'ndcg_cut_10', 'ndcg_cut_20', 'ndcg_cut_100',
                                                           'ndcg_cut_1000'})
    results_ndcg = evaluator_ndcg.evaluate(run)

    results = {}
    for qid, _ in results_ndcg.items():
        results[qid] = {}
        for measure, score in results_ndcg[qid].items():
            results[qid][mapping[measure]] = score

    for q_id, pid_rel in qrel.items():
        for p_id, rel in pid_rel.items():
            if int(rel) >= args.rel_scale:
                qrel[q_id][p_id] = 1
            else:
                qrel[q_id][p_id] = 0

    evaluator_general = pytrec_eval.RelevanceEvaluator(qrel, {'map_cut_10', 'map_cut_100', 'map_cut_1000', 'recall_5', 'recall_20',
                                                              'recall_100', 'recall_1000', 'P_1', 'P_3', 'P_5', 'P_10',
                                                              "P_100"})
    results_general = evaluator_general.evaluate(run)

    for qid, _ in results.items():
        for measure, score in results_general[qid].items():
            results[qid][mapping[measure]] = score

    evaluator_rr = pytrec_eval.RelevanceEvaluator(qrel, {'recip_rank'})
    results_rr_5 = evaluator_rr.evaluate(run_5)
    results_rr_10 = evaluator_rr.evaluate(run_10)
    results_rr_20 = evaluator_rr.evaluate(run_20)
    results_rr_100 = evaluator_rr.evaluate(run_100)

    for qid, _ in results.items():
        results[qid][mapping["mrr_5"]] = results_rr_5[qid]['recip_rank']
        results[qid][mapping["mrr_10"]] = results_rr_10[qid]['recip_rank']
        results[qid][mapping["mrr_20"]] = results_rr_20[qid]['recip_rank']
        results[qid][mapping["mrr_100"]] = results_rr_100[qid]['recip_rank']

    for measure in mapping.values():
        overall = pytrec_eval.compute_aggregated_measure(measure, [result[measure] for result in results.values()])
        print('{}: {:.4f}'.format(measure, overall))
        avg[measure]=overall


    if "procis.test" in args.qrels_dir:
        # npdcg
        npdcg = []
        cuttoffs = [5, 10, 20]

        conv = {}
        for qid in qrel.keys():
            conv_id = qid.split("_")[0]
            if conv_id not in conv:
                conv[conv_id] = []
            conv[conv_id].append(qid)

        conv_max_turn = {}
        for conv_id in conv.keys():
            conv_max_turn[conv_id] =[]
            for qid in conv[conv_id]:
                conv_max_turn[conv_id].append(int(qid.split("_")[1]))
        for conv_id in conv_max_turn.keys():
            conv_max_turn[conv_id]=max(conv_max_turn[conv_id])

        # queries_proactive [[q11,q12,...],[q21,q22,...]...]
        for conv_id in conv.keys():
            retrieved = []

            for rank in range(1,(conv_max_turn[conv_id]+1)):
            #for qid in conv[conv_id]:
                qid = f"{conv_id}_{rank}"

                if qid in conv[conv_id]:
                    # per turn
                    retrieved_docs = [did for did, score in sorted(run[qid].items(), key=lambda item: item[1], reverse=True)]
                    correct_docs = [(did, socre) for did, socre in qrel[qid].items()]
                else:
                    retrieved_docs = []
                    correct_docs = []

                retrieved.append({'retrieved_docs': retrieved_docs, 'correct_docs': correct_docs})

            npdcg.append(calculate_npdcg(retrieved, [5, 10, 20]))

        # calculate average npdcg per cutoff, calculate_npdcg returns dict
        npdcg_avg = {c: sum([npdcg[i][c] for i in range(len(npdcg))]) / len(npdcg) for c in cuttoffs}
        for c in npdcg_avg.keys():
            print('npdcg@{}: {:.4f}'.format(c, npdcg_avg[c]))
            avg[f"npdcg@{c}"]=npdcg_avg[c]

    if args.output_path is not None:
        args.output_dir = "/".join(args.output_path.split("/")[0:-1])
        if not os.path.exists(args.output_dir):
            os.makedirs(args.output_dir)

        with open(f"{args.output_path}", 'w') as w:
            w.write(json.dumps(avg))