BULZRtyp

1# -*- coding: utf-8 -*-
2
3
4'''
5edgelistとlabellistを0始まりのIDに変換する
6'''
7
8import os
9import sys
10#from snlocest.largedict import LargeDict
11
12
13def parse_args():
14    import argparse
15    parser = argparse.ArgumentParser()
16    parser.add_argument('mode', help='処理のモード')
17    parser.add_argument('--tablepath', required=True)
18    parser.add_argument('inputfiles', nargs='+')
19    return parser.parse_args()
20
21
22def load_table(filepath):
23    user2id = dict() #LargeDict()
24    with open(filepath, 'r') as fd:
25        for line in fd:
26            user_id, idx = line.rstrip().split('\t')
27            user2id[user_id] = int(idx)
28    idcnt = max(user2id.values()) + 1
29    return user2id, idcnt
30
31
32if __name__ == '__main__':
33    args = parse_args()
34
35    if args.mode == 'table':
36        # Generate convert table from inputfiles and save to tablepath
37        if os.path.exists(args.tablepath):
38            print('Table path already exists', file=sys.stderr)
39            sys.exit(1)
40        # create table
41        user2id = dict() #LargeDict()
42        idcnt = 0
43        for filepath in args.inputfiles:
44            with open(filepath, 'r') as inputfile:
45                for line in inputfile:
46                    tokens = line.rstrip().split('\t')
47                    user_id = tokens[0]
48                    if user_id not in user2id:
49                        user2id[user_id] = idcnt
50                        idcnt += 1
51                    user_id = tokens[1]
52                    if user_id not in user2id:
53                        user2id[user_id] = idcnt
54                        idcnt += 1
55        # save table
56        with open(args.tablepath, 'w') as fd:
57            for k, v in user2id.items():
58                print(k, v, sep='\t', file=fd)
59
60    elif args.mode == 'edgelist':
61        # Convert edgelist using table
62        user2id, idcnt = load_table(args.tablepath)
63
64        for filepath in args.inputfiles:
65            with open(filepath, 'r') as fd:
66                for line in fd:
67                    row = line.rstrip().split('\t')
68                    src = row[0]
69                    dst = row[1]
70                    print(user2id[src], user2id[dst], *row[2:], sep='\t')
71
72    elif args.mode == 'label':
73        # Convert label file
74        user2id, idcnt = load_table(args.tablepath)
75
76        for filepath in args.inputfiles:
77            with open(filepath, 'r') as fd:
78                for line in fd:
79                    row = line.rstrip().split('\t')
80                    label = row[0]
81                    print(user2id[label], *row[1:], sep='\t')
82
83    else:
84        print('Invalid mode. Choose "table" or "edgelist" or "label"', file=sys.stderr)