-
Notifications
You must be signed in to change notification settings - Fork 3
/
prepare.py
157 lines (131 loc) · 5.13 KB
/
prepare.py
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
import glob
import logging
import os
import os.path
from typing import Dict
os.environ['FOR_DISABLE_CONSOLE_CTRL_HANDLER'] = '1'
import _gin_bugfix
import gin
import pandas as pd
import numpy as np
from absl import flags
from absl import app
from tqdm import tqdm
from pytorch_lightning import seed_everything
from ariadne.preprocessing import DataProcessor
from ariadne.parsing import parse_df
FLAGS = flags.FLAGS
flags.DEFINE_string(
name='config', default=None,
help='Path to the config file to use.'
)
flags.DEFINE_enum(
name='log', default='INFO',
enum_values=['INFO', 'DEBUG', 'WARNING', 'ERROR', 'CRITICAL'],
help='Level of logging'
)
def setup_logger(logger_dir, preprocessor_name):
# create logger
logger = logging.getLogger()
logger.setLevel(logging.DEBUG)
# create console handler and set level to debug
logger_dir = os.path.join(logger_dir, "logs_"+preprocessor_name)
os.makedirs(logger_dir, exist_ok=True)
fh = logging.FileHandler('%s/prepare_%s.log' % (logger_dir, preprocessor_name))
fh.setLevel(logging.DEBUG)
# create formatter
formatter = logging.Formatter('%(asctime)s [%(levelname)s] %(name)s: %(message)s')
# add formatter to ch
fh.setFormatter(formatter)
# add ch to logger
logger.addHandler(fh)
LOGGER = logging.getLogger('ariadne.prepare')
def parse_single_arr_arg(arr_arg):
if '..' in arr_arg:
args = arr_arg.split('..')
assert len(args) == 2, "It should have form '%num%..%num%' ."
return np.arange(int(args[0]), int(args[1])), False
if ':' in arr_arg:
return -1, True
return [int(arr_arg)], False
# todo: made universal parse
@gin.configurable
def parse(input_file_mask,
csv_params: Dict[str, object],
events_quantity,
filter_func=None,
input_file_list=None
):
assert not (input_file_mask and input_file_list), 'specify only input_file_mask or input_file_list'
if input_file_list:
files_list = []
for mask in input_file_list:
files_list += glob.glob(mask)
elif input_file_mask:
files_list = glob.glob(input_file_mask)
assert len(files_list) > 0, f"no files found matching mask {input_file_mask}"
assert isinstance(events_quantity, str), 'events_quantity should be a str. see comments in config to set it ' \
'correctly. Got: %r with type %r ' % (
events_quantity, type(events_quantity))
event_idxs, parse_all = parse_single_arr_arg(events_quantity)
LOGGER.info(f"[Parse]: matched {len(files_list)} files:")
for idx, elem in enumerate(files_list):
LOGGER.info("[Parse]: started parsing CSV #%d (%s):" % (idx, elem))
parsed_df = parse_df(elem, **csv_params)
if filter_func:
parsed_df = filter_func(parsed_df)
LOGGER.info("[Parse]: finished parsing CSV...")
if not parse_all:
res = np.array(event_idxs)
yield parsed_df[parsed_df.event.isin(res)].copy(), os.path.basename(elem)
return
else:
yield parsed_df, os.path.basename(elem)
return
# endof TODO
@gin.configurable
def preprocess(
target_processor: DataProcessor.__class__,
output_dir: str,
ignore_asserts: bool,
random_seed=None,
):
os.makedirs(output_dir, exist_ok=True)
setup_logger(output_dir, target_processor.__name__)
LOGGER.info("GOT config: \n======config======\n %s \n========config=======" % gin.config_str())
if random_seed is not None:
LOGGER.info('Setting random seed to %d', random_seed)
seed_everything(random_seed)
for data_df, basename in parse():
LOGGER.info("[Preprocess]: started processing a df with %d rows:" % len(data_df))
processor: DataProcessor = target_processor(data_df=data_df,
output_dir=output_dir)
generator = processor.generate_chunks_iterable()
preprocessed_chunks = []
try:
for (idx, df_chunk) in tqdm(generator):
try:
data_chunk = processor.construct_chunk(df_chunk)
except AssertionError as ex:
if ignore_asserts:
LOGGER.warning("GOT ASSERT %r on idx %d" % (ex, idx))
continue
else:
raise ex
preprocessed_chunks.append(
processor.preprocess_chunk(chunk=data_chunk, idx=basename)
)
except KeyboardInterrupt as ex:
LOGGER.warning("BREAKING by interrupt. got %d processed chunks" % len(preprocessed_chunks))
processed_data = processor.postprocess_chunks(preprocessed_chunks)
processor.save_on_disk(processed_data)
def main(argv):
del argv
if FLAGS.config is None:
raise SystemError("Expected valid path to the GIN-config file supplied as '--config %PATH%' parameter")
gin.parse_config(open(FLAGS.config))
LOGGER.setLevel(FLAGS.log)
preprocess()
LOGGER.info("end processing")
if __name__ == '__main__':
app.run(main)