1#!/usr/bin/env python
2
3# Licensed to the Apache Software Foundation (ASF) under one
4# or more contributor license agreements.  See the NOTICE file
5# distributed with this work for additional information
6# regarding copyright ownership.  The ASF licenses this file
7# to you under the Apache License, Version 2.0 (the
8# "License"); you may not use this file except in compliance
9# with the License.  You may obtain a copy of the License at
10#
11#   http://www.apache.org/licenses/LICENSE-2.0
12#
13# Unless required by applicable law or agreed to in writing,
14# software distributed under the License is distributed on an
15# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
16# KIND, either express or implied.  See the License for the
17# specific language governing permissions and limitations
18# under the License.
19
20"""
21parse mxnet output log into a markdown table
22"""
23import argparse
24import sys
25import re
26
27parser = argparse.ArgumentParser(description='Parse mxnet output log')
28parser.add_argument('logfile', nargs=1, type=str,
29                    help = 'the log file for parsing')
30parser.add_argument('--format', type=str, default='markdown',
31                    choices = ['markdown', 'none'],
32                    help = 'the format of the parsed outout')
33parser.add_argument('--metric-names', type=str, nargs="+", default = ['accuracy'],
34                    help='names of metrics in log which should be parsed')
35args = parser.parse_args()
36
37with open(args.logfile[0]) as f:
38    lines = f.readlines()
39
40res = [re.compile('.*Epoch\[(\d+)\] Train-'+s+'.*=([.\d]+)') for s in args.metric_names]\
41     + [re.compile('.*Epoch\[(\d+)\] Validation-'+s+'.*=([.\d]+)') for s in args.metric_names]\
42     + [re.compile('.*Epoch\[(\d+)\] Time.*=([.\d]+)')]
43
44data = {}
45for l in lines:
46    i = 0
47    for r in res:
48        m = r.match(l)
49        if m is not None:
50            break
51        i += 1
52    if m is None:
53        continue
54
55    assert len(m.groups()) == 2
56    epoch = int(m.groups()[0])
57    val = float(m.groups()[1])
58
59    if epoch not in data:
60        data[epoch] = [0] * len(res) * 2
61
62    data[epoch][i*2] += val
63    data[epoch][i*2+1] += 1
64
65if args.format == 'markdown':
66    print("| epoch | " + " | ".join(['train-'+s for s in args.metric_names]) + " | " + " | ".join(['val-'+s for s in args.metric_names]) + " | time |")
67    print("| --- "*(len(res)+1) + "|")
68    for k, v in data.items():
69        print("| %2d | " % (k+1)\
70              + " | ".join(["%f" % (v[2*j]/v[2*j+1]) for j in range(2*len(args.metric_names))])\
71              + " | %.1f |" % (v[-2]/v[-1]))
72elif args.format == 'none':
73    print("\t".join(['epoch'] + ['train-' + s for s in args.metric_names] + ['val-' + s for s in args.metric_names] + ['time']))
74    for k, v in data.items():
75        print("\t".join(["%2d" % (k+1)] + ["%f" % (v[2*j]/v[2*j+1]) for j in range(2*len(args.metric_names))] + ["%.1f" % (v[-2]/v[-1])]))
76