-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathwmt16_en_de.sh
182 lines (151 loc) · 7.06 KB
/
wmt16_en_de.sh
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
#! /usr/bin/env bash
# Copyright 2017 Google Inc.
#
# 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.
set -e
BASE_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )/../.." && pwd )"
OUTPUT_DIR=${OUTPUT_DIR:-$HOME/nmt_data/wmt16_de_en}
echo "Writing to ${OUTPUT_DIR}. To change this, set the OUTPUT_DIR environment variable."
OUTPUT_DIR_DATA="${OUTPUT_DIR}/data"
mkdir -p $OUTPUT_DIR_DATA
echo "Downloading Europarl v7. This may take a while..."
wget -nc -nv -O ${OUTPUT_DIR_DATA}/europarl-v7-de-en.tgz \
http://www.statmt.org/europarl/v7/de-en.tgz
echo "Downloading Common Crawl corpus. This may take a while..."
wget -nc -nv -O ${OUTPUT_DIR_DATA}/common-crawl.tgz \
http://www.statmt.org/wmt13/training-parallel-commoncrawl.tgz
echo "Downloading News Commentary v11. This may take a while..."
wget -nc -nv -O ${OUTPUT_DIR_DATA}/nc-v11.tgz \
http://data.statmt.org/wmt16/translation-task/training-parallel-nc-v11.tgz
echo "Downloading dev/test sets"
wget -nc -nv -O ${OUTPUT_DIR_DATA}/dev.tgz \
http://data.statmt.org/wmt16/translation-task/dev.tgz
wget -nc -nv -O ${OUTPUT_DIR_DATA}/test.tgz \
http://data.statmt.org/wmt16/translation-task/test.tgz
# Extract everything
echo "Extracting all files..."
mkdir -p "${OUTPUT_DIR_DATA}/europarl-v7-de-en"
tar -xvzf "${OUTPUT_DIR_DATA}/europarl-v7-de-en.tgz" -C "${OUTPUT_DIR_DATA}/europarl-v7-de-en"
mkdir -p "${OUTPUT_DIR_DATA}/common-crawl"
tar -xvzf "${OUTPUT_DIR_DATA}/common-crawl.tgz" -C "${OUTPUT_DIR_DATA}/common-crawl"
mkdir -p "${OUTPUT_DIR_DATA}/nc-v11"
tar -xvzf "${OUTPUT_DIR_DATA}/nc-v11.tgz" -C "${OUTPUT_DIR_DATA}/nc-v11"
mkdir -p "${OUTPUT_DIR_DATA}/dev"
tar -xvzf "${OUTPUT_DIR_DATA}/dev.tgz" -C "${OUTPUT_DIR_DATA}/dev"
mkdir -p "${OUTPUT_DIR_DATA}/test"
tar -xvzf "${OUTPUT_DIR_DATA}/test.tgz" -C "${OUTPUT_DIR_DATA}/test"
# Concatenate Training data
cat "${OUTPUT_DIR_DATA}/europarl-v7-de-en/europarl-v7.de-en.en" \
"${OUTPUT_DIR_DATA}/common-crawl/commoncrawl.de-en.en" \
"${OUTPUT_DIR_DATA}/nc-v11/training-parallel-nc-v11/news-commentary-v11.de-en.en" \
> "${OUTPUT_DIR}/train.en"
wc -l "${OUTPUT_DIR}/train.en"
cat "${OUTPUT_DIR_DATA}/europarl-v7-de-en/europarl-v7.de-en.de" \
"${OUTPUT_DIR_DATA}/common-crawl/commoncrawl.de-en.de" \
"${OUTPUT_DIR_DATA}/nc-v11/training-parallel-nc-v11/news-commentary-v11.de-en.de" \
> "${OUTPUT_DIR}/train.de"
wc -l "${OUTPUT_DIR}/train.de"
# Clone Moses
if [ ! -d "${OUTPUT_DIR}/mosesdecoder" ]; then
echo "Cloning moses for data processing"
git clone https://github.com/moses-smt/mosesdecoder.git "${OUTPUT_DIR}/mosesdecoder"
fi
# Convert SGM files
# Convert newstest2014 data into raw text format
${OUTPUT_DIR}/mosesdecoder/scripts/ems/support/input-from-sgm.perl \
< ${OUTPUT_DIR_DATA}/dev/dev/newstest2014-deen-src.de.sgm \
> ${OUTPUT_DIR_DATA}/dev/dev/newstest2014.de
${OUTPUT_DIR}/mosesdecoder/scripts/ems/support/input-from-sgm.perl \
< ${OUTPUT_DIR_DATA}/dev/dev/newstest2014-deen-ref.en.sgm \
> ${OUTPUT_DIR_DATA}/dev/dev/newstest2014.en
# Convert newstest2015 data into raw text format
${OUTPUT_DIR}/mosesdecoder/scripts/ems/support/input-from-sgm.perl \
< ${OUTPUT_DIR_DATA}/dev/dev/newstest2015-deen-src.de.sgm \
> ${OUTPUT_DIR_DATA}/dev/dev/newstest2015.de
${OUTPUT_DIR}/mosesdecoder/scripts/ems/support/input-from-sgm.perl \
< ${OUTPUT_DIR_DATA}/dev/dev/newstest2015-deen-ref.en.sgm \
> ${OUTPUT_DIR_DATA}/dev/dev/newstest2015.en
# Convert newstest2016 data into raw text format
${OUTPUT_DIR}/mosesdecoder/scripts/ems/support/input-from-sgm.perl \
< ${OUTPUT_DIR_DATA}/test/test/newstest2016-deen-src.de.sgm \
> ${OUTPUT_DIR_DATA}/test/test/newstest2016.de
${OUTPUT_DIR}/mosesdecoder/scripts/ems/support/input-from-sgm.perl \
< ${OUTPUT_DIR_DATA}/test/test/newstest2016-deen-ref.en.sgm \
> ${OUTPUT_DIR_DATA}/test/test/newstest2016.en
# Copy dev/test data to output dir
cp ${OUTPUT_DIR_DATA}/dev/dev/newstest20*.de ${OUTPUT_DIR}
cp ${OUTPUT_DIR_DATA}/dev/dev/newstest20*.en ${OUTPUT_DIR}
cp ${OUTPUT_DIR_DATA}/test/test/newstest20*.de ${OUTPUT_DIR}
cp ${OUTPUT_DIR_DATA}/test/test/newstest20*.en ${OUTPUT_DIR}
# Tokenize data
for f in ${OUTPUT_DIR}/*.de; do
echo "Tokenizing $f..."
${OUTPUT_DIR}/mosesdecoder/scripts/tokenizer/tokenizer.perl -q -l de -threads 8 < $f > ${f%.*}.tok.de
done
for f in ${OUTPUT_DIR}/*.en; do
echo "Tokenizing $f..."
${OUTPUT_DIR}/mosesdecoder/scripts/tokenizer/tokenizer.perl -q -l en -threads 8 < $f > ${f%.*}.tok.en
done
# Clean all corpora
for f in ${OUTPUT_DIR}/*.en; do
fbase=${f%.*}
echo "Cleaning ${fbase}..."
${OUTPUT_DIR}/mosesdecoder/scripts/training/clean-corpus-n.perl $fbase de en "${fbase}.clean" 1 80
done
# Create character vocabulary (on tokenized data)
${BASE_DIR}/bin/tools/generate_vocab.py --delimiter "" \
< ${OUTPUT_DIR}/train.tok.clean.en \
> ${OUTPUT_DIR}/vocab.tok.char.en
${BASE_DIR}/bin/tools/generate_vocab.py --delimiter "" \
< ${OUTPUT_DIR}/train.tok.clean.de \
> ${OUTPUT_DIR}/vocab.tok.char.de
# Create character vocabulary (on non-tokenized data)
${BASE_DIR}/bin/tools/generate_vocab.py --delimiter "" \
< ${OUTPUT_DIR}/train.clean.en \
> ${OUTPUT_DIR}/vocab.char.en
${BASE_DIR}/bin/tools/generate_vocab.py --delimiter "" \
< ${OUTPUT_DIR}/train.clean.de \
> ${OUTPUT_DIR}/vocab.char.de
# Create vocabulary for EN data
$BASE_DIR/bin/tools/generate_vocab.py \
--max_vocab_size 50000 \
< ${OUTPUT_DIR}/train.tok.clean.en \
> ${OUTPUT_DIR}/vocab.50k.en \
# Create vocabulary for DE data
$BASE_DIR/bin/tools/generate_vocab.py \
--max_vocab_size 50000 \
< ${OUTPUT_DIR}/train.tok.clean.de \
> ${OUTPUT_DIR}/vocab.50k.de \
# Generate Subword Units (BPE)
# Clone Subword NMT
if [ ! -d "${OUTPUT_DIR}/subword-nmt" ]; then
git clone https://github.com/rsennrich/subword-nmt.git "${OUTPUT_DIR}/subword-nmt"
fi
# Learn Shared BPE
for merge_ops in 32000; do
echo "Learning BPE with merge_ops=${merge_ops}. This may take a while..."
cat "${OUTPUT_DIR}/train.tok.clean.de" "${OUTPUT_DIR}/train.tok.clean.en" | \
${OUTPUT_DIR}/subword-nmt/learn_bpe.py -s $merge_ops > "${OUTPUT_DIR}/bpe.${merge_ops}"
echo "Apply BPE with merge_ops=${merge_ops} to tokenized files..."
for lang in en de; do
for f in ${OUTPUT_DIR}/*.tok.${lang} ${OUTPUT_DIR}/*.tok.clean.${lang}; do
outfile="${f%.*}.bpe.${merge_ops}.${lang}"
${OUTPUT_DIR}/subword-nmt/apply_bpe.py -c "${OUTPUT_DIR}/bpe.${merge_ops}" < $f > "${outfile}"
echo ${outfile}
done
done
# Create vocabulary file for BPE
cat "${OUTPUT_DIR}/train.tok.clean.bpe.${merge_ops}.en" "${OUTPUT_DIR}/train.tok.clean.bpe.${merge_ops}.de" | \
${OUTPUT_DIR}/subword-nmt/get_vocab.py | cut -f1 -d ' ' > "${OUTPUT_DIR}/vocab.bpe.${merge_ops}"
done
echo "All done."