Mercurial > repos > stevecassidy > nltktools

annotate g_pos.py @ 3:0df72a8ab095 draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit f2432aaedd36ae7662873623d8861d0982dffdd2
author stevecassidy
date Mon, 20 Nov 2017 22:52:11 -0500
parents a47980ef2b96
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
3
0df72a8ab095 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit f2432aaedd36ae7662873623d8861d0982dffdd2
stevecassidy
parents: 2
diff changeset
1 from __future__ import print_function, unicode_literals
0
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
2 import nltk
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
3 import argparse
3
0df72a8ab095 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit f2432aaedd36ae7662873623d8861d0982dffdd2
stevecassidy
parents: 2
diff changeset
4 import io
2
a47980ef2b96 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit b5b26e9118f2ad8af109d606746b39a5588f0511-dirty
stevecassidy
parents: 1
diff changeset
5
a47980ef2b96 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit b5b26e9118f2ad8af109d606746b39a5588f0511-dirty
stevecassidy
parents: 1
diff changeset
6 nltk.download('averaged_perceptron_tagger', quiet=True)
3
0df72a8ab095 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit f2432aaedd36ae7662873623d8861d0982dffdd2
stevecassidy
parents: 2
diff changeset
7 nltk.download('punkt', quiet=True)
2
a47980ef2b96 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit b5b26e9118f2ad8af109d606746b39a5588f0511-dirty
stevecassidy
parents: 1
diff changeset
8
0
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
9
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
10 def arguments():
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
11 parser = argparse.ArgumentParser(description="tokenize a text")
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
12 parser.add_argument('--input', required=True, action="store", type=str, help="input text file")
2
a47980ef2b96 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit b5b26e9118f2ad8af109d606746b39a5588f0511-dirty
stevecassidy
parents: 1
diff changeset
13 parser.add_argument('--output', required=True, action="store", type=str, help="output file path")
a47980ef2b96 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit b5b26e9118f2ad8af109d606746b39a5588f0511-dirty
stevecassidy
parents: 1
diff changeset
14 return parser.parse_args()
1
fb617586f4b2 planemo upload commit a81826fe44f09a3710a35c183aa88b745aeec064-dirty
stevecassidy
parents: 0
diff changeset
15
0
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
16
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
17 def postag(in_file, out_file):
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
18 """Input: a text file with one token per line
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
19 Output: a version of the text with Part of Speech tags written as word/TAG
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
20 """
3
0df72a8ab095 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit f2432aaedd36ae7662873623d8861d0982dffdd2
stevecassidy
parents: 2
diff changeset
21 with open(in_file, 'rb') as fd:
1
fb617586f4b2 planemo upload commit a81826fe44f09a3710a35c183aa88b745aeec064-dirty
stevecassidy
parents: 0
diff changeset
22 text = fd.read()
3
0df72a8ab095 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit f2432aaedd36ae7662873623d8861d0982dffdd2
stevecassidy
parents: 2
diff changeset
23 text = text.decode('utf-8')
1
fb617586f4b2 planemo upload commit a81826fe44f09a3710a35c183aa88b745aeec064-dirty
stevecassidy
parents: 0
diff changeset
24
0
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
25 sentences = nltk.sent_tokenize(text)
2
a47980ef2b96 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit b5b26e9118f2ad8af109d606746b39a5588f0511-dirty
stevecassidy
parents: 1
diff changeset
26
3
0df72a8ab095 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit f2432aaedd36ae7662873623d8861d0982dffdd2
stevecassidy
parents: 2
diff changeset
27 with io.open(out_file, 'w') as output:
1
fb617586f4b2 planemo upload commit a81826fe44f09a3710a35c183aa88b745aeec064-dirty
stevecassidy
parents: 0
diff changeset
28 for sentence in sentences:
fb617586f4b2 planemo upload commit a81826fe44f09a3710a35c183aa88b745aeec064-dirty
stevecassidy
parents: 0
diff changeset
29 tokens = nltk.word_tokenize(sentence)
fb617586f4b2 planemo upload commit a81826fe44f09a3710a35c183aa88b745aeec064-dirty
stevecassidy
parents: 0
diff changeset
30 postags = nltk.pos_tag(tokens)
fb617586f4b2 planemo upload commit a81826fe44f09a3710a35c183aa88b745aeec064-dirty
stevecassidy
parents: 0
diff changeset
31 for postag in postags:
fb617586f4b2 planemo upload commit a81826fe44f09a3710a35c183aa88b745aeec064-dirty
stevecassidy
parents: 0
diff changeset
32 # print postag
3
0df72a8ab095 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit f2432aaedd36ae7662873623d8861d0982dffdd2
stevecassidy
parents: 2
diff changeset
33 p = "%s/%s " % postag
0df72a8ab095 planemo upload for repository https://github.com/Alveo/alveo-galaxy-tools commit f2432aaedd36ae7662873623d8861d0982dffdd2
stevecassidy
parents: 2
diff changeset
34 output.write(p)
1
fb617586f4b2 planemo upload commit a81826fe44f09a3710a35c183aa88b745aeec064-dirty
stevecassidy
parents: 0
diff changeset
35 output.write('\n')
0
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
36
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
37
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
38 if __name__ == '__main__':
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
39 args = arguments()
e991d4e60c17 planemo upload commit 0203cb3a0b40d9348674b2b098af805e2986abca-dirty
stevecassidy
parents:
diff changeset
40 postag(args.input, args.output)