Mercurial > repos > rmarenco > hubarchivecreator

annotate util/santitizer.py @ 32:6d1f2bb5c1d1 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/goeckslab/hub-archive-creator commit 5a0c39ed61b4408ba65a759aede807df260d097d
author yating-l
date Sun, 26 Aug 2018 16:44:27 -0400
parents 7e8a8b732db3
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
29
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
1 #!/usr/bin/python
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
2 # -*- coding: utf8 -*-
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
3
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
4 """
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
5 This class handles the subprocess calls of the different tools used
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
6 in HubArchiveCreator
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
7 """
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
8
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
9 import logging
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
10 import os
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
11 import subprocess
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
12 import sys
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
13 import string
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
14 import tempfile
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
15
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
16
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
17 def prefixTrackName(filename):
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
18 """
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
19 santitize trackName. Because track name must begin with a letter and
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
20 contain only the following chars: [a-zA-Z0-9_].
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
21 See the "track" Common settings at:
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
22 https://genome.ucsc.edu/goldenpath/help/trackDb/trackDbHub.html#bigPsl_-_Pairwise_Alignments
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
23 skip the santitization for cytoBandIdeo track
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
24 """
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
25 if filename == 'cytoBandIdeo':
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
26 return filename
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
27 valid_chars = "_%s%s" % (string.ascii_letters, string.digits)
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
28 sanitize_name = ''.join([c if c in valid_chars else '_' for c in filename])
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
29 sanitize_name = "gonramp_" + sanitize_name
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
30 return sanitize_name
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
31
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
32 def sanitize_name_input(string_to_sanitize):
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
33 """
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
34 Sanitize the string passed in parameter by replacing '/' and ' ' by '_'
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
35
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
36 :param string_to_sanitize:
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
37 :return :
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
38
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
39 :Example:
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
40
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
41 >>> sanitize_name_input('this/is an//example')
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
42 this_is_an__example
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
43 """
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
44 return string_to_sanitize \
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
45 .replace("/", "_") \
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
46 .replace(" ", "_")
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
47
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
48 def sanitize_name_inputs(inputs_data):
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
49 """
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
50 Sanitize value of the keys "name" of the dictionary passed in parameter.
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
51
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
52 Because sometimes output from Galaxy, or even just file name, from user inputs, have spaces.
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
53 Also, it can contain '/' character and could break the use of os.path function.
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
54
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
55 :param inputs_data: dict[string, dict[string, string]]
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
56 """
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
57 for key in inputs_data:
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
58 inputs_data[key]["name"] = sanitize_name_input(inputs_data[key]["name"])
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
59
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
60 def sanitize_group_name(group_name):
7e8a8b732db3 planemo upload for repository https://github.com/goeckslab/hub-archive-creator commit 1a81ebd0ddea950b84af3fc830e9267a4814b29f
yating-l
parents:
diff changeset
61 return group_name.lower().replace(' ', '_')