Mercurial > repos > rmarenco > hubarchivecreator

comparison util/subtools.py @ 1:fb5e60d4d18a draft

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 64cfc08088d11f6818c1b4e5514ef9e67969eaff-dirty
author rmarenco
date Wed, 13 Jul 2016 13:36:37 -0400
parents
children acc233161f50
comparison
equal deleted inserted replaced
0:0f3bc17e5ede 1:fb5e60d4d18a
1 #!/usr/bin/python
2 # -*- coding: utf8 -*-
3
4 """
5 This class handles the subprocess calls of the different tools used
6 in HubArchiveCreator
7 """
8
9 import os
10 import subprocess
11
12
13 def _handleExceptionAndCheckCall(array_call, **kwargs):
14 """
15 This class handle exceptions and call the tool.
16 It maps the signature of subprocess.check_call:
17 See https://docs.python.org/2/library/subprocess.html#subprocess.check_call
18 """
19 stdin = kwargs.get('stdin')
20 stdout = kwargs.get('stdout')
21 stderr = kwargs.get('stderr')
22 shell = kwargs.get('shell')
23 try:
24 p = subprocess.check_call(array_call, stdin=stdin, stdout=stdout, stderr=stderr, shell=shell)
25 except subprocess.CalledProcessError:
26 raise
27 return p
28
29
30 def twoBitInfo(two_bit_file_name, two_bit_info_file):
31 """
32 Call twoBitInfo and write the result into twoBit_info_file
33 :param two_bit_file_name:
34 :param two_bit_info_file:
35 :return the subprocess.check_call return object:
36 """
37 array_call = ['twoBitInfo', two_bit_file_name, two_bit_info_file]
38 p = _handleExceptionAndCheckCall(array_call)
39 return p
40
41
42 def faToTwoBit(fasta_file_name, mySpecieFolder):
43 """
44 This function call faToTwoBit UCSC tool, and return the twoBitFile
45 :param fasta_file_name:
46 :param mySpecieFolder:
47 :return:
48 """
49 baseNameFasta = os.path.basename(fasta_file_name)
50 suffixTwoBit, extensionTwoBit = os.path.splitext(baseNameFasta)
51 nameTwoBit = suffixTwoBit + '.2bit'
52
53 with open(os.path.join(mySpecieFolder, nameTwoBit), 'w') as twoBitFile:
54 array_call = ['faToTwoBit', fasta_file_name, twoBitFile.name]
55 _handleExceptionAndCheckCall(array_call)
56
57 return twoBitFile
58
59
60 def gtfToGenePred(input_gtf_file_name, gene_pred_file_name):
61 """
62 Call gtfToGenePred and write the result into gene_pred_file_name
63 :param input_gtf_file_name:
64 :param gene_pred_file_name:
65 :return:
66 """
67 array_call = ['gtfToGenePred', input_gtf_file_name, gene_pred_file_name]
68 p = _handleExceptionAndCheckCall(array_call)
69 return p
70
71
72 def gff3ToGenePred(input_gff3_file_name, gene_pred_file_name):
73 """
74 Call gff3ToGenePred and write the result into gene_pred_file_name
75 :param input_gff3_file_name:
76 :param gene_pred_file_name:
77 :return:
78 """
79 array_call = ['gff3ToGenePred', input_gff3_file_name, gene_pred_file_name]
80 p = _handleExceptionAndCheckCall(array_call)
81 return p
82
83
84 def genePredToBed(gene_pred_file_name, unsorted_bed_file_name):
85 """
86 Call genePredToBed and write the result into unsorted_bed_file_name
87 :param gene_pred_file_name:
88 :param unsorted_bed_file_name:
89 :return:
90 """
91 array_call = ['genePredToBed', gene_pred_file_name, unsorted_bed_file_name]
92 p = _handleExceptionAndCheckCall(array_call)
93 return p
94
95
96 def sort(unsorted_bed_file_name, sorted_bed_file_name):
97 """
98 Call sort with -k1,1 -k2,2n on unsorted_bed_file_name and write the result into sorted_bed_file_name
99 :param unsorted_bed_file_name:
100 :param sorted_bed_file_name:
101 :return:
102 """
103 array_call = ['sort', '-k', '1,1', '-k', '2,2n', unsorted_bed_file_name, '-o', sorted_bed_file_name]
104 p = _handleExceptionAndCheckCall(array_call)
105 return p
106
107
108 def sortChromSizes(two_bit_info_file_name, chrom_sizes_file_name):
109 """
110 Call sort with -k2rn on two_bit_info_file_name and write the result into chrom_sizes_file_name
111 :param two_bit_info_file_name:
112 :param chrom_sizes_file_name:
113 :return:
114 """
115 array_call = ['sort', '-k2rn', two_bit_info_file_name, '-o', chrom_sizes_file_name]
116 p = _handleExceptionAndCheckCall(array_call)
117 return p
118
119
120 def bedToBigBed(sorted_bed_file_name, chrom_sizes_file_name, big_bed_file_name, typeOption=None, autoSql=None):
121 """
122 Call bedToBigBed on sorted_bed_file_name, using chrom_sizes_file_name and write the result into big_bed_file_name
123 :param sorted_bed_file_name:
124 :param chrom_sizes_file_name:
125 :param big_bed_file_name:
126 :return:
127 """
128 array_call = ['bedToBigBed', sorted_bed_file_name, chrom_sizes_file_name, big_bed_file_name]
129 if typeOption:
130 array_call.append(typeOption)
131 if autoSql:
132 array_call.append(autoSql)
133
134 p = _handleExceptionAndCheckCall(array_call)
135 return p
136
137
138 def sortBam(input_bam_file_name, output_sorted_bam_name):
139 """
140 Call samtools on input_bam_file_name and output the result in output_sorted_bam_name
141 :param input_bam_file_name:
142 :param output_sorted_bam_name:
143 :return:
144 """
145 array_call = ['samtools', 'sort', input_bam_file_name, '-o', output_sorted_bam_name]
146 p = _handleExceptionAndCheckCall(array_call)
147 return p
148
149
150 def createBamIndex(input_sorted_bam_file_name, output_name_index_name):
151 """
152 Call `samtools index` on imput_sorted_bam_file_name and output the result in output_name_index_name
153 :param input_sorted_bam_file_name:
154 :param output_name_index_name:
155 :return:
156 """
157 array_call = ['samtools', 'index', input_sorted_bam_file_name, output_name_index_name]
158 p = _handleExceptionAndCheckCall(array_call)
159 return p