Mercurial > repos > rmarenco > hubarchivecreator

comparison Gtf.py @ 10:acc233161f50 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 1b1063f90004764bcf504f4340738eca5c4b1f9d
author rmarenco
date Thu, 21 Jul 2016 05:58:51 -0400
parents fb5e60d4d18a
children d05236b15f81
comparison
equal deleted inserted replaced
9:4f9847539a28 10:acc233161f50
9 from TrackDb import TrackDb 9 from TrackDb import TrackDb
10 from util import subtools 10 from util import subtools
11 11
12 12
13 class Gtf( Datatype ): 13 class Gtf( Datatype ):
14 def __init__( self, input_gtf_false_path, data_gtf, 14 def __init__( self, input_gtf_false_path, data_gtf):
15 input_fasta_file, extra_files_path, tool_directory ): 15
16 super(Gtf, self).__init__( input_fasta_file=input_fasta_file, 16 super(Gtf, self).__init__()
17 extra_files_path=extra_files_path,
18 tool_directory=tool_directory )
19 17
20 self.track = None 18 self.track = None
21 19
22 self.input_gtf_false_path = input_gtf_false_path 20 self.input_gtf_false_path = input_gtf_false_path
23 self.name_gtf = data_gtf["name"] 21 self.name_gtf = data_gtf["name"]
28 # TODO: See if we need these temporary files as part of the generated files 26 # TODO: See if we need these temporary files as part of the generated files
29 genePredFile = tempfile.NamedTemporaryFile(bufsize=0, suffix=".genePred") 27 genePredFile = tempfile.NamedTemporaryFile(bufsize=0, suffix=".genePred")
30 unsortedBedFile = tempfile.NamedTemporaryFile(bufsize=0, suffix=".unsortedBed") 28 unsortedBedFile = tempfile.NamedTemporaryFile(bufsize=0, suffix=".unsortedBed")
31 sortedBedFile = tempfile.NamedTemporaryFile(suffix=".sortedBed") 29 sortedBedFile = tempfile.NamedTemporaryFile(suffix=".sortedBed")
32 30
33 twoBitInfoFile = tempfile.NamedTemporaryFile(bufsize=0)
34 chromSizesFile = tempfile.NamedTemporaryFile(bufsize=0, suffix=".chrom.sizes")
35
36 # GtfToGenePred 31 # GtfToGenePred
37 subtools.gtfToGenePred(self.input_gtf_false_path, genePredFile.name) 32 subtools.gtfToGenePred(self.input_gtf_false_path, genePredFile.name)
38 33
39 # TODO: From there, refactor because common use with Gff3.py 34 # TODO: From there, refactor because common use with Gff3.py
40 # genePredToBed processing 35 # genePredToBed processing
41 subtools.genePredToBed(genePredFile.name, unsortedBedFile.name) 36 subtools.genePredToBed(genePredFile.name, unsortedBedFile.name)
42 37
43 # Sort processing 38 # Sort processing
44 subtools.sort(unsortedBedFile.name, sortedBedFile.name) 39 subtools.sort(unsortedBedFile.name, sortedBedFile.name)
45 40
46 # TODO: Chehck if the twoBitInfo / ChromSizes is redundant and make an intermediate class
47 # Generate the twoBitInfo
48 subtools.twoBitInfo(self.twoBitFile.name, twoBitInfoFile.name)
49
50 # Then we get the output to generate the chromSizes
51 # TODO: Check if no errors
52 subtools.sortChromSizes(twoBitInfoFile.name, chromSizesFile.name)
53
54 # bedToBigBed processing 41 # bedToBigBed processing
55 # TODO: Change the name of the bb, to tool + genome + possible adding if multiple + .bb 42 # TODO: Change the name of the bb, to tool + genome + possible adding if multiple + .bb
56 trackName = "".join( ( self.name_gtf, ".bb") ) 43 trackName = "".join( ( self.name_gtf, ".bb") )
57 myBigBedFilePath = os.path.join(self.myTrackFolderPath, trackName) 44 myBigBedFilePath = os.path.join(self.myTrackFolderPath, trackName)
58 with open(myBigBedFilePath, 'w') as bigBedFile: 45 with open(myBigBedFilePath, 'w') as bigBedFile:
59 subtools.bedToBigBed(sortedBedFile.name, chromSizesFile.name, bigBedFile.name) 46 subtools.bedToBigBed(sortedBedFile.name, self.chromSizesFile.name, bigBedFile.name)
60 47
61 # Create the Track Object 48 # Create the Track Object
62 dataURL = "tracks/%s" % trackName 49 dataURL = "tracks/%s" % trackName
63 50
64 trackDb = TrackDb( 51 trackDb = TrackDb(