Mercurial > repos > sblanck > mpagenomics

comparison preprocess.R @ 0:4d539083cf7f draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/sblanck/MPAgenomics4Galaxy/tree/master/mpagenomics_wrappers commit 689d0d8dc899a683ee18700ef385753559850233-dirty
author sblanck
date Tue, 12 May 2020 10:40:36 -0400
parents
children 3fcbb8030fcc
comparison
equal deleted inserted replaced
-1:000000000000 0:4d539083cf7f
1 #!/usr/bin/env Rscript
2 # setup R error handling to go to stderr
3 options( show.error.messages=F, error = function () { cat( geterrmessage(), file=stderr() ); q( "no", 1, F ) } )
4
5 # we need that to not crash galaxy with an UTF8 error on German LC settings.
6 loc <- Sys.setlocale("LC_MESSAGES", "en_US.UTF-8")
7
8 library("optparse")
9
10 ##### Read options
11 option_list=list(
12 make_option("--summary",type="character",default=NULL, dest="summary"),
13 make_option("--dataSetName",type="character",default=NULL, dest="dataSetName"),
14 make_option("--new_file_path",type="character",default=NULL, dest="new_file_path"),
15 make_option("--inputcdffull_name",type="character",default=NULL, dest="inputcdffull_name"),
16 make_option("--inputufl_name",type="character",default=NULL, dest="inputufl_name"),
17 make_option("--inputugp_name",type="character",default=NULL, dest="inputugp_name"),
18 make_option("--inputacs_name",type="character",default=NULL, dest="inputacs_name"),
19 make_option("--inputcdffull",type="character",default=NULL, dest="inputcdffull"),
20 make_option("--inputufl",type="character",default=NULL, dest="inputufl"),
21 make_option("--inputugp",type="character",default=NULL, dest="inputugp"),
22 make_option("--inputacs",type="character",default=NULL, dest="inputacs"),
23 make_option("--tumorcsv",type="character",default=NULL, dest="tumorcsv"),
24 make_option("--settingsType",type="character",default=NULL, dest="settingsType"),
25 make_option("--outputgraph",type="character",default=NULL, dest="outputgraph"),
26 make_option("--zipfigures",type="character",default=NULL, dest="zipfigures"),
27 make_option("--outputlog",type="character",default=NULL, dest="outputlog"),
28 make_option("--log",type="character",default=NULL, dest="log"),
29 make_option("--user_id",type="character",default=NULL, dest="user_id"),
30 make_option("--input",type="character",default=NULL, dest="input")
31 );
32
33 opt_parser = OptionParser(option_list=option_list);
34 opt = parse_args(opt_parser);
35
36 if(is.null(opt$input)){
37 print_help(opt_parser)
38 stop("input required.", call.=FALSE)
39 }
40
41 #loading libraries
42
43 summary=opt$summary
44 dataSetName=opt$dataSetName
45 newFilePath=opt$new_file_path
46 inputCDFName=opt$inputcdffull_name
47 inputUFLName=opt$inputufl_name
48 inputUGPName=opt$inputugp_name
49 inputACSName=opt$inputacs_name
50 inputCDF=opt$inputcdffull
51 inputUFL=opt$inputufl
52 inputUGP=opt$inputugp
53 inputACS=opt$inputacs
54 tumorcsv=opt$tumorcsv
55 settingsType=opt$settingsType
56 outputGraph=opt$outputgraph
57 zipfigures=opt$zipfigures
58 outputlog=opt$outputlog
59 log=opt$log
60 userId=opt$user_id
61
62 destinationPath=file.path(newFilePath, userId, dataSetName)
63 mpagenomicsDir = file.path(newFilePath,"mpagenomics",userId)
64 dataDir = file.path(newFilePath, userId)
65 chipDir = file.path(newFilePath,"mpagenomics",userId,"annotationData","chipTypes")
66 createArchitecture=TRUE
67
68 if (dir.exists(chipDir))
69 system(paste0("rm -r ", chipDir))
70
71 if (!dir.exists(mpagenomicsDir))
72 dir.create(mpagenomicsDir, showWarnings = TRUE, recursive = TRUE)
73
74 if (!dir.exists(dataDir))
75 dir.create(dataDir, showWarnings = TRUE, recursive = TRUE)
76
77 listInput <- trimws( unlist( strsplit(trimws(opt$input), ",") ) )
78 if(length(listInput)<2){
79 stop("To few .CEL files selected : At least 2 .CEL files are required", call.=FALSE)
80 }
81
82
83 celList=vector()
84 celFileNameList=vector()
85
86 for (i in 1:length(listInput))
87 {
88 inputFileInfo <- unlist( strsplit( listInput[i], ';' ) )
89 celList=c(celList,inputFileInfo[1])
90 celFileNameList=c(celFileNameList,inputFileInfo[2])
91 }
92
93
94 for (i in 1:length(celFileNameList))
95 {
96 source = celList[i]
97 destination=file.path(dataDir,celFileNameList[i])
98 file.copy(source, destination)
99 }
100 split=unlist(strsplit(inputCDFName,",",fixed=TRUE))
101 tag=NULL
102 if (length(split) != 0) {
103 chipType=split[1]
104 tagExt=split[2]
105 tag=unlist(strsplit(tagExt,".",fixed=TRUE))[1]
106 } else {
107 chipType=split[1]
108 }
109
110 if(!file.exists(file.path(dataDir,inputCDFName)))
111 file.symlink(inputCDF,file.path(dataDir,inputCDFName))
112 if(!file.exists(file.path(dataDir,inputACSName)))
113 file.symlink(inputACS,file.path(dataDir,inputACSName))
114 if(!file.exists(file.path(dataDir,inputUFLName)))
115 file.symlink(inputUFL,file.path(dataDir,inputUFLName))
116 if(!file.exists(file.path(dataDir,inputUGPName)))
117 file.symlink(inputUGP,file.path(dataDir,inputUGPName))
118
119 fig_dir = file.path("mpagenomics", userId, "figures", dataSetName, "signal")
120 abs_fig_dir = file.path(newFilePath, fig_dir)
121
122 chip=chipType
123 dataset=dataSetName
124 workdir=mpagenomicsDir
125 celPath=dataDir
126 chipPath=dataDir
127 tumor=tumorcsv
128 outputgraph=type.convert(outputGraph)
129
130
131 library(MPAgenomics)
132 setwd(workdir)
133
134 if (outputlog){
135 sinklog <- file(log, open = "wt")
136 sink(sinklog ,type = "output")
137 sink(sinklog, type = "message")
138 }
139
140 if (settingsType=="standard")
141 {
142 signalPreProcess(dataSetName=dataset, chipType=chip, dataSetPath=celPath,chipFilesPath=chipPath, path=workdir,createArchitecture=createArchitecture, savePlot=outputgraph, tags=tag)
143 } else {
144 signalPreProcess(dataSetName=dataset, chipType=chip, dataSetPath=celPath,chipFilesPath=chipPath, normalTumorArray=tumor, path=workdir,createArchitecture=createArchitecture, savePlot=outputgraph, tags=tag)
145 }
146 setwd(abs_fig_dir)
147 files2zip <- dir(abs_fig_dir)
148 zip(zipfile = "figures.zip", files = files2zip)
149 file.rename("figures.zip",zipfigures)
150 summarydf=data.frame(celFileNameList,rep(dataSetName,length(celFileNameList)),rep(chipType,length(celFileNameList)))
151 write.table(summarydf,file=summary,quote=FALSE,row.names=FALSE,col.names=FALSE,sep="\t")
152
153 if (outputlog){
154 sink(type="output")
155 sink(type="message")
156 close(sinklog)
157 }
158