37
+ − 1 # rgToolFactory.py
+ − 2 # see https://bitbucket.org/fubar/galaxytoolfactory/wiki/Home
+ − 3 #
+ − 4 # copyright ross lazarus (ross stop lazarus at gmail stop com) May 2012
+ − 5 #
+ − 6 # all rights reserved
+ − 7 # Licensed under the LGPL
+ − 8 # suggestions for improvement and bug fixes welcome at https://bitbucket.org/fubar/galaxytoolfactory/wiki/Home
+ − 9 #
+ − 10 # august 2014
+ − 11 # added some additional output datatypes
+ − 12 # john added citations
+ − 13 # march 2014
+ − 14 # had to remove dependencies because cross toolshed dependencies are not possible - can't pre-specify a toolshed url for graphicsmagick and ghostscript
+ − 15 # grrrrr - night before a demo
+ − 16 # added dependencies to a tool_dependencies.xml if html page generated so generated tool is properly portable
+ − 17 #
+ − 18 # added ghostscript and graphicsmagick as dependencies
+ − 19 # fixed a wierd problem where gs was trying to use the new_files_path from universe (database/tmp) as ./database/tmp
+ − 20 # errors ensued
+ − 21 #
+ − 22 # august 2013
+ − 23 # found a problem with GS if $TMP or $TEMP missing - now inject /tmp and warn
+ − 24 #
+ − 25 # july 2013
+ − 26 # added ability to combine images and individual log files into html output
+ − 27 # just make sure there's a log file foo.log and it will be output
+ − 28 # together with all images named like "foo_*.pdf
+ − 29 # otherwise old format for html
+ − 30 #
+ − 31 # January 2013
+ − 32 # problem pointed out by Carlos Borroto
+ − 33 # added escaping for <>$ - thought I did that ages ago...
+ − 34 #
+ − 35 # August 11 2012
+ − 36 # changed to use shell=False and cl as a sequence
+ − 37
+ − 38 # This is a Galaxy tool factory for simple scripts in python, R or whatever ails ye.
+ − 39 # It also serves as the wrapper for the new tool.
+ − 40 #
+ − 41 # you paste and run your script
+ − 42 # Only works for simple scripts that read one input from the history.
+ − 43 # Optionally can write one new history dataset,
+ − 44 # and optionally collect any number of outputs into links on an autogenerated HTML page.
+ − 45
+ − 46 # DO NOT install on a public or important site - please.
+ − 47
+ − 48 # installed generated tools are fine if the script is safe.
+ − 49 # They just run normally and their user cannot do anything unusually insecure
+ − 50 # but please, practice safe toolshed.
+ − 51 # Read the fucking code before you install any tool
+ − 52 # especially this one
+ − 53
+ − 54 # After you get the script working on some test data, you can
+ − 55 # optionally generate a toolshed compatible gzip file
+ − 56 # containing your script safely wrapped as an ordinary Galaxy script in your local toolshed for
+ − 57 # safe and largely automated installation in a production Galaxy.
+ − 58
+ − 59 # If you opt for an HTML output, you get all the script outputs arranged
+ − 60 # as a single Html history item - all output files are linked, thumbnails for all the pdfs.
+ − 61 # Ugly but really inexpensive.
+ − 62 #
+ − 63 # Patches appreciated please.
+ − 64 #
+ − 65 #
+ − 66 # long route to June 2012 product
+ − 67 # Behold the awesome power of Galaxy and the toolshed with the tool factory to bind them
+ − 68 # derived from an integrated script model
+ − 69 # called rgBaseScriptWrapper.py
+ − 70 # Note to the unwary:
+ − 71 # This tool allows arbitrary scripting on your Galaxy as the Galaxy user
+ − 72 # There is nothing stopping a malicious user doing whatever they choose
+ − 73 # Extremely dangerous!!
+ − 74 # Totally insecure. So, trusted users only
+ − 75 #
+ − 76 # preferred model is a developer using their throw away workstation instance - ie a private site.
+ − 77 # no real risk. The universe_wsgi.ini admin_users string is checked - only admin users are permitted to run this tool.
+ − 78 #
+ − 79
+ − 80 import sys
+ − 81 import shutil
+ − 82 import subprocess
+ − 83 import os
+ − 84 import time
+ − 85 import tempfile
+ − 86 import optparse
+ − 87 import tarfile
+ − 88 import re
+ − 89 import shutil
+ − 90 import math
+ − 91
+ − 92 progname = os.path.split(sys.argv[0])[1]
+ − 93 myversion = 'V001.1 March 2014'
+ − 94 verbose = False
+ − 95 debug = False
+ − 96 toolFactoryURL = 'https://bitbucket.org/fubar/galaxytoolfactory'
+ − 97
+ − 98 # if we do html we need these dependencies specified in a tool_dependencies.xml file and referred to in the generated
+ − 99 # tool xml
+ − 100 toolhtmldepskel = """<?xml version="1.0"?>
+ − 101 <tool_dependency>
+ − 102 <package name="ghostscript" version="9.10">
+ − 103 <repository name="package_ghostscript_9_10" owner="devteam" prior_installation_required="True" />
+ − 104 </package>
+ − 105 <package name="graphicsmagick" version="1.3.18">
+ − 106 <repository name="package_graphicsmagick_1_3" owner="iuc" prior_installation_required="True" />
+ − 107 </package>
+ − 108 <readme>
+ − 109 %s
+ − 110 </readme>
+ − 111 </tool_dependency>
+ − 112 """
+ − 113
+ − 114 protorequirements = """<requirements>
+ − 115 <requirement type="package" version="9.10">ghostscript</requirement>
+ − 116 <requirement type="package" version="1.3.18">graphicsmagick</requirement>
+ − 117 </requirements>"""
+ − 118
+ − 119 def timenow():
+ − 120 """return current time as a string
+ − 121 """
+ − 122 return time.strftime('%d/%m/%Y %H:%M:%S', time.localtime(time.time()))
+ − 123
+ − 124 html_escape_table = {
+ − 125 "&": "&",
+ − 126 ">": ">",
+ − 127 "<": "<",
+ − 128 "$": "\$"
+ − 129 }
+ − 130
+ − 131 def html_escape(text):
+ − 132 """Produce entities within text."""
+ − 133 return "".join(html_escape_table.get(c,c) for c in text)
+ − 134
+ − 135 def cmd_exists(cmd):
+ − 136 return subprocess.call("type " + cmd, shell=True,
+ − 137 stdout=subprocess.PIPE, stderr=subprocess.PIPE) == 0
+ − 138
+ − 139
+ − 140 def parse_citations(citations_text):
+ − 141 """
+ − 142 """
+ − 143 citations = [c for c in citations_text.split("**ENTRY**") if c.strip()]
+ − 144 citation_tuples = []
+ − 145 for citation in citations:
+ − 146 if citation.startswith("doi"):
+ − 147 citation_tuples.append( ("doi", citation[len("doi"):].strip() ) )
+ − 148 else:
+ − 149 citation_tuples.append( ("bibtex", citation[len("bibtex"):].strip() ) )
+ − 150 return citation_tuples
+ − 151
+ − 152
+ − 153 class ScriptRunner:
+ − 154 """class is a wrapper for an arbitrary script
+ − 155 """
+ − 156
+ − 157 def __init__(self,opts=None,treatbashSpecial=True):
+ − 158 """
+ − 159 cleanup inputs, setup some outputs
+ − 160
+ − 161 """
+ − 162 self.useGM = cmd_exists('gm')
+ − 163 self.useIM = cmd_exists('convert')
+ − 164 self.useGS = cmd_exists('gs')
+ − 165 self.temp_warned = False # we want only one warning if $TMP not set
+ − 166 self.treatbashSpecial = treatbashSpecial
+ − 167 if opts.output_dir: # simplify for the tool tarball
+ − 168 os.chdir(opts.output_dir)
+ − 169 self.thumbformat = 'png'
+ − 170 self.opts = opts
+ − 171 self.toolname = re.sub('[^a-zA-Z0-9_]+', '', opts.tool_name) # a sanitizer now does this but..
+ − 172 self.toolid = self.toolname
+ − 173 self.myname = sys.argv[0] # get our name because we write ourselves out as a tool later
+ − 174 self.pyfile = self.myname # crude but efficient - the cruft won't hurt much
+ − 175 self.xmlfile = '%s.xml' % self.toolname
+ − 176 s = open(self.opts.script_path,'r').readlines()
+ − 177 s = [x.rstrip() for x in s] # remove pesky dos line endings if needed
+ − 178 self.script = '\n'.join(s)
+ − 179 fhandle,self.sfile = tempfile.mkstemp(prefix=self.toolname,suffix=".%s" % (opts.interpreter))
+ − 180 tscript = open(self.sfile,'w') # use self.sfile as script source for Popen
+ − 181 tscript.write(self.script)
+ − 182 tscript.close()
+ − 183 self.indentedScript = '\n'.join([' %s' % html_escape(x) for x in s]) # for restructured text in help
+ − 184 self.escapedScript = '\n'.join([html_escape(x) for x in s])
+ − 185 self.elog = os.path.join(self.opts.output_dir,"%s_error.log" % self.toolname)
+ − 186 if opts.output_dir: # may not want these complexities
+ − 187 self.tlog = os.path.join(self.opts.output_dir,"%s_runner.log" % self.toolname)
+ − 188 art = '%s.%s' % (self.toolname,opts.interpreter)
+ − 189 artpath = os.path.join(self.opts.output_dir,art) # need full path
+ − 190 artifact = open(artpath,'w') # use self.sfile as script source for Popen
+ − 191 artifact.write(self.script)
+ − 192 artifact.close()
+ − 193 self.cl = []
+ − 194 self.html = []
+ − 195 a = self.cl.append
+ − 196 a(opts.interpreter)
+ − 197 if self.treatbashSpecial and opts.interpreter in ['bash','sh']:
+ − 198 a(self.sfile)
+ − 199 else:
+ − 200 a('-') # stdin
+ − 201 a(opts.input_tab)
+ − 202 a(opts.output_tab)
+ − 203 self.outFormats = 'tabular' # TODO make this an option at tool generation time
+ − 204 self.inputFormats = 'tabular,txt' # TODO make this an option at tool generation time
+ − 205 self.test1Input = '%s_test1_input.xls' % self.toolname
+ − 206 self.test1Output = '%s_test1_output.xls' % self.toolname
+ − 207 self.test1HTML = '%s_test1_output.html' % self.toolname
+ − 208
+ − 209 def makeXML(self):
+ − 210 """
+ − 211 Create a Galaxy xml tool wrapper for the new script as a string to write out
+ − 212 fixme - use templating or something less fugly than this example of what we produce
+ − 213
+ − 214 <tool id="reverse" name="reverse" version="0.01">
+ − 215 <description>a tabular file</description>
+ − 216 <command interpreter="python">
+ − 217 reverse.py --script_path "$runMe" --interpreter "python"
+ − 218 --tool_name "reverse" --input_tab "$input1" --output_tab "$tab_file"
+ − 219 </command>
+ − 220 <inputs>
+ − 221 <param name="input1" type="data" format="tabular" label="Select a suitable input file from your history"/><param name="job_name" type="text" label="Supply a name for the outputs to remind you what they contain" value="reverse"/>
+ − 222
+ − 223 </inputs>
+ − 224 <outputs>
+ − 225 <data format="tabular" name="tab_file" label="${job_name}"/>
+ − 226
+ − 227 </outputs>
+ − 228 <help>
+ − 229
+ − 230 **What it Does**
+ − 231
+ − 232 Reverse the columns in a tabular file
+ − 233
+ − 234 </help>
+ − 235 <configfiles>
+ − 236 <configfile name="runMe">
+ − 237
+ − 238 # reverse order of columns in a tabular file
+ − 239 import sys
+ − 240 inp = sys.argv[1]
+ − 241 outp = sys.argv[2]
+ − 242 i = open(inp,'r')
+ − 243 o = open(outp,'w')
+ − 244 for row in i:
+ − 245 rs = row.rstrip().split('\t')
+ − 246 rs.reverse()
+ − 247 o.write('\t'.join(rs))
+ − 248 o.write('\n')
+ − 249 i.close()
+ − 250 o.close()
+ − 251
+ − 252
+ − 253 </configfile>
+ − 254 </configfiles>
+ − 255 </tool>
+ − 256
+ − 257 """
+ − 258 newXML="""<tool id="%(toolid)s" name="%(toolname)s" version="%(tool_version)s">
+ − 259 %(tooldesc)s
+ − 260 %(requirements)s
+ − 261 <command interpreter="python">
+ − 262 %(command)s
+ − 263 </command>
+ − 264 <inputs>
+ − 265 %(inputs)s
+ − 266 </inputs>
+ − 267 <outputs>
+ − 268 %(outputs)s
+ − 269 </outputs>
+ − 270 <configfiles>
+ − 271 <configfile name="runMe">
+ − 272 %(script)s
+ − 273 </configfile>
+ − 274 </configfiles>
+ − 275
+ − 276 %(tooltests)s
+ − 277
+ − 278 <help>
+ − 279
+ − 280 %(help)s
+ − 281
+ − 282 </help>
+ − 283 <citations>
+ − 284 %(citations)s
+ − 285 <citation type="doi">10.1093/bioinformatics/bts573</citation>
+ − 286 </citations>
+ − 287 </tool>""" # needs a dict with toolname, toolid, interpreter, scriptname, command, inputs as a multi line string ready to write, outputs ditto, help ditto
+ − 288
+ − 289 newCommand="""
+ − 290 %(toolname)s.py --script_path "$runMe" --interpreter "%(interpreter)s"
+ − 291 --tool_name "%(toolname)s" %(command_inputs)s %(command_outputs)s """
+ − 292 # may NOT be an input or htmlout - appended later
+ − 293 tooltestsTabOnly = """
+ − 294 <tests>
+ − 295 <test>
+ − 296 <param name="input1" value="%(test1Input)s" ftype="tabular"/>
+ − 297 <param name="job_name" value="test1"/>
+ − 298 <param name="runMe" value="$runMe"/>
+ − 299 <output name="tab_file" file="%(test1Output)s" ftype="tabular"/>
+ − 300 </test>
+ − 301 </tests>
+ − 302 """
+ − 303 tooltestsHTMLOnly = """
+ − 304 <tests>
+ − 305 <test>
+ − 306 <param name="input1" value="%(test1Input)s" ftype="tabular"/>
+ − 307 <param name="job_name" value="test1"/>
+ − 308 <param name="runMe" value="$runMe"/>
+ − 309 <output name="html_file" file="%(test1HTML)s" ftype="html" lines_diff="5"/>
+ − 310 </test>
+ − 311 </tests>
+ − 312 """
+ − 313 tooltestsBoth = """<tests>
+ − 314 <test>
+ − 315 <param name="input1" value="%(test1Input)s" ftype="tabular"/>
+ − 316 <param name="job_name" value="test1"/>
+ − 317 <param name="runMe" value="$runMe"/>
+ − 318 <output name="tab_file" file="%(test1Output)s" ftype="tabular" />
+ − 319 <output name="html_file" file="%(test1HTML)s" ftype="html" lines_diff="10"/>
+ − 320 </test>
+ − 321 </tests>
+ − 322 """
+ − 323 xdict = {}
+ − 324 xdict['requirements'] = ''
+ − 325 if self.opts.make_HTML:
+ − 326 if self.opts.include_dependencies == "yes":
+ − 327 xdict['requirements'] = protorequirements
+ − 328 xdict['tool_version'] = self.opts.tool_version
+ − 329 xdict['test1Input'] = self.test1Input
+ − 330 xdict['test1HTML'] = self.test1HTML
+ − 331 xdict['test1Output'] = self.test1Output
+ − 332 if self.opts.make_HTML and self.opts.output_tab <> 'None':
+ − 333 xdict['tooltests'] = tooltestsBoth % xdict
+ − 334 elif self.opts.make_HTML:
+ − 335 xdict['tooltests'] = tooltestsHTMLOnly % xdict
+ − 336 else:
+ − 337 xdict['tooltests'] = tooltestsTabOnly % xdict
+ − 338 xdict['script'] = self.escapedScript
+ − 339 # configfile is least painful way to embed script to avoid external dependencies
+ − 340 # but requires escaping of <, > and $ to avoid Mako parsing
+ − 341 if self.opts.help_text:
+ − 342 helptext = open(self.opts.help_text,'r').readlines()
+ − 343 helptext = [html_escape(x) for x in helptext] # must html escape here too - thanks to Marius van den Beek
+ − 344 xdict['help'] = ''.join([x for x in helptext])
+ − 345 else:
+ − 346 xdict['help'] = 'Please ask the tool author (%s) for help as none was supplied at tool generation\n' % (self.opts.user_email)
+ − 347 if self.opts.citations:
+ − 348 citationstext = open(self.opts.citations,'r').read()
+ − 349 citation_tuples = parse_citations(citationstext)
+ − 350 citations_xml = ""
+ − 351 for citation_type, citation_content in citation_tuples:
+ − 352 citation_xml = """<citation type="%s">%s</citation>""" % (citation_type, html_escape(citation_content))
+ − 353 citations_xml += citation_xml
+ − 354 xdict['citations'] = citations_xml
+ − 355 else:
+ − 356 xdict['citations'] = ""
+ − 357 coda = ['**Script**','Pressing execute will run the following code over your input file and generate some outputs in your history::']
+ − 358 coda.append('\n')
+ − 359 coda.append(self.indentedScript)
+ − 360 coda.append('\n**Attribution**\nThis Galaxy tool was created by %s at %s\nusing the Galaxy Tool Factory.\n' % (self.opts.user_email,timenow()))
+ − 361 coda.append('See %s for details of that project' % (toolFactoryURL))
+ − 362 coda.append('Please cite: Creating re-usable tools from scripts: The Galaxy Tool Factory. Ross Lazarus; Antony Kaspi; Mark Ziemann; The Galaxy Team. ')
+ − 363 coda.append('Bioinformatics 2012; doi: 10.1093/bioinformatics/bts573\n')
+ − 364 xdict['help'] = '%s\n%s' % (xdict['help'],'\n'.join(coda))
+ − 365 if self.opts.tool_desc:
+ − 366 xdict['tooldesc'] = '<description>%s</description>' % self.opts.tool_desc
+ − 367 else:
+ − 368 xdict['tooldesc'] = ''
+ − 369 xdict['command_outputs'] = ''
+ − 370 xdict['outputs'] = ''
+ − 371 if self.opts.input_tab <> 'None':
+ − 372 xdict['command_inputs'] = '--input_tab "$input1" ' # the space may matter a lot if we append something
+ − 373 xdict['inputs'] = '<param name="input1" type="data" format="%s" label="Select a suitable input file from your history"/> \n' % self.inputFormats
+ − 374 else:
+ − 375 xdict['command_inputs'] = '' # assume no input - eg a random data generator
+ − 376 xdict['inputs'] = ''
+ − 377 xdict['inputs'] += '<param name="job_name" type="text" label="Supply a name for the outputs to remind you what they contain" value="%s"/> \n' % self.toolname
+ − 378 xdict['toolname'] = self.toolname
+ − 379 xdict['toolid'] = self.toolid
+ − 380 xdict['interpreter'] = self.opts.interpreter
+ − 381 xdict['scriptname'] = self.sfile
+ − 382 if self.opts.make_HTML:
+ − 383 xdict['command_outputs'] += ' --output_dir "$html_file.files_path" --output_html "$html_file" --make_HTML "yes"'
+ − 384 xdict['outputs'] += ' <data format="html" name="html_file" label="${job_name}.html"/>\n'
+ − 385 else:
+ − 386 xdict['command_outputs'] += ' --output_dir "./"'
+ − 387 if self.opts.output_tab <> 'None':
+ − 388 xdict['command_outputs'] += ' --output_tab "$tab_file"'
+ − 389 xdict['outputs'] += ' <data format="%s" name="tab_file" label="${job_name}"/>\n' % self.outFormats
+ − 390 xdict['command'] = newCommand % xdict
+ − 391 xmls = newXML % xdict
+ − 392 xf = open(self.xmlfile,'w')
+ − 393 xf.write(xmls)
+ − 394 xf.write('\n')
+ − 395 xf.close()
+ − 396 # ready for the tarball
+ − 397
+ − 398
+ − 399 def makeTooltar(self):
+ − 400 """
+ − 401 a tool is a gz tarball with eg
+ − 402 /toolname/tool.xml /toolname/tool.py /toolname/test-data/test1_in.foo ...
+ − 403 """
+ − 404 retval = self.run()
+ − 405 if retval:
+ − 406 print >> sys.stderr,'## Run failed. Cannot build yet. Please fix and retry'
+ − 407 sys.exit(1)
+ − 408 tdir = self.toolname
+ − 409 os.mkdir(tdir)
+ − 410 self.makeXML()
+ − 411 if self.opts.make_HTML:
+ − 412 if self.opts.help_text:
+ − 413 hlp = open(self.opts.help_text,'r').read()
+ − 414 else:
+ − 415 hlp = 'Please ask the tool author for help as none was supplied at tool generation\n'
+ − 416 if self.opts.include_dependencies:
+ − 417 tooldepcontent = toolhtmldepskel % hlp
+ − 418 depf = open(os.path.join(tdir,'tool_dependencies.xml'),'w')
+ − 419 depf.write(tooldepcontent)
+ − 420 depf.write('\n')
+ − 421 depf.close()
+ − 422 if self.opts.input_tab <> 'None': # no reproducible test otherwise? TODO: maybe..
+ − 423 testdir = os.path.join(tdir,'test-data')
+ − 424 os.mkdir(testdir) # make tests directory
+ − 425 shutil.copyfile(self.opts.input_tab,os.path.join(testdir,self.test1Input))
+ − 426 if self.opts.output_tab <> 'None':
+ − 427 shutil.copyfile(self.opts.output_tab,os.path.join(testdir,self.test1Output))
+ − 428 if self.opts.make_HTML:
+ − 429 shutil.copyfile(self.opts.output_html,os.path.join(testdir,self.test1HTML))
+ − 430 if self.opts.output_dir:
+ − 431 shutil.copyfile(self.tlog,os.path.join(testdir,'test1_out.log'))
+ − 432 outpif = '%s.py' % self.toolname # new name
+ − 433 outpiname = os.path.join(tdir,outpif) # path for the tool tarball
+ − 434 pyin = os.path.basename(self.pyfile) # our name - we rewrite ourselves (TM)
+ − 435 notes = ['# %s - a self annotated version of %s generated by running %s\n' % (outpiname,pyin,pyin),]
+ − 436 notes.append('# to make a new Galaxy tool called %s\n' % self.toolname)
+ − 437 notes.append('# User %s at %s\n' % (self.opts.user_email,timenow()))
+ − 438 pi = open(self.pyfile,'r').readlines() # our code becomes new tool wrapper (!) - first Galaxy worm
+ − 439 notes += pi
+ − 440 outpi = open(outpiname,'w')
+ − 441 outpi.write(''.join(notes))
+ − 442 outpi.write('\n')
+ − 443 outpi.close()
+ − 444 stname = os.path.join(tdir,self.sfile)
+ − 445 if not os.path.exists(stname):
+ − 446 shutil.copyfile(self.sfile, stname)
+ − 447 xtname = os.path.join(tdir,self.xmlfile)
+ − 448 if not os.path.exists(xtname):
+ − 449 shutil.copyfile(self.xmlfile,xtname)
+ − 450 tarpath = "%s.gz" % self.toolname
+ − 451 tar = tarfile.open(tarpath, "w:gz")
+ − 452 tar.add(tdir,arcname=self.toolname)
+ − 453 tar.close()
+ − 454 shutil.copyfile(tarpath,self.opts.new_tool)
+ − 455 shutil.rmtree(tdir)
+ − 456 ## TODO: replace with optional direct upload to local toolshed?
+ − 457 return retval
+ − 458
+ − 459
+ − 460 def compressPDF(self,inpdf=None,thumbformat='png'):
+ − 461 """need absolute path to pdf
+ − 462 note that GS gets confoozled if no $TMP or $TEMP
+ − 463 so we set it
+ − 464 """
+ − 465 assert os.path.isfile(inpdf), "## Input %s supplied to %s compressPDF not found" % (inpdf,self.myName)
+ − 466 hlog = os.path.join(self.opts.output_dir,"compress_%s.txt" % os.path.basename(inpdf))
+ − 467 sto = open(hlog,'a')
+ − 468 our_env = os.environ.copy()
+ − 469 our_tmp = our_env.get('TMP',None)
+ − 470 if not our_tmp:
+ − 471 our_tmp = our_env.get('TEMP',None)
+ − 472 if not (our_tmp and os.path.exists(our_tmp)):
+ − 473 newtmp = os.path.join(self.opts.output_dir,'tmp')
+ − 474 try:
+ − 475 os.mkdir(newtmp)
+ − 476 except:
+ − 477 sto.write('## WARNING - cannot make %s - it may exist or permissions need fixing\n' % newtmp)
+ − 478 our_env['TEMP'] = newtmp
+ − 479 if not self.temp_warned:
+ − 480 sto.write('## WARNING - no $TMP or $TEMP!!! Please fix - using %s temporarily\n' % newtmp)
+ − 481 self.temp_warned = True
+ − 482 outpdf = '%s_compressed' % inpdf
+ − 483 cl = ["gs", "-sDEVICE=pdfwrite", "-dNOPAUSE", "-dUseCIEColor", "-dBATCH","-dPDFSETTINGS=/printer", "-sOutputFile=%s" % outpdf,inpdf]
+ − 484 x = subprocess.Popen(cl,stdout=sto,stderr=sto,cwd=self.opts.output_dir,env=our_env)
+ − 485 retval1 = x.wait()
+ − 486 sto.close()
+ − 487 if retval1 == 0:
+ − 488 os.unlink(inpdf)
+ − 489 shutil.move(outpdf,inpdf)
+ − 490 os.unlink(hlog)
+ − 491 hlog = os.path.join(self.opts.output_dir,"thumbnail_%s.txt" % os.path.basename(inpdf))
+ − 492 sto = open(hlog,'w')
+ − 493 outpng = '%s.%s' % (os.path.splitext(inpdf)[0],thumbformat)
+ − 494 if self.useGM:
+ − 495 cl2 = ['gm', 'convert', inpdf, outpng]
+ − 496 else: # assume imagemagick
+ − 497 cl2 = ['convert', inpdf, outpng]
+ − 498 x = subprocess.Popen(cl2,stdout=sto,stderr=sto,cwd=self.opts.output_dir,env=our_env)
+ − 499 retval2 = x.wait()
+ − 500 sto.close()
+ − 501 if retval2 == 0:
+ − 502 os.unlink(hlog)
+ − 503 retval = retval1 or retval2
+ − 504 return retval
+ − 505
+ − 506
+ − 507 def getfSize(self,fpath,outpath):
+ − 508 """
+ − 509 format a nice file size string
+ − 510 """
+ − 511 size = ''
+ − 512 fp = os.path.join(outpath,fpath)
+ − 513 if os.path.isfile(fp):
+ − 514 size = '0 B'
+ − 515 n = float(os.path.getsize(fp))
+ − 516 if n > 2**20:
+ − 517 size = '%1.1f MB' % (n/2**20)
+ − 518 elif n > 2**10:
+ − 519 size = '%1.1f KB' % (n/2**10)
+ − 520 elif n > 0:
+ − 521 size = '%d B' % (int(n))
+ − 522 return size
+ − 523
+ − 524 def makeHtml(self):
+ − 525 """ Create an HTML file content to list all the artifacts found in the output_dir
+ − 526 """
+ − 527
+ − 528 galhtmlprefix = """<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+ − 529 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
+ − 530 <head> <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+ − 531 <meta name="generator" content="Galaxy %s tool output - see http://g2.trac.bx.psu.edu/" />
+ − 532 <title></title>
+ − 533 <link rel="stylesheet" href="/static/style/base.css" type="text/css" />
+ − 534 </head>
+ − 535 <body>
+ − 536 <div class="toolFormBody">
+ − 537 """
+ − 538 galhtmlattr = """<hr/><div class="infomessage">This tool (%s) was generated by the <a href="https://bitbucket.org/fubar/galaxytoolfactory/overview">Galaxy Tool Factory</a></div><br/>"""
+ − 539 galhtmlpostfix = """</div></body></html>\n"""
+ − 540
+ − 541 flist = os.listdir(self.opts.output_dir)
+ − 542 flist = [x for x in flist if x <> 'Rplots.pdf']
+ − 543 flist.sort()
+ − 544 html = []
+ − 545 html.append(galhtmlprefix % progname)
+ − 546 html.append('<div class="infomessage">Galaxy Tool "%s" run at %s</div><br/>' % (self.toolname,timenow()))
+ − 547 fhtml = []
+ − 548 if len(flist) > 0:
+ − 549 logfiles = [x for x in flist if x.lower().endswith('.log')] # log file names determine sections
+ − 550 logfiles.sort()
+ − 551 logfiles = [x for x in logfiles if os.path.abspath(x) <> os.path.abspath(self.tlog)]
+ − 552 logfiles.append(os.path.abspath(self.tlog)) # make it the last one
+ − 553 pdflist = []
+ − 554 npdf = len([x for x in flist if os.path.splitext(x)[-1].lower() == '.pdf'])
+ − 555 for rownum,fname in enumerate(flist):
+ − 556 dname,e = os.path.splitext(fname)
+ − 557 sfsize = self.getfSize(fname,self.opts.output_dir)
+ − 558 if e.lower() == '.pdf' : # compress and make a thumbnail
+ − 559 thumb = '%s.%s' % (dname,self.thumbformat)
+ − 560 pdff = os.path.join(self.opts.output_dir,fname)
+ − 561 retval = self.compressPDF(inpdf=pdff,thumbformat=self.thumbformat)
+ − 562 if retval == 0:
+ − 563 pdflist.append((fname,thumb))
+ − 564 else:
+ − 565 pdflist.append((fname,fname))
+ − 566 if (rownum+1) % 2 == 0:
+ − 567 fhtml.append('<tr class="odd_row"><td><a href="%s">%s</a></td><td>%s</td></tr>' % (fname,fname,sfsize))
+ − 568 else:
+ − 569 fhtml.append('<tr><td><a href="%s">%s</a></td><td>%s</td></tr>' % (fname,fname,sfsize))
+ − 570 for logfname in logfiles: # expect at least tlog - if more
+ − 571 if os.path.abspath(logfname) == os.path.abspath(self.tlog): # handled later
+ − 572 sectionname = 'All tool run'
+ − 573 if (len(logfiles) > 1):
+ − 574 sectionname = 'Other'
+ − 575 ourpdfs = pdflist
+ − 576 else:
+ − 577 realname = os.path.basename(logfname)
+ − 578 sectionname = os.path.splitext(realname)[0].split('_')[0] # break in case _ added to log
+ − 579 ourpdfs = [x for x in pdflist if os.path.basename(x[0]).split('_')[0] == sectionname]
+ − 580 pdflist = [x for x in pdflist if os.path.basename(x[0]).split('_')[0] <> sectionname] # remove
+ − 581 nacross = 1
+ − 582 npdf = len(ourpdfs)
+ − 583
+ − 584 if npdf > 0:
+ − 585 nacross = math.sqrt(npdf) ## int(round(math.log(npdf,2)))
+ − 586 if int(nacross)**2 != npdf:
+ − 587 nacross += 1
+ − 588 nacross = int(nacross)
+ − 589 width = min(400,int(1200/nacross))
+ − 590 html.append('<div class="toolFormTitle">%s images and outputs</div>' % sectionname)
+ − 591 html.append('(Click on a thumbnail image to download the corresponding original PDF image)<br/>')
+ − 592 ntogo = nacross # counter for table row padding with empty cells
+ − 593 html.append('<div><table class="simple" cellpadding="2" cellspacing="2">\n<tr>')
+ − 594 for i,paths in enumerate(ourpdfs):
+ − 595 fname,thumb = paths
+ − 596 s= """<td><a href="%s"><img src="%s" title="Click to download a PDF of %s" hspace="5" width="%d"
+ − 597 alt="Image called %s"/></a></td>\n""" % (fname,thumb,fname,width,fname)
+ − 598 if ((i+1) % nacross == 0):
+ − 599 s += '</tr>\n'
+ − 600 ntogo = 0
+ − 601 if i < (npdf - 1): # more to come
+ − 602 s += '<tr>'
+ − 603 ntogo = nacross
+ − 604 else:
+ − 605 ntogo -= 1
+ − 606 html.append(s)
+ − 607 if html[-1].strip().endswith('</tr>'):
+ − 608 html.append('</table></div>\n')
+ − 609 else:
+ − 610 if ntogo > 0: # pad
+ − 611 html.append('<td> </td>'*ntogo)
+ − 612 html.append('</tr></table></div>\n')
+ − 613 logt = open(logfname,'r').readlines()
+ − 614 logtext = [x for x in logt if x.strip() > '']
+ − 615 html.append('<div class="toolFormTitle">%s log output</div>' % sectionname)
+ − 616 if len(logtext) > 1:
+ − 617 html.append('\n<pre>\n')
+ − 618 html += logtext
+ − 619 html.append('\n</pre>\n')
+ − 620 else:
+ − 621 html.append('%s is empty<br/>' % logfname)
+ − 622 if len(fhtml) > 0:
+ − 623 fhtml.insert(0,'<div><table class="colored" cellpadding="3" cellspacing="3"><tr><th>Output File Name (click to view)</th><th>Size</th></tr>\n')
+ − 624 fhtml.append('</table></div><br/>')
+ − 625 html.append('<div class="toolFormTitle">All output files available for downloading</div>\n')
+ − 626 html += fhtml # add all non-pdf files to the end of the display
+ − 627 else:
+ − 628 html.append('<div class="warningmessagelarge">### Error - %s returned no files - please confirm that parameters are sane</div>' % self.opts.interpreter)
+ − 629 html.append(galhtmlpostfix)
+ − 630 htmlf = file(self.opts.output_html,'w')
+ − 631 htmlf.write('\n'.join(html))
+ − 632 htmlf.write('\n')
+ − 633 htmlf.close()
+ − 634 self.html = html
+ − 635
+ − 636
+ − 637 def run(self):
+ − 638 """
+ − 639 scripts must be small enough not to fill the pipe!
+ − 640 """
+ − 641 if self.treatbashSpecial and self.opts.interpreter in ['bash','sh']:
+ − 642 retval = self.runBash()
+ − 643 else:
+ − 644 if self.opts.output_dir:
+ − 645 ste = open(self.elog,'w')
+ − 646 sto = open(self.tlog,'w')
+ − 647 sto.write('## Toolfactory generated command line = %s\n' % ' '.join(self.cl))
+ − 648 sto.flush()
+ − 649 p = subprocess.Popen(self.cl,shell=False,stdout=sto,stderr=ste,stdin=subprocess.PIPE,cwd=self.opts.output_dir)
+ − 650 else:
+ − 651 p = subprocess.Popen(self.cl,shell=False,stdin=subprocess.PIPE)
+ − 652 p.stdin.write(self.script)
+ − 653 p.stdin.close()
+ − 654 retval = p.wait()
+ − 655 if self.opts.output_dir:
+ − 656 sto.close()
+ − 657 ste.close()
+ − 658 err = open(self.elog,'r').readlines()
+ − 659 if retval <> 0 and err: # problem
+ − 660 print >> sys.stderr,err
+ − 661 if self.opts.make_HTML:
+ − 662 self.makeHtml()
+ − 663 return retval
+ − 664
+ − 665 def runBash(self):
+ − 666 """
+ − 667 cannot use - for bash so use self.sfile
+ − 668 """
+ − 669 if self.opts.output_dir:
+ − 670 s = '## Toolfactory generated command line = %s\n' % ' '.join(self.cl)
+ − 671 sto = open(self.tlog,'w')
+ − 672 sto.write(s)
+ − 673 sto.flush()
+ − 674 p = subprocess.Popen(self.cl,shell=False,stdout=sto,stderr=sto,cwd=self.opts.output_dir)
+ − 675 else:
+ − 676 p = subprocess.Popen(self.cl,shell=False)
+ − 677 retval = p.wait()
+ − 678 if self.opts.output_dir:
+ − 679 sto.close()
+ − 680 if self.opts.make_HTML:
+ − 681 self.makeHtml()
+ − 682 return retval
+ − 683
+ − 684
+ − 685 def main():
+ − 686 u = """
+ − 687 This is a Galaxy wrapper. It expects to be called by a special purpose tool.xml as:
+ − 688 <command interpreter="python">rgBaseScriptWrapper.py --script_path "$scriptPath" --tool_name "foo" --interpreter "Rscript"
+ − 689 </command>
+ − 690 """
+ − 691 op = optparse.OptionParser()
+ − 692 a = op.add_option
+ − 693 a('--script_path',default=None)
+ − 694 a('--tool_name',default=None)
+ − 695 a('--interpreter',default=None)
+ − 696 a('--output_dir',default='./')
+ − 697 a('--output_html',default=None)
+ − 698 a('--input_tab',default="None")
+ − 699 a('--output_tab',default="None")
+ − 700 a('--user_email',default='Unknown')
+ − 701 a('--bad_user',default=None)
+ − 702 a('--make_Tool',default=None)
+ − 703 a('--make_HTML',default=None)
+ − 704 a('--help_text',default=None)
+ − 705 a('--citations',default=None)
+ − 706 a('--tool_desc',default=None)
+ − 707 a('--new_tool',default=None)
+ − 708 a('--tool_version',default=None)
+ − 709 a('--include_dependencies',default=None)
+ − 710 opts, args = op.parse_args()
+ − 711 assert not opts.bad_user,'UNAUTHORISED: %s is NOT authorized to use this tool until Galaxy admin adds %s to admin_users in universe_wsgi.ini' % (opts.bad_user,opts.bad_user)
+ − 712 assert opts.tool_name,'## Tool Factory expects a tool name - eg --tool_name=DESeq'
+ − 713 assert opts.interpreter,'## Tool Factory wrapper expects an interpreter - eg --interpreter=Rscript'
+ − 714 assert os.path.isfile(opts.script_path),'## Tool Factory wrapper expects a script path - eg --script_path=foo.R'
+ − 715 if opts.output_dir:
+ − 716 try:
+ − 717 os.makedirs(opts.output_dir)
+ − 718 except:
+ − 719 pass
+ − 720 r = ScriptRunner(opts)
+ − 721 if opts.make_Tool:
+ − 722 retcode = r.makeTooltar()
+ − 723 else:
+ − 724 retcode = r.run()
+ − 725 os.unlink(r.sfile)
+ − 726 if retcode:
+ − 727 sys.exit(retcode) # indicate failure to job runner
+ − 728
+ − 729
+ − 730 if __name__ == "__main__":
+ − 731 main()
+ − 732
+ − 733
