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 #
41
+ − 10 # August 2014
+ − 11 # merged John Chilton's citation addition and ideas from Marius van den Beek to enable arbitrary
+ − 12 # data types for input and output - thanks!
+ − 13 #
37
+ − 14 # march 2014
+ − 15 # had to remove dependencies because cross toolshed dependencies are not possible - can't pre-specify a toolshed url for graphicsmagick and ghostscript
+ − 16 # grrrrr - night before a demo
+ − 17 # added dependencies to a tool_dependencies.xml if html page generated so generated tool is properly portable
+ − 18 #
+ − 19 # added ghostscript and graphicsmagick as dependencies
+ − 20 # fixed a wierd problem where gs was trying to use the new_files_path from universe (database/tmp) as ./database/tmp
+ − 21 # errors ensued
+ − 22 #
+ − 23 # august 2013
+ − 24 # found a problem with GS if $TMP or $TEMP missing - now inject /tmp and warn
+ − 25 #
+ − 26 # july 2013
+ − 27 # added ability to combine images and individual log files into html output
+ − 28 # just make sure there's a log file foo.log and it will be output
+ − 29 # together with all images named like "foo_*.pdf
+ − 30 # otherwise old format for html
+ − 31 #
+ − 32 # January 2013
+ − 33 # problem pointed out by Carlos Borroto
+ − 34 # added escaping for <>$ - thought I did that ages ago...
+ − 35 #
+ − 36 # August 11 2012
+ − 37 # changed to use shell=False and cl as a sequence
+ − 38
+ − 39 # This is a Galaxy tool factory for simple scripts in python, R or whatever ails ye.
+ − 40 # It also serves as the wrapper for the new tool.
+ − 41 #
+ − 42 # you paste and run your script
+ − 43 # Only works for simple scripts that read one input from the history.
+ − 44 # Optionally can write one new history dataset,
+ − 45 # and optionally collect any number of outputs into links on an autogenerated HTML page.
+ − 46
+ − 47 # DO NOT install on a public or important site - please.
+ − 48
+ − 49 # installed generated tools are fine if the script is safe.
+ − 50 # They just run normally and their user cannot do anything unusually insecure
+ − 51 # but please, practice safe toolshed.
+ − 52 # Read the fucking code before you install any tool
+ − 53 # especially this one
+ − 54
+ − 55 # After you get the script working on some test data, you can
+ − 56 # optionally generate a toolshed compatible gzip file
+ − 57 # containing your script safely wrapped as an ordinary Galaxy script in your local toolshed for
+ − 58 # safe and largely automated installation in a production Galaxy.
+ − 59
+ − 60 # If you opt for an HTML output, you get all the script outputs arranged
+ − 61 # as a single Html history item - all output files are linked, thumbnails for all the pdfs.
+ − 62 # Ugly but really inexpensive.
+ − 63 #
+ − 64 # Patches appreciated please.
+ − 65 #
+ − 66 #
+ − 67 # long route to June 2012 product
+ − 68 # Behold the awesome power of Galaxy and the toolshed with the tool factory to bind them
+ − 69 # derived from an integrated script model
+ − 70 # called rgBaseScriptWrapper.py
+ − 71 # Note to the unwary:
+ − 72 # This tool allows arbitrary scripting on your Galaxy as the Galaxy user
+ − 73 # There is nothing stopping a malicious user doing whatever they choose
+ − 74 # Extremely dangerous!!
+ − 75 # Totally insecure. So, trusted users only
+ − 76 #
+ − 77 # preferred model is a developer using their throw away workstation instance - ie a private site.
+ − 78 # no real risk. The universe_wsgi.ini admin_users string is checked - only admin users are permitted to run this tool.
+ − 79 #
+ − 80
+ − 81 import sys
+ − 82 import shutil
+ − 83 import subprocess
+ − 84 import os
+ − 85 import time
+ − 86 import tempfile
+ − 87 import optparse
+ − 88 import tarfile
+ − 89 import re
+ − 90 import shutil
+ − 91 import math
+ − 92
+ − 93 progname = os.path.split(sys.argv[0])[1]
+ − 94 myversion = 'V001.1 March 2014'
+ − 95 verbose = False
+ − 96 debug = False
+ − 97 toolFactoryURL = 'https://bitbucket.org/fubar/galaxytoolfactory'
+ − 98
+ − 99 # if we do html we need these dependencies specified in a tool_dependencies.xml file and referred to in the generated
+ − 100 # tool xml
+ − 101 toolhtmldepskel = """<?xml version="1.0"?>
+ − 102 <tool_dependency>
+ − 103 <package name="ghostscript" version="9.10">
+ − 104 <repository name="package_ghostscript_9_10" owner="devteam" prior_installation_required="True" />
+ − 105 </package>
+ − 106 <package name="graphicsmagick" version="1.3.18">
+ − 107 <repository name="package_graphicsmagick_1_3" owner="iuc" prior_installation_required="True" />
+ − 108 </package>
+ − 109 <readme>
+ − 110 %s
+ − 111 </readme>
+ − 112 </tool_dependency>
+ − 113 """
+ − 114
+ − 115 protorequirements = """<requirements>
+ − 116 <requirement type="package" version="9.10">ghostscript</requirement>
+ − 117 <requirement type="package" version="1.3.18">graphicsmagick</requirement>
+ − 118 </requirements>"""
+ − 119
+ − 120 def timenow():
+ − 121 """return current time as a string
+ − 122 """
+ − 123 return time.strftime('%d/%m/%Y %H:%M:%S', time.localtime(time.time()))
+ − 124
+ − 125 html_escape_table = {
+ − 126 "&": "&",
+ − 127 ">": ">",
+ − 128 "<": "<",
+ − 129 "$": "\$"
+ − 130 }
+ − 131
+ − 132 def html_escape(text):
+ − 133 """Produce entities within text."""
+ − 134 return "".join(html_escape_table.get(c,c) for c in text)
+ − 135
+ − 136 def cmd_exists(cmd):
+ − 137 return subprocess.call("type " + cmd, shell=True,
+ − 138 stdout=subprocess.PIPE, stderr=subprocess.PIPE) == 0
+ − 139
+ − 140
+ − 141 def parse_citations(citations_text):
+ − 142 """
+ − 143 """
+ − 144 citations = [c for c in citations_text.split("**ENTRY**") if c.strip()]
+ − 145 citation_tuples = []
+ − 146 for citation in citations:
+ − 147 if citation.startswith("doi"):
+ − 148 citation_tuples.append( ("doi", citation[len("doi"):].strip() ) )
+ − 149 else:
+ − 150 citation_tuples.append( ("bibtex", citation[len("bibtex"):].strip() ) )
+ − 151 return citation_tuples
+ − 152
+ − 153
+ − 154 class ScriptRunner:
+ − 155 """class is a wrapper for an arbitrary script
+ − 156 """
+ − 157
+ − 158 def __init__(self,opts=None,treatbashSpecial=True):
+ − 159 """
+ − 160 cleanup inputs, setup some outputs
+ − 161
+ − 162 """
+ − 163 self.useGM = cmd_exists('gm')
+ − 164 self.useIM = cmd_exists('convert')
+ − 165 self.useGS = cmd_exists('gs')
+ − 166 self.temp_warned = False # we want only one warning if $TMP not set
+ − 167 self.treatbashSpecial = treatbashSpecial
+ − 168 if opts.output_dir: # simplify for the tool tarball
+ − 169 os.chdir(opts.output_dir)
+ − 170 self.thumbformat = 'png'
+ − 171 self.opts = opts
+ − 172 self.toolname = re.sub('[^a-zA-Z0-9_]+', '', opts.tool_name) # a sanitizer now does this but..
+ − 173 self.toolid = self.toolname
+ − 174 self.myname = sys.argv[0] # get our name because we write ourselves out as a tool later
+ − 175 self.pyfile = self.myname # crude but efficient - the cruft won't hurt much
+ − 176 self.xmlfile = '%s.xml' % self.toolname
+ − 177 s = open(self.opts.script_path,'r').readlines()
+ − 178 s = [x.rstrip() for x in s] # remove pesky dos line endings if needed
+ − 179 self.script = '\n'.join(s)
+ − 180 fhandle,self.sfile = tempfile.mkstemp(prefix=self.toolname,suffix=".%s" % (opts.interpreter))
+ − 181 tscript = open(self.sfile,'w') # use self.sfile as script source for Popen
+ − 182 tscript.write(self.script)
+ − 183 tscript.close()
+ − 184 self.indentedScript = '\n'.join([' %s' % html_escape(x) for x in s]) # for restructured text in help
+ − 185 self.escapedScript = '\n'.join([html_escape(x) for x in s])
+ − 186 self.elog = os.path.join(self.opts.output_dir,"%s_error.log" % self.toolname)
+ − 187 if opts.output_dir: # may not want these complexities
+ − 188 self.tlog = os.path.join(self.opts.output_dir,"%s_runner.log" % self.toolname)
+ − 189 art = '%s.%s' % (self.toolname,opts.interpreter)
+ − 190 artpath = os.path.join(self.opts.output_dir,art) # need full path
+ − 191 artifact = open(artpath,'w') # use self.sfile as script source for Popen
+ − 192 artifact.write(self.script)
+ − 193 artifact.close()
+ − 194 self.cl = []
+ − 195 self.html = []
+ − 196 a = self.cl.append
+ − 197 a(opts.interpreter)
+ − 198 if self.treatbashSpecial and opts.interpreter in ['bash','sh']:
+ − 199 a(self.sfile)
+ − 200 else:
+ − 201 a('-') # stdin
+ − 202 a(opts.input_tab)
+ − 203 a(opts.output_tab)
41
+ − 204 self.outputFormat = self.opts.output_format
+ − 205 self.inputFormats = self.opts.input_formats
37
+ − 206 self.test1Input = '%s_test1_input.xls' % self.toolname
+ − 207 self.test1Output = '%s_test1_output.xls' % self.toolname
+ − 208 self.test1HTML = '%s_test1_output.html' % self.toolname
+ − 209
+ − 210 def makeXML(self):
+ − 211 """
+ − 212 Create a Galaxy xml tool wrapper for the new script as a string to write out
+ − 213 fixme - use templating or something less fugly than this example of what we produce
+ − 214
+ − 215 <tool id="reverse" name="reverse" version="0.01">
+ − 216 <description>a tabular file</description>
+ − 217 <command interpreter="python">
+ − 218 reverse.py --script_path "$runMe" --interpreter "python"
+ − 219 --tool_name "reverse" --input_tab "$input1" --output_tab "$tab_file"
+ − 220 </command>
+ − 221 <inputs>
+ − 222 <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"/>
+ − 223
+ − 224 </inputs>
+ − 225 <outputs>
+ − 226 <data format="tabular" name="tab_file" label="${job_name}"/>
+ − 227
+ − 228 </outputs>
+ − 229 <help>
+ − 230
+ − 231 **What it Does**
+ − 232
+ − 233 Reverse the columns in a tabular file
+ − 234
+ − 235 </help>
+ − 236 <configfiles>
+ − 237 <configfile name="runMe">
+ − 238
+ − 239 # reverse order of columns in a tabular file
+ − 240 import sys
+ − 241 inp = sys.argv[1]
+ − 242 outp = sys.argv[2]
+ − 243 i = open(inp,'r')
+ − 244 o = open(outp,'w')
+ − 245 for row in i:
+ − 246 rs = row.rstrip().split('\t')
+ − 247 rs.reverse()
+ − 248 o.write('\t'.join(rs))
+ − 249 o.write('\n')
+ − 250 i.close()
+ − 251 o.close()
+ − 252
+ − 253
+ − 254 </configfile>
+ − 255 </configfiles>
+ − 256 </tool>
+ − 257
+ − 258 """
+ − 259 newXML="""<tool id="%(toolid)s" name="%(toolname)s" version="%(tool_version)s">
+ − 260 %(tooldesc)s
+ − 261 %(requirements)s
+ − 262 <command interpreter="python">
+ − 263 %(command)s
+ − 264 </command>
+ − 265 <inputs>
+ − 266 %(inputs)s
+ − 267 </inputs>
+ − 268 <outputs>
+ − 269 %(outputs)s
+ − 270 </outputs>
+ − 271 <configfiles>
+ − 272 <configfile name="runMe">
+ − 273 %(script)s
+ − 274 </configfile>
+ − 275 </configfiles>
+ − 276
+ − 277 %(tooltests)s
+ − 278
+ − 279 <help>
+ − 280
+ − 281 %(help)s
+ − 282
+ − 283 </help>
+ − 284 <citations>
+ − 285 %(citations)s
+ − 286 <citation type="doi">10.1093/bioinformatics/bts573</citation>
+ − 287 </citations>
+ − 288 </tool>""" # needs a dict with toolname, toolid, interpreter, scriptname, command, inputs as a multi line string ready to write, outputs ditto, help ditto
+ − 289
+ − 290 newCommand="""
+ − 291 %(toolname)s.py --script_path "$runMe" --interpreter "%(interpreter)s"
+ − 292 --tool_name "%(toolname)s" %(command_inputs)s %(command_outputs)s """
+ − 293 # may NOT be an input or htmlout - appended later
+ − 294 tooltestsTabOnly = """
+ − 295 <tests>
+ − 296 <test>
41
+ − 297 <param name="input1" value="%(test1Input)s" ftype="%(inputFormats)s"/>
37
+ − 298 <param name="job_name" value="test1"/>
+ − 299 <param name="runMe" value="$runMe"/>
41
+ − 300 <output name="tab_file" file="%(test1Output)s" ftype="%(outputFormat)s"/>
37
+ − 301 </test>
+ − 302 </tests>
+ − 303 """
+ − 304 tooltestsHTMLOnly = """
+ − 305 <tests>
+ − 306 <test>
41
+ − 307 <param name="input1" value="%(test1Input)s" ftype="%(inputFormats)s"/>
37
+ − 308 <param name="job_name" value="test1"/>
+ − 309 <param name="runMe" value="$runMe"/>
+ − 310 <output name="html_file" file="%(test1HTML)s" ftype="html" lines_diff="5"/>
+ − 311 </test>
+ − 312 </tests>
+ − 313 """
+ − 314 tooltestsBoth = """<tests>
+ − 315 <test>
41
+ − 316 <param name="input1" value="%(test1Input)s" ftype="%(inputFormats)s"/>
37
+ − 317 <param name="job_name" value="test1"/>
+ − 318 <param name="runMe" value="$runMe"/>
41
+ − 319 <output name="tab_file" file="%(test1Output)s" ftype="%(outputFormat)s" />
37
+ − 320 <output name="html_file" file="%(test1HTML)s" ftype="html" lines_diff="10"/>
+ − 321 </test>
+ − 322 </tests>
+ − 323 """
+ − 324 xdict = {}
41
+ − 325 xdict['outputFormat'] = self.outputFormat
+ − 326 xdict['inputFormats'] = self.inputFormats
37
+ − 327 xdict['requirements'] = ''
+ − 328 if self.opts.make_HTML:
+ − 329 if self.opts.include_dependencies == "yes":
+ − 330 xdict['requirements'] = protorequirements
+ − 331 xdict['tool_version'] = self.opts.tool_version
+ − 332 xdict['test1Input'] = self.test1Input
+ − 333 xdict['test1HTML'] = self.test1HTML
+ − 334 xdict['test1Output'] = self.test1Output
+ − 335 if self.opts.make_HTML and self.opts.output_tab <> 'None':
+ − 336 xdict['tooltests'] = tooltestsBoth % xdict
+ − 337 elif self.opts.make_HTML:
+ − 338 xdict['tooltests'] = tooltestsHTMLOnly % xdict
+ − 339 else:
+ − 340 xdict['tooltests'] = tooltestsTabOnly % xdict
+ − 341 xdict['script'] = self.escapedScript
+ − 342 # configfile is least painful way to embed script to avoid external dependencies
+ − 343 # but requires escaping of <, > and $ to avoid Mako parsing
+ − 344 if self.opts.help_text:
+ − 345 helptext = open(self.opts.help_text,'r').readlines()
+ − 346 helptext = [html_escape(x) for x in helptext] # must html escape here too - thanks to Marius van den Beek
+ − 347 xdict['help'] = ''.join([x for x in helptext])
+ − 348 else:
+ − 349 xdict['help'] = 'Please ask the tool author (%s) for help as none was supplied at tool generation\n' % (self.opts.user_email)
+ − 350 if self.opts.citations:
+ − 351 citationstext = open(self.opts.citations,'r').read()
+ − 352 citation_tuples = parse_citations(citationstext)
+ − 353 citations_xml = ""
+ − 354 for citation_type, citation_content in citation_tuples:
+ − 355 citation_xml = """<citation type="%s">%s</citation>""" % (citation_type, html_escape(citation_content))
+ − 356 citations_xml += citation_xml
+ − 357 xdict['citations'] = citations_xml
+ − 358 else:
+ − 359 xdict['citations'] = ""
+ − 360 coda = ['**Script**','Pressing execute will run the following code over your input file and generate some outputs in your history::']
+ − 361 coda.append('\n')
+ − 362 coda.append(self.indentedScript)
+ − 363 coda.append('\n**Attribution**\nThis Galaxy tool was created by %s at %s\nusing the Galaxy Tool Factory.\n' % (self.opts.user_email,timenow()))
+ − 364 coda.append('See %s for details of that project' % (toolFactoryURL))
+ − 365 coda.append('Please cite: Creating re-usable tools from scripts: The Galaxy Tool Factory. Ross Lazarus; Antony Kaspi; Mark Ziemann; The Galaxy Team. ')
+ − 366 coda.append('Bioinformatics 2012; doi: 10.1093/bioinformatics/bts573\n')
+ − 367 xdict['help'] = '%s\n%s' % (xdict['help'],'\n'.join(coda))
+ − 368 if self.opts.tool_desc:
+ − 369 xdict['tooldesc'] = '<description>%s</description>' % self.opts.tool_desc
+ − 370 else:
+ − 371 xdict['tooldesc'] = ''
+ − 372 xdict['command_outputs'] = ''
+ − 373 xdict['outputs'] = ''
+ − 374 if self.opts.input_tab <> 'None':
+ − 375 xdict['command_inputs'] = '--input_tab "$input1" ' # the space may matter a lot if we append something
+ − 376 xdict['inputs'] = '<param name="input1" type="data" format="%s" label="Select a suitable input file from your history"/> \n' % self.inputFormats
+ − 377 else:
+ − 378 xdict['command_inputs'] = '' # assume no input - eg a random data generator
+ − 379 xdict['inputs'] = ''
+ − 380 xdict['inputs'] += '<param name="job_name" type="text" label="Supply a name for the outputs to remind you what they contain" value="%s"/> \n' % self.toolname
+ − 381 xdict['toolname'] = self.toolname
+ − 382 xdict['toolid'] = self.toolid
+ − 383 xdict['interpreter'] = self.opts.interpreter
+ − 384 xdict['scriptname'] = self.sfile
+ − 385 if self.opts.make_HTML:
+ − 386 xdict['command_outputs'] += ' --output_dir "$html_file.files_path" --output_html "$html_file" --make_HTML "yes"'
+ − 387 xdict['outputs'] += ' <data format="html" name="html_file" label="${job_name}.html"/>\n'
+ − 388 else:
+ − 389 xdict['command_outputs'] += ' --output_dir "./"'
+ − 390 if self.opts.output_tab <> 'None':
+ − 391 xdict['command_outputs'] += ' --output_tab "$tab_file"'
41
+ − 392 xdict['outputs'] += ' <data format="%s" name="tab_file" label="${job_name}"/>\n' % self.outputFormat
37
+ − 393 xdict['command'] = newCommand % xdict
+ − 394 xmls = newXML % xdict
+ − 395 xf = open(self.xmlfile,'w')
+ − 396 xf.write(xmls)
+ − 397 xf.write('\n')
+ − 398 xf.close()
+ − 399 # ready for the tarball
+ − 400
+ − 401
+ − 402 def makeTooltar(self):
+ − 403 """
+ − 404 a tool is a gz tarball with eg
+ − 405 /toolname/tool.xml /toolname/tool.py /toolname/test-data/test1_in.foo ...
+ − 406 """
+ − 407 retval = self.run()
+ − 408 if retval:
+ − 409 print >> sys.stderr,'## Run failed. Cannot build yet. Please fix and retry'
+ − 410 sys.exit(1)
+ − 411 tdir = self.toolname
+ − 412 os.mkdir(tdir)
+ − 413 self.makeXML()
+ − 414 if self.opts.make_HTML:
+ − 415 if self.opts.help_text:
+ − 416 hlp = open(self.opts.help_text,'r').read()
+ − 417 else:
+ − 418 hlp = 'Please ask the tool author for help as none was supplied at tool generation\n'
+ − 419 if self.opts.include_dependencies:
+ − 420 tooldepcontent = toolhtmldepskel % hlp
+ − 421 depf = open(os.path.join(tdir,'tool_dependencies.xml'),'w')
+ − 422 depf.write(tooldepcontent)
+ − 423 depf.write('\n')
+ − 424 depf.close()
+ − 425 if self.opts.input_tab <> 'None': # no reproducible test otherwise? TODO: maybe..
+ − 426 testdir = os.path.join(tdir,'test-data')
+ − 427 os.mkdir(testdir) # make tests directory
+ − 428 shutil.copyfile(self.opts.input_tab,os.path.join(testdir,self.test1Input))
+ − 429 if self.opts.output_tab <> 'None':
+ − 430 shutil.copyfile(self.opts.output_tab,os.path.join(testdir,self.test1Output))
+ − 431 if self.opts.make_HTML:
+ − 432 shutil.copyfile(self.opts.output_html,os.path.join(testdir,self.test1HTML))
+ − 433 if self.opts.output_dir:
+ − 434 shutil.copyfile(self.tlog,os.path.join(testdir,'test1_out.log'))
+ − 435 outpif = '%s.py' % self.toolname # new name
+ − 436 outpiname = os.path.join(tdir,outpif) # path for the tool tarball
+ − 437 pyin = os.path.basename(self.pyfile) # our name - we rewrite ourselves (TM)
+ − 438 notes = ['# %s - a self annotated version of %s generated by running %s\n' % (outpiname,pyin,pyin),]
+ − 439 notes.append('# to make a new Galaxy tool called %s\n' % self.toolname)
+ − 440 notes.append('# User %s at %s\n' % (self.opts.user_email,timenow()))
+ − 441 pi = open(self.pyfile,'r').readlines() # our code becomes new tool wrapper (!) - first Galaxy worm
+ − 442 notes += pi
+ − 443 outpi = open(outpiname,'w')
+ − 444 outpi.write(''.join(notes))
+ − 445 outpi.write('\n')
+ − 446 outpi.close()
+ − 447 stname = os.path.join(tdir,self.sfile)
+ − 448 if not os.path.exists(stname):
+ − 449 shutil.copyfile(self.sfile, stname)
+ − 450 xtname = os.path.join(tdir,self.xmlfile)
+ − 451 if not os.path.exists(xtname):
+ − 452 shutil.copyfile(self.xmlfile,xtname)
+ − 453 tarpath = "%s.gz" % self.toolname
+ − 454 tar = tarfile.open(tarpath, "w:gz")
+ − 455 tar.add(tdir,arcname=self.toolname)
+ − 456 tar.close()
+ − 457 shutil.copyfile(tarpath,self.opts.new_tool)
+ − 458 shutil.rmtree(tdir)
+ − 459 ## TODO: replace with optional direct upload to local toolshed?
+ − 460 return retval
+ − 461
+ − 462
+ − 463 def compressPDF(self,inpdf=None,thumbformat='png'):
+ − 464 """need absolute path to pdf
+ − 465 note that GS gets confoozled if no $TMP or $TEMP
+ − 466 so we set it
+ − 467 """
+ − 468 assert os.path.isfile(inpdf), "## Input %s supplied to %s compressPDF not found" % (inpdf,self.myName)
+ − 469 hlog = os.path.join(self.opts.output_dir,"compress_%s.txt" % os.path.basename(inpdf))
+ − 470 sto = open(hlog,'a')
+ − 471 our_env = os.environ.copy()
+ − 472 our_tmp = our_env.get('TMP',None)
+ − 473 if not our_tmp:
+ − 474 our_tmp = our_env.get('TEMP',None)
+ − 475 if not (our_tmp and os.path.exists(our_tmp)):
+ − 476 newtmp = os.path.join(self.opts.output_dir,'tmp')
+ − 477 try:
+ − 478 os.mkdir(newtmp)
+ − 479 except:
+ − 480 sto.write('## WARNING - cannot make %s - it may exist or permissions need fixing\n' % newtmp)
+ − 481 our_env['TEMP'] = newtmp
+ − 482 if not self.temp_warned:
+ − 483 sto.write('## WARNING - no $TMP or $TEMP!!! Please fix - using %s temporarily\n' % newtmp)
+ − 484 self.temp_warned = True
+ − 485 outpdf = '%s_compressed' % inpdf
+ − 486 cl = ["gs", "-sDEVICE=pdfwrite", "-dNOPAUSE", "-dUseCIEColor", "-dBATCH","-dPDFSETTINGS=/printer", "-sOutputFile=%s" % outpdf,inpdf]
+ − 487 x = subprocess.Popen(cl,stdout=sto,stderr=sto,cwd=self.opts.output_dir,env=our_env)
+ − 488 retval1 = x.wait()
+ − 489 sto.close()
+ − 490 if retval1 == 0:
+ − 491 os.unlink(inpdf)
+ − 492 shutil.move(outpdf,inpdf)
+ − 493 os.unlink(hlog)
+ − 494 hlog = os.path.join(self.opts.output_dir,"thumbnail_%s.txt" % os.path.basename(inpdf))
+ − 495 sto = open(hlog,'w')
+ − 496 outpng = '%s.%s' % (os.path.splitext(inpdf)[0],thumbformat)
+ − 497 if self.useGM:
+ − 498 cl2 = ['gm', 'convert', inpdf, outpng]
+ − 499 else: # assume imagemagick
+ − 500 cl2 = ['convert', inpdf, outpng]
+ − 501 x = subprocess.Popen(cl2,stdout=sto,stderr=sto,cwd=self.opts.output_dir,env=our_env)
+ − 502 retval2 = x.wait()
+ − 503 sto.close()
+ − 504 if retval2 == 0:
+ − 505 os.unlink(hlog)
+ − 506 retval = retval1 or retval2
+ − 507 return retval
+ − 508
+ − 509
+ − 510 def getfSize(self,fpath,outpath):
+ − 511 """
+ − 512 format a nice file size string
+ − 513 """
+ − 514 size = ''
+ − 515 fp = os.path.join(outpath,fpath)
+ − 516 if os.path.isfile(fp):
+ − 517 size = '0 B'
+ − 518 n = float(os.path.getsize(fp))
+ − 519 if n > 2**20:
+ − 520 size = '%1.1f MB' % (n/2**20)
+ − 521 elif n > 2**10:
+ − 522 size = '%1.1f KB' % (n/2**10)
+ − 523 elif n > 0:
+ − 524 size = '%d B' % (int(n))
+ − 525 return size
+ − 526
+ − 527 def makeHtml(self):
+ − 528 """ Create an HTML file content to list all the artifacts found in the output_dir
+ − 529 """
+ − 530
+ − 531 galhtmlprefix = """<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
+ − 532 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
+ − 533 <head> <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
+ − 534 <meta name="generator" content="Galaxy %s tool output - see http://g2.trac.bx.psu.edu/" />
+ − 535 <title></title>
+ − 536 <link rel="stylesheet" href="/static/style/base.css" type="text/css" />
+ − 537 </head>
+ − 538 <body>
+ − 539 <div class="toolFormBody">
+ − 540 """
+ − 541 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/>"""
+ − 542 galhtmlpostfix = """</div></body></html>\n"""
+ − 543
+ − 544 flist = os.listdir(self.opts.output_dir)
+ − 545 flist = [x for x in flist if x <> 'Rplots.pdf']
+ − 546 flist.sort()
+ − 547 html = []
+ − 548 html.append(galhtmlprefix % progname)
+ − 549 html.append('<div class="infomessage">Galaxy Tool "%s" run at %s</div><br/>' % (self.toolname,timenow()))
+ − 550 fhtml = []
+ − 551 if len(flist) > 0:
+ − 552 logfiles = [x for x in flist if x.lower().endswith('.log')] # log file names determine sections
+ − 553 logfiles.sort()
+ − 554 logfiles = [x for x in logfiles if os.path.abspath(x) <> os.path.abspath(self.tlog)]
+ − 555 logfiles.append(os.path.abspath(self.tlog)) # make it the last one
+ − 556 pdflist = []
+ − 557 npdf = len([x for x in flist if os.path.splitext(x)[-1].lower() == '.pdf'])
+ − 558 for rownum,fname in enumerate(flist):
+ − 559 dname,e = os.path.splitext(fname)
+ − 560 sfsize = self.getfSize(fname,self.opts.output_dir)
+ − 561 if e.lower() == '.pdf' : # compress and make a thumbnail
+ − 562 thumb = '%s.%s' % (dname,self.thumbformat)
+ − 563 pdff = os.path.join(self.opts.output_dir,fname)
+ − 564 retval = self.compressPDF(inpdf=pdff,thumbformat=self.thumbformat)
+ − 565 if retval == 0:
+ − 566 pdflist.append((fname,thumb))
+ − 567 else:
+ − 568 pdflist.append((fname,fname))
+ − 569 if (rownum+1) % 2 == 0:
+ − 570 fhtml.append('<tr class="odd_row"><td><a href="%s">%s</a></td><td>%s</td></tr>' % (fname,fname,sfsize))
+ − 571 else:
+ − 572 fhtml.append('<tr><td><a href="%s">%s</a></td><td>%s</td></tr>' % (fname,fname,sfsize))
+ − 573 for logfname in logfiles: # expect at least tlog - if more
+ − 574 if os.path.abspath(logfname) == os.path.abspath(self.tlog): # handled later
+ − 575 sectionname = 'All tool run'
+ − 576 if (len(logfiles) > 1):
+ − 577 sectionname = 'Other'
+ − 578 ourpdfs = pdflist
+ − 579 else:
+ − 580 realname = os.path.basename(logfname)
+ − 581 sectionname = os.path.splitext(realname)[0].split('_')[0] # break in case _ added to log
+ − 582 ourpdfs = [x for x in pdflist if os.path.basename(x[0]).split('_')[0] == sectionname]
+ − 583 pdflist = [x for x in pdflist if os.path.basename(x[0]).split('_')[0] <> sectionname] # remove
+ − 584 nacross = 1
+ − 585 npdf = len(ourpdfs)
+ − 586
+ − 587 if npdf > 0:
+ − 588 nacross = math.sqrt(npdf) ## int(round(math.log(npdf,2)))
+ − 589 if int(nacross)**2 != npdf:
+ − 590 nacross += 1
+ − 591 nacross = int(nacross)
+ − 592 width = min(400,int(1200/nacross))
+ − 593 html.append('<div class="toolFormTitle">%s images and outputs</div>' % sectionname)
+ − 594 html.append('(Click on a thumbnail image to download the corresponding original PDF image)<br/>')
+ − 595 ntogo = nacross # counter for table row padding with empty cells
+ − 596 html.append('<div><table class="simple" cellpadding="2" cellspacing="2">\n<tr>')
+ − 597 for i,paths in enumerate(ourpdfs):
+ − 598 fname,thumb = paths
+ − 599 s= """<td><a href="%s"><img src="%s" title="Click to download a PDF of %s" hspace="5" width="%d"
+ − 600 alt="Image called %s"/></a></td>\n""" % (fname,thumb,fname,width,fname)
+ − 601 if ((i+1) % nacross == 0):
+ − 602 s += '</tr>\n'
+ − 603 ntogo = 0
+ − 604 if i < (npdf - 1): # more to come
+ − 605 s += '<tr>'
+ − 606 ntogo = nacross
+ − 607 else:
+ − 608 ntogo -= 1
+ − 609 html.append(s)
+ − 610 if html[-1].strip().endswith('</tr>'):
+ − 611 html.append('</table></div>\n')
+ − 612 else:
+ − 613 if ntogo > 0: # pad
+ − 614 html.append('<td> </td>'*ntogo)
+ − 615 html.append('</tr></table></div>\n')
+ − 616 logt = open(logfname,'r').readlines()
+ − 617 logtext = [x for x in logt if x.strip() > '']
+ − 618 html.append('<div class="toolFormTitle">%s log output</div>' % sectionname)
+ − 619 if len(logtext) > 1:
+ − 620 html.append('\n<pre>\n')
+ − 621 html += logtext
+ − 622 html.append('\n</pre>\n')
+ − 623 else:
+ − 624 html.append('%s is empty<br/>' % logfname)
+ − 625 if len(fhtml) > 0:
+ − 626 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')
+ − 627 fhtml.append('</table></div><br/>')
+ − 628 html.append('<div class="toolFormTitle">All output files available for downloading</div>\n')
+ − 629 html += fhtml # add all non-pdf files to the end of the display
+ − 630 else:
+ − 631 html.append('<div class="warningmessagelarge">### Error - %s returned no files - please confirm that parameters are sane</div>' % self.opts.interpreter)
+ − 632 html.append(galhtmlpostfix)
+ − 633 htmlf = file(self.opts.output_html,'w')
+ − 634 htmlf.write('\n'.join(html))
+ − 635 htmlf.write('\n')
+ − 636 htmlf.close()
+ − 637 self.html = html
+ − 638
+ − 639
+ − 640 def run(self):
+ − 641 """
+ − 642 scripts must be small enough not to fill the pipe!
+ − 643 """
+ − 644 if self.treatbashSpecial and self.opts.interpreter in ['bash','sh']:
+ − 645 retval = self.runBash()
+ − 646 else:
+ − 647 if self.opts.output_dir:
45
+ − 648 ste = open(self.elog,'a')
+ − 649 sto = open(self.tlog,'a')
37
+ − 650 sto.write('## Toolfactory generated command line = %s\n' % ' '.join(self.cl))
+ − 651 sto.flush()
+ − 652 p = subprocess.Popen(self.cl,shell=False,stdout=sto,stderr=ste,stdin=subprocess.PIPE,cwd=self.opts.output_dir)
+ − 653 else:
+ − 654 p = subprocess.Popen(self.cl,shell=False,stdin=subprocess.PIPE)
+ − 655 p.stdin.write(self.script)
+ − 656 p.stdin.close()
+ − 657 retval = p.wait()
+ − 658 if self.opts.output_dir:
+ − 659 sto.close()
+ − 660 ste.close()
+ − 661 err = open(self.elog,'r').readlines()
+ − 662 if retval <> 0 and err: # problem
+ − 663 print >> sys.stderr,err
+ − 664 if self.opts.make_HTML:
+ − 665 self.makeHtml()
+ − 666 return retval
+ − 667
+ − 668 def runBash(self):
+ − 669 """
+ − 670 cannot use - for bash so use self.sfile
+ − 671 """
+ − 672 if self.opts.output_dir:
+ − 673 s = '## Toolfactory generated command line = %s\n' % ' '.join(self.cl)
+ − 674 sto = open(self.tlog,'w')
+ − 675 sto.write(s)
+ − 676 sto.flush()
+ − 677 p = subprocess.Popen(self.cl,shell=False,stdout=sto,stderr=sto,cwd=self.opts.output_dir)
+ − 678 else:
+ − 679 p = subprocess.Popen(self.cl,shell=False)
+ − 680 retval = p.wait()
+ − 681 if self.opts.output_dir:
+ − 682 sto.close()
+ − 683 if self.opts.make_HTML:
+ − 684 self.makeHtml()
+ − 685 return retval
+ − 686
+ − 687
+ − 688 def main():
+ − 689 u = """
+ − 690 This is a Galaxy wrapper. It expects to be called by a special purpose tool.xml as:
+ − 691 <command interpreter="python">rgBaseScriptWrapper.py --script_path "$scriptPath" --tool_name "foo" --interpreter "Rscript"
+ − 692 </command>
+ − 693 """
+ − 694 op = optparse.OptionParser()
+ − 695 a = op.add_option
+ − 696 a('--script_path',default=None)
+ − 697 a('--tool_name',default=None)
+ − 698 a('--interpreter',default=None)
+ − 699 a('--output_dir',default='./')
+ − 700 a('--output_html',default=None)
+ − 701 a('--input_tab',default="None")
41
+ − 702 a('--input_formats',default="tabular,text")
37
+ − 703 a('--output_tab',default="None")
41
+ − 704 a('--output_format',default="tabular")
37
+ − 705 a('--user_email',default='Unknown')
+ − 706 a('--bad_user',default=None)
+ − 707 a('--make_Tool',default=None)
+ − 708 a('--make_HTML',default=None)
+ − 709 a('--help_text',default=None)
+ − 710 a('--citations',default=None)
+ − 711 a('--tool_desc',default=None)
+ − 712 a('--new_tool',default=None)
+ − 713 a('--tool_version',default=None)
+ − 714 a('--include_dependencies',default=None)
+ − 715 opts, args = op.parse_args()
+ − 716 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)
+ − 717 assert opts.tool_name,'## Tool Factory expects a tool name - eg --tool_name=DESeq'
+ − 718 assert opts.interpreter,'## Tool Factory wrapper expects an interpreter - eg --interpreter=Rscript'
+ − 719 assert os.path.isfile(opts.script_path),'## Tool Factory wrapper expects a script path - eg --script_path=foo.R'
+ − 720 if opts.output_dir:
+ − 721 try:
+ − 722 os.makedirs(opts.output_dir)
+ − 723 except:
+ − 724 pass
+ − 725 r = ScriptRunner(opts)
+ − 726 if opts.make_Tool:
+ − 727 retcode = r.makeTooltar()
+ − 728 else:
+ − 729 retcode = r.run()
+ − 730 os.unlink(r.sfile)
+ − 731 if retcode:
+ − 732 sys.exit(retcode) # indicate failure to job runner
+ − 733
+ − 734
+ − 735 if __name__ == "__main__":
+ − 736 main()
+ − 737
+ − 738