Mercurial > repos > fubar > tool_factory_2
comparison rgToolFactory2.py @ 0:c34063ab3735 draft
Initial commit of code in iuc github repository
author | fubar |
---|---|
date | Thu, 01 Jan 2015 21:58:00 -0500 |
parents | |
children | 6a3c292412fa |
comparison
equal
deleted
inserted
replaced
-1:000000000000 | 0:c34063ab3735 |
---|---|
1 # rgToolFactoryMultIn.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 # sept 2014 added additional params from | |
11 # https://bitbucket.org/mvdbeek/dockertoolfactory/src/d4863bcf7b521532c7e8c61b6333840ba5393f73/DockerToolFactory.py?at=default | |
12 # passing them is complex | |
13 # and they are restricted to NOT contain commas or double quotes to ensure that they can be safely passed together on | |
14 # the toolfactory command line as a comma delimited double quoted string for parsing and passing to the script | |
15 # see examples on this tool form | |
16 | |
17 # august 2014 | |
18 | |
19 # Allows arbitrary number of input files | |
20 # NOTE positional parameters are now passed to script | |
21 # and output (may be "None") is *before* arbitrary number of inputs | |
22 # | |
23 # march 2014 | |
24 # had to remove dependencies because cross toolshed dependencies are not possible - can't pre-specify a toolshed url for graphicsmagick and ghostscript | |
25 # grrrrr - night before a demo | |
26 # added dependencies to a tool_dependencies.xml if html page generated so generated tool is properly portable | |
27 # | |
28 # added ghostscript and graphicsmagick as dependencies | |
29 # fixed a wierd problem where gs was trying to use the new_files_path from universe (database/tmp) as ./database/tmp | |
30 # errors ensued | |
31 # | |
32 # august 2013 | |
33 # found a problem with GS if $TMP or $TEMP missing - now inject /tmp and warn | |
34 # | |
35 # july 2013 | |
36 # added ability to combine images and individual log files into html output | |
37 # just make sure there's a log file foo.log and it will be output | |
38 # together with all images named like "foo_*.pdf | |
39 # otherwise old format for html | |
40 # | |
41 # January 2013 | |
42 # problem pointed out by Carlos Borroto | |
43 # added escaping for <>$ - thought I did that ages ago... | |
44 # | |
45 # August 11 2012 | |
46 # changed to use shell=False and cl as a sequence | |
47 | |
48 # This is a Galaxy tool factory for simple scripts in python, R or whatever ails ye. | |
49 # It also serves as the wrapper for the new tool. | |
50 # | |
51 # you paste and run your script | |
52 # Only works for simple scripts that read one input from the history. | |
53 # Optionally can write one new history dataset, | |
54 # and optionally collect any number of outputs into links on an autogenerated HTML page. | |
55 | |
56 # DO NOT install on a public or important site - please. | |
57 | |
58 # installed generated tools are fine if the script is safe. | |
59 # They just run normally and their user cannot do anything unusually insecure | |
60 # but please, practice safe toolshed. | |
61 # Read the fucking code before you install any tool | |
62 # especially this one | |
63 | |
64 # After you get the script working on some test data, you can | |
65 # optionally generate a toolshed compatible gzip file | |
66 # containing your script safely wrapped as an ordinary Galaxy script in your local toolshed for | |
67 # safe and largely automated installation in a production Galaxy. | |
68 | |
69 # If you opt for an HTML output, you get all the script outputs arranged | |
70 # as a single Html history item - all output files are linked, thumbnails for all the pdfs. | |
71 # Ugly but really inexpensive. | |
72 # | |
73 # Patches appreciated please. | |
74 # | |
75 # | |
76 # long route to June 2012 product | |
77 # Behold the awesome power of Galaxy and the toolshed with the tool factory to bind them | |
78 # derived from an integrated script model | |
79 # called rgBaseScriptWrapper.py | |
80 # Note to the unwary: | |
81 # This tool allows arbitrary scripting on your Galaxy as the Galaxy user | |
82 # There is nothing stopping a malicious user doing whatever they choose | |
83 # Extremely dangerous!! | |
84 # Totally insecure. So, trusted users only | |
85 # | |
86 # preferred model is a developer using their throw away workstation instance - ie a private site. | |
87 # no real risk. The universe_wsgi.ini admin_users string is checked - only admin users are permitted to run this tool. | |
88 # | |
89 | |
90 import sys | |
91 import shutil | |
92 import subprocess | |
93 import os | |
94 import time | |
95 import tempfile | |
96 import optparse | |
97 import tarfile | |
98 import re | |
99 import shutil | |
100 import math | |
101 | |
102 progname = os.path.split(sys.argv[0])[1] | |
103 myversion = 'V001.1 March 2014' | |
104 verbose = False | |
105 debug = False | |
106 toolFactoryURL = 'https://bitbucket.org/fubar/galaxytoolfactory' | |
107 | |
108 # if we do html we need these dependencies specified in a tool_dependencies.xml file and referred to in the generated | |
109 # tool xml | |
110 toolhtmldepskel = """<?xml version="1.0"?> | |
111 <tool_dependency> | |
112 <package name="ghostscript" version="9.10"> | |
113 <repository name="package_ghostscript_9_10" owner="devteam" prior_installation_required="True" /> | |
114 </package> | |
115 <package name="graphicsmagick" version="1.3.18"> | |
116 <repository name="package_graphicsmagick_1_3" owner="iuc" prior_installation_required="True" /> | |
117 </package> | |
118 <readme> | |
119 %s | |
120 </readme> | |
121 </tool_dependency> | |
122 """ | |
123 | |
124 toolhtmldepskel = """<?xml version="1.0"?> | |
125 <tool_dependency> | |
126 <readme> | |
127 %s | |
128 </readme> | |
129 </tool_dependency> | |
130 """ | |
131 | |
132 protorequirements = """<requirements> | |
133 <requirement type="package" version="9.10">ghostscript</requirement> | |
134 <requirement type="package" version="1.3.18">graphicsmagick</requirement> | |
135 </requirements>""" | |
136 | |
137 def timenow(): | |
138 """return current time as a string | |
139 """ | |
140 return time.strftime('%d/%m/%Y %H:%M:%S', time.localtime(time.time())) | |
141 | |
142 html_escape_table = { | |
143 "&": "&", | |
144 ">": ">", | |
145 "<": "<", | |
146 "$": "\$" | |
147 } | |
148 | |
149 def html_escape(text): | |
150 """Produce entities within text.""" | |
151 return "".join(html_escape_table.get(c,c) for c in text) | |
152 | |
153 def cmd_exists(cmd): | |
154 return subprocess.call("type " + cmd, shell=True, | |
155 stdout=subprocess.PIPE, stderr=subprocess.PIPE) == 0 | |
156 | |
157 def parse_citations(citations_text): | |
158 """ | |
159 """ | |
160 citations = [c for c in citations_text.split("**ENTRY**") if c.strip()] | |
161 citation_tuples = [] | |
162 for citation in citations: | |
163 if citation.startswith("doi"): | |
164 citation_tuples.append( ("doi", citation[len("doi"):].strip() ) ) | |
165 else: | |
166 citation_tuples.append( ("bibtex", citation[len("bibtex"):].strip() ) ) | |
167 return citation_tuples | |
168 | |
169 | |
170 class ScriptRunner: | |
171 """class is a wrapper for an arbitrary script | |
172 """ | |
173 | |
174 def __init__(self,opts=None,treatbashSpecial=True): | |
175 """ | |
176 cleanup inputs, setup some outputs | |
177 | |
178 """ | |
179 self.useGM = cmd_exists('gm') | |
180 self.useIM = cmd_exists('convert') | |
181 self.useGS = cmd_exists('gs') | |
182 self.temp_warned = False # we want only one warning if $TMP not set | |
183 self.treatbashSpecial = treatbashSpecial | |
184 if opts.output_dir: # simplify for the tool tarball | |
185 os.chdir(opts.output_dir) | |
186 self.thumbformat = 'png' | |
187 self.opts = opts | |
188 self.toolname = re.sub('[^a-zA-Z0-9_]+', '', opts.tool_name) # a sanitizer now does this but.. | |
189 self.toolid = self.toolname | |
190 self.myname = sys.argv[0] # get our name because we write ourselves out as a tool later | |
191 self.pyfile = self.myname # crude but efficient - the cruft won't hurt much | |
192 self.xmlfile = '%s.xml' % self.toolname | |
193 s = open(self.opts.script_path,'r').readlines() | |
194 s = [x.rstrip() for x in s] # remove pesky dos line endings if needed | |
195 self.script = '\n'.join(s) | |
196 fhandle,self.sfile = tempfile.mkstemp(prefix=self.toolname,suffix=".%s" % (opts.interpreter)) | |
197 tscript = open(self.sfile,'w') # use self.sfile as script source for Popen | |
198 tscript.write(self.script) | |
199 tscript.close() | |
200 self.indentedScript = '\n'.join([' %s' % html_escape(x) for x in s]) # for restructured text in help | |
201 self.escapedScript = '\n'.join([html_escape(x) for x in s]) | |
202 self.elog = os.path.join(self.opts.output_dir,"%s_error.log" % self.toolname) | |
203 if opts.output_dir: # may not want these complexities | |
204 self.tlog = os.path.join(self.opts.output_dir,"%s_runner.log" % self.toolname) | |
205 art = '%s.%s' % (self.toolname,opts.interpreter) | |
206 artpath = os.path.join(self.opts.output_dir,art) # need full path | |
207 artifact = open(artpath,'w') # use self.sfile as script source for Popen | |
208 artifact.write(self.script) | |
209 artifact.close() | |
210 self.cl = [] | |
211 self.html = [] | |
212 self.test1Inputs = [] # now a list | |
213 a = self.cl.append | |
214 a(opts.interpreter) | |
215 if self.treatbashSpecial and opts.interpreter in ['bash','sh']: | |
216 a(self.sfile) | |
217 else: | |
218 a('-') # stdin | |
219 # if multiple inputs - positional or need to distinguish them with cl params | |
220 if opts.input_tab: | |
221 tests = [] | |
222 for i,intab in enumerate(opts.input_tab): # if multiple, make tests | |
223 if intab.find(',') <> -1: | |
224 (gpath,uname) = intab.split(',') | |
225 else: | |
226 gpath = uname = intab | |
227 tests.append(os.path.basename(gpath)) | |
228 self.test1Inputs = '<param name="input_tab" value="%s" />' % (','.join(tests)) | |
229 else: | |
230 self.test1Inputs = '' | |
231 # we always pass path,name pairs in using python optparse append | |
232 # but the command line has to be different | |
233 self.infile_paths = ','.join([x.split(',')[0] for x in self.opts.input_tab]) | |
234 self.infile_names = ','.join([x.split(',')[1] for x in self.opts.input_tab]) | |
235 if self.opts.interpreter == 'python': | |
236 # yes, this is how additional parameters are always passed in python - to the TF itself and to | |
237 # scripts to avoid having unknown parameter names (yes, they can be parsed but...) on the command line | |
238 a('--INPATHS "%s"' % (self.infile_paths)) | |
239 a('--INNAMES "%s"' % (self.infile_names)) | |
240 if self.opts.output_tab: | |
241 a('--OUTPATH "%s"' % self.opts.output_tab) | |
242 for p in opts.additional_parameters: | |
243 p = p.replace('"','') | |
244 psplit=p.split(',') | |
245 param = psplit[0] | |
246 value = psplit[1] | |
247 a('--additional_parameters "%s,%s"' % (param,value)) | |
248 if (self.opts.interpreter == 'Rscript'): | |
249 # pass params on command line | |
250 a('INPATHS "%s"' % self.infile_paths) | |
251 a('INNAMES "%s"' % self.infile_names) | |
252 if self.opts.output_tab: | |
253 a('OUTPATH "%s"' % self.opts.output_tab) | |
254 for param in opts.additional_parameters: | |
255 param, value=param.split(',') | |
256 a('%s="%s"' % (param,value)) | |
257 if (self.opts.interpreter == 'perl'): | |
258 # pass params on command line | |
259 a('%s' % self.infile_paths) | |
260 a('%s' % self.infile_names) | |
261 if self.opts.output_tab: | |
262 a('%s' % self.opts.output_tab) | |
263 for param in opts.additional_parameters: | |
264 param, value=param.split(',') | |
265 if (value.find(' ') <> -1): | |
266 a('%s="%s"' % (param,value)) | |
267 else: | |
268 a('%s=%s' % (param,value)) | |
269 | |
270 if self.opts.interpreter == 'sh' or self.opts.interpreter == 'bash': | |
271 # more is better - now move all params into environment AND drop on to command line. | |
272 self.cl.insert(0,'env') | |
273 self.cl.insert(1,'INPATHS=%s' % (self.infile_paths)) | |
274 self.cl.insert(2,'INNAMES=%s' % (self.infile_names)) | |
275 if self.opts.output_tab: | |
276 self.cl.insert(3,'OUTPATH=%s' % (self.opts.output_tab)) | |
277 a('OUTPATH=%s' % (self.opts.output_tab)) | |
278 # sets those environment variables for the script | |
279 # additional params appear in CL - yes, it's confusing | |
280 for i,param in enumerate(opts.additional_parameters): | |
281 psplit = param.split(',') | |
282 n = psplit[0] | |
283 v = psplit[1] | |
284 if (v.find(' ') <> -1): | |
285 a('%s="%s"' % (n,v)) | |
286 self.cl.insert(4+i,'%s="%s"' % (n,v)) | |
287 else: | |
288 a('%s=%s' % (n,v)) | |
289 self.cl.insert(4+i,'%s=%s' % (n,v)) | |
290 | |
291 | |
292 self.outFormats = opts.output_format | |
293 self.inputFormats = opts.input_formats | |
294 self.test1Output = '%s_test1_output.xls' % self.toolname | |
295 self.test1HTML = '%s_test1_output.html' % self.toolname | |
296 | |
297 def makeXML(self): | |
298 """ | |
299 Create a Galaxy xml tool wrapper for the new script as a string to write out | |
300 fixme - use templating or something less fugly than this example of what we produce | |
301 | |
302 <tool id="reverse" name="reverse" version="0.01"> | |
303 <description>a tabular file</description> | |
304 <command interpreter="python"> | |
305 reverse.py --script_path "$runMe" --interpreter "python" | |
306 --tool_name "reverse" --input_tab "$input1" --output_tab "$tab_file" | |
307 </command> | |
308 <inputs> | |
309 <param name="input1" type="data" format="tabular" label="Select one or more input files from your history"/> | |
310 <param name="job_name" type="text" label="Supply a name for the outputs to remind you what they contain" value="reverse"/> | |
311 </inputs> | |
312 <outputs> | |
313 <data format="tabular" name="tab_file" label="${job_name}"/> | |
314 | |
315 </outputs> | |
316 <help> | |
317 | |
318 **What it Does** | |
319 | |
320 Reverse the columns in a tabular file | |
321 | |
322 </help> | |
323 <configfiles> | |
324 <configfile name="runMe"> | |
325 | |
326 # reverse order of columns in a tabular file | |
327 import sys | |
328 inp = sys.argv[1] | |
329 outp = sys.argv[2] | |
330 i = open(inp,'r') | |
331 o = open(outp,'w') | |
332 for row in i: | |
333 rs = row.rstrip().split('\t') | |
334 rs.reverse() | |
335 o.write('\t'.join(rs)) | |
336 o.write('\n') | |
337 i.close() | |
338 o.close() | |
339 | |
340 | |
341 </configfile> | |
342 </configfiles> | |
343 </tool> | |
344 | |
345 """ | |
346 newXML="""<tool id="%(toolid)s" name="%(toolname)s" version="%(tool_version)s"> | |
347 %(tooldesc)s | |
348 %(requirements)s | |
349 <command interpreter="python"> | |
350 %(command)s | |
351 </command> | |
352 <inputs> | |
353 %(inputs)s | |
354 %(additionalInputs)s | |
355 </inputs> | |
356 <outputs> | |
357 %(outputs)s | |
358 </outputs> | |
359 <configfiles> | |
360 <configfile name="runMe"> | |
361 %(script)s | |
362 </configfile> | |
363 </configfiles> | |
364 <tests> | |
365 %(tooltests)s | |
366 </tests> | |
367 <help> | |
368 | |
369 %(help)s | |
370 | |
371 </help> | |
372 <citations> | |
373 %(citations)s | |
374 <citation type="doi">10.1093/bioinformatics/bts573</citation> | |
375 </citations> | |
376 </tool>""" # needs a dict with toolname, toolid, interpreter, scriptname, command, inputs as a multi line string ready to write, outputs ditto, help ditto | |
377 | |
378 newCommand=""" | |
379 %(toolname)s.py --script_path "$runMe" --interpreter "%(interpreter)s" | |
380 --tool_name "%(toolname)s" | |
381 %(command_inputs)s | |
382 %(command_outputs)s | |
383 """ | |
384 # may NOT be an input or htmlout - appended later | |
385 tooltestsTabOnly = """ | |
386 <test> | |
387 %(test1Inputs)s | |
388 <param name="job_name" value="test1"/> | |
389 <param name="runMe" value="$runMe"/> | |
390 <output name="output1="%(test1Output)s" ftype="tabular"/> | |
391 %(additionalParams)s | |
392 </test> | |
393 """ | |
394 tooltestsHTMLOnly = """ | |
395 <test> | |
396 %(test1Inputs)s | |
397 <param name="job_name" value="test1"/> | |
398 <param name="runMe" value="$runMe"/> | |
399 %(additionalParams)s | |
400 <output name="html_file" file="%(test1HTML)s" ftype="html" lines_diff="5"/> | |
401 </test> | |
402 """ | |
403 tooltestsBoth = """ | |
404 <test> | |
405 %(test1Inputs)s | |
406 <param name="job_name" value="test1"/> | |
407 <param name="runMe" value="$runMe"/> | |
408 %(additionalParams)s | |
409 <output name="output1" file="%(test1Output)s" ftype="tabular" /> | |
410 <output name="html_file" file="%(test1HTML)s" ftype="html" lines_diff="10"/> | |
411 </test> | |
412 """ | |
413 xdict = {} | |
414 xdict['additionalParams'] = '' | |
415 xdict['additionalInputs'] = '' | |
416 if self.opts.additional_parameters: | |
417 if self.opts.edit_additional_parameters: # add to new tool form with default value set to original value | |
418 xdict['additionalInputs'] = '\n'.join(['<param name="%s" value="%s" label="%s" help="%s" type="%s"/>' % (x.split(',')[0],x.split(',')[1],x.split(',')[2], | |
419 x.split(',')[3], x.split(',')[4]) for x in self.opts.additional_parameters]) | |
420 xdict['additionalParams'] = '\n'.join(['<param name="%s" value="%s" />' % (x.split(',')[0],x.split(',')[1]) for x in self.opts.additional_parameters]) | |
421 xdict['requirements'] = '' | |
422 if self.opts.make_HTML: | |
423 if self.opts.include_dependencies == "yes": | |
424 xdict['requirements'] = protorequirements | |
425 xdict['tool_version'] = self.opts.tool_version | |
426 xdict['test1HTML'] = self.test1HTML | |
427 xdict['test1Output'] = self.test1Output | |
428 xdict['test1Inputs'] = self.test1Inputs | |
429 if self.opts.make_HTML and self.opts.output_tab: | |
430 xdict['tooltests'] = tooltestsBoth % xdict | |
431 elif self.opts.make_HTML: | |
432 xdict['tooltests'] = tooltestsHTMLOnly % xdict | |
433 else: | |
434 xdict['tooltests'] = tooltestsTabOnly % xdict | |
435 xdict['script'] = self.escapedScript | |
436 # configfile is least painful way to embed script to avoid external dependencies | |
437 # but requires escaping of <, > and $ to avoid Mako parsing | |
438 if self.opts.help_text: | |
439 helptext = open(self.opts.help_text,'r').readlines() | |
440 helptext = [html_escape(x) for x in helptext] # must html escape here too - thanks to Marius van den Beek | |
441 xdict['help'] = ''.join([x for x in helptext]) | |
442 else: | |
443 xdict['help'] = 'Please ask the tool author (%s) for help as none was supplied at tool generation\n' % (self.opts.user_email) | |
444 coda = ['**Script**','Pressing execute will run the following code over your input file and generate some outputs in your history::'] | |
445 coda.append('\n') | |
446 coda.append(self.indentedScript) | |
447 coda.append('\n**Attribution**\nThis Galaxy tool was created by %s at %s\nusing the Galaxy Tool Factory.\n' % (self.opts.user_email,timenow())) | |
448 coda.append('See %s for details of that project' % (toolFactoryURL)) | |
449 coda.append('Please cite: Creating re-usable tools from scripts: The Galaxy Tool Factory. Ross Lazarus; Antony Kaspi; Mark Ziemann; The Galaxy Team. ') | |
450 coda.append('Bioinformatics 2012; doi: 10.1093/bioinformatics/bts573\n') | |
451 xdict['help'] = '%s\n%s' % (xdict['help'],'\n'.join(coda)) | |
452 if self.opts.tool_desc: | |
453 xdict['tooldesc'] = '<description>%s</description>' % self.opts.tool_desc | |
454 else: | |
455 xdict['tooldesc'] = '' | |
456 xdict['command_outputs'] = '' | |
457 xdict['outputs'] = '' | |
458 if self.opts.input_tab <> 'None': | |
459 cins = ['\n',] | |
460 cins.append('#for intab in $input1:') | |
461 cins.append('--input_tab "$intab,$intab.name"') | |
462 cins.append('#end for\n') | |
463 xdict['command_inputs'] = '\n'.join(cins) | |
464 xdict['inputs'] = '''<param name="input_tab" multiple="true" type="data" format="%s" label="Select one or more %s input files from your history" | |
465 help="Multiple inputs may be selected assuming the script can deal with them..."/> \n''' % (self.inputFormats,self.inputFormats) | |
466 else: | |
467 xdict['command_inputs'] = '' # assume no input - eg a random data generator | |
468 xdict['inputs'] = '' | |
469 if (len(self.opts.additional_parameters) > 0): | |
470 cins = ['\n',] | |
471 for params in self.opts.additional_parameters: | |
472 if self.opts.edit_additional_parameters: | |
473 psplit = params.split(',') # name,value... | |
474 psplit[1] = '$%s' % psplit[0] # replace with form value | |
475 cins.append('--additional_parameters "%s"' % ','.join(psplit)) | |
476 else: | |
477 cins.append('--additional_parameters "%s"' % params) | |
478 xdict['command_inputs'] = '%s\n%s' % (xdict['command_inputs'],'\n'.join(cins)) | |
479 xdict['inputs'] += '<param name="job_name" type="text" size="60" label="Supply a name for the outputs to remind you what they contain" value="%s"/> \n' % self.toolname | |
480 xdict['toolname'] = self.toolname | |
481 xdict['toolid'] = self.toolid | |
482 xdict['interpreter'] = self.opts.interpreter | |
483 xdict['scriptname'] = self.sfile | |
484 if self.opts.make_HTML: | |
485 xdict['command_outputs'] += ' --output_dir "$html_file.files_path" --output_html "$html_file" --make_HTML "yes"' | |
486 xdict['outputs'] += ' <data format="html" name="html_file" label="${job_name}.html"/>\n' | |
487 else: | |
488 xdict['command_outputs'] += ' --output_dir "./"' | |
489 if self.opts.output_tab: | |
490 xdict['command_outputs'] += ' --output_tab "$tab_file"' | |
491 xdict['outputs'] += ' <data format="%s" name="output1" label="${job_name}"/>\n' % self.outFormats | |
492 xdict['command'] = newCommand % xdict | |
493 if self.opts.citations: | |
494 citationstext = open(self.opts.citations,'r').read() | |
495 citation_tuples = parse_citations(citationstext) | |
496 citations_xml = "" | |
497 for citation_type, citation_content in citation_tuples: | |
498 citation_xml = """<citation type="%s">%s</citation>""" % (citation_type, html_escape(citation_content)) | |
499 citations_xml += citation_xml | |
500 xdict['citations'] = citations_xml | |
501 else: | |
502 xdict['citations'] = "" | |
503 xmls = newXML % xdict | |
504 xf = open(self.xmlfile,'w') | |
505 xf.write(xmls) | |
506 xf.write('\n') | |
507 xf.close() | |
508 # ready for the tarball | |
509 | |
510 | |
511 def makeTooltar(self): | |
512 """ | |
513 a tool is a gz tarball with eg | |
514 /toolname/tool.xml /toolname/tool.py /toolname/test-data/test1_in.foo ... | |
515 """ | |
516 retval = self.run() | |
517 if retval: | |
518 print >> sys.stderr,'## Run failed. Cannot build yet. Please fix and retry' | |
519 sys.exit(1) | |
520 tdir = self.toolname | |
521 os.mkdir(tdir) | |
522 self.makeXML() | |
523 if self.opts.make_HTML: | |
524 if self.opts.help_text: | |
525 hlp = open(self.opts.help_text,'r').read() | |
526 else: | |
527 hlp = 'Please ask the tool author for help as none was supplied at tool generation\n' | |
528 if self.opts.include_dependencies == "yes": | |
529 tooldepcontent = toolhtmldepskel % hlp | |
530 else: | |
531 tooldepcontent = emptytoolhtmldepskel % hlp | |
532 depf = open(os.path.join(tdir,'tool_dependencies.xml'),'w') | |
533 depf.write(tooldepcontent) | |
534 depf.write('\n') | |
535 depf.close() | |
536 if self.opts.input_tab <> 'None': # no reproducible test otherwise? TODO: maybe.. | |
537 testdir = os.path.join(tdir,'test-data') | |
538 os.mkdir(testdir) # make tests directory | |
539 for i,intab in enumerate(self.opts.input_tab): | |
540 si = self.opts.input_tab[i] | |
541 if si.find(',') <> -1: | |
542 s = si.split(',')[0] | |
543 si = s | |
544 dest = os.path.join(testdir,os.path.basename(si)) | |
545 if si <> dest: | |
546 shutil.copyfile(si,dest) | |
547 if self.opts.output_tab <> None: | |
548 shutil.copyfile(self.opts.output_tab,os.path.join(testdir,self.test1Output)) | |
549 if self.opts.make_HTML: | |
550 shutil.copyfile(self.opts.output_html,os.path.join(testdir,self.test1HTML)) | |
551 if self.opts.output_dir: | |
552 shutil.copyfile(self.tlog,os.path.join(testdir,'test1_out.log')) | |
553 outpif = '%s.py' % self.toolname # new name | |
554 outpiname = os.path.join(tdir,outpif) # path for the tool tarball | |
555 pyin = os.path.basename(self.pyfile) # our name - we rewrite ourselves (TM) | |
556 notes = ['# %s - a self annotated version of %s generated by running %s\n' % (outpiname,pyin,pyin),] | |
557 notes.append('# to make a new Galaxy tool called %s\n' % self.toolname) | |
558 notes.append('# User %s at %s\n' % (self.opts.user_email,timenow())) | |
559 pi = open(self.pyfile,'r').readlines() # our code becomes new tool wrapper (!) - first Galaxy worm | |
560 notes += pi | |
561 outpi = open(outpiname,'w') | |
562 outpi.write(''.join(notes)) | |
563 outpi.write('\n') | |
564 outpi.close() | |
565 stname = os.path.join(tdir,self.sfile) | |
566 if not os.path.exists(stname): | |
567 shutil.copyfile(self.sfile, stname) | |
568 xtname = os.path.join(tdir,self.xmlfile) | |
569 if not os.path.exists(xtname): | |
570 shutil.copyfile(self.xmlfile,xtname) | |
571 tarpath = "%s.tar.gz" % self.toolname | |
572 tar = tarfile.open(tarpath, "w:gz") | |
573 tar.add(tdir,arcname='%s' % self.toolname) | |
574 tar.close() | |
575 shutil.copyfile(tarpath,self.opts.new_tool) | |
576 shutil.rmtree(tdir) | |
577 ## TODO: replace with optional direct upload to local toolshed? | |
578 return retval | |
579 | |
580 | |
581 def compressPDF(self,inpdf=None,thumbformat='png'): | |
582 """need absolute path to pdf | |
583 note that GS gets confoozled if no $TMP or $TEMP | |
584 so we set it | |
585 """ | |
586 assert os.path.isfile(inpdf), "## Input %s supplied to %s compressPDF not found" % (inpdf,self.myName) | |
587 hlog = os.path.join(self.opts.output_dir,"compress_%s.txt" % os.path.basename(inpdf)) | |
588 sto = open(hlog,'a') | |
589 our_env = os.environ.copy() | |
590 our_tmp = our_env.get('TMP',None) | |
591 if not our_tmp: | |
592 our_tmp = our_env.get('TEMP',None) | |
593 if not (our_tmp and os.path.exists(our_tmp)): | |
594 newtmp = os.path.join(self.opts.output_dir,'tmp') | |
595 try: | |
596 os.mkdir(newtmp) | |
597 except: | |
598 sto.write('## WARNING - cannot make %s - it may exist or permissions need fixing\n' % newtmp) | |
599 our_env['TEMP'] = newtmp | |
600 if not self.temp_warned: | |
601 sto.write('## WARNING - no $TMP or $TEMP!!! Please fix - using %s temporarily\n' % newtmp) | |
602 self.temp_warned = True | |
603 outpdf = '%s_compressed' % inpdf | |
604 cl = ["gs", "-sDEVICE=pdfwrite", "-dNOPAUSE", "-dUseCIEColor", "-dBATCH","-dPDFSETTINGS=/printer", "-sOutputFile=%s" % outpdf,inpdf] | |
605 x = subprocess.Popen(cl,stdout=sto,stderr=sto,cwd=self.opts.output_dir,env=our_env) | |
606 retval1 = x.wait() | |
607 sto.close() | |
608 if retval1 == 0: | |
609 os.unlink(inpdf) | |
610 shutil.move(outpdf,inpdf) | |
611 os.unlink(hlog) | |
612 hlog = os.path.join(self.opts.output_dir,"thumbnail_%s.txt" % os.path.basename(inpdf)) | |
613 sto = open(hlog,'w') | |
614 outpng = '%s.%s' % (os.path.splitext(inpdf)[0],thumbformat) | |
615 if self.useGM: | |
616 cl2 = ['gm', 'convert', inpdf, outpng] | |
617 else: # assume imagemagick | |
618 cl2 = ['convert', inpdf, outpng] | |
619 x = subprocess.Popen(cl2,stdout=sto,stderr=sto,cwd=self.opts.output_dir,env=our_env) | |
620 retval2 = x.wait() | |
621 sto.close() | |
622 if retval2 == 0: | |
623 os.unlink(hlog) | |
624 retval = retval1 or retval2 | |
625 return retval | |
626 | |
627 | |
628 def getfSize(self,fpath,outpath): | |
629 """ | |
630 format a nice file size string | |
631 """ | |
632 size = '' | |
633 fp = os.path.join(outpath,fpath) | |
634 if os.path.isfile(fp): | |
635 size = '0 B' | |
636 n = float(os.path.getsize(fp)) | |
637 if n > 2**20: | |
638 size = '%1.1f MB' % (n/2**20) | |
639 elif n > 2**10: | |
640 size = '%1.1f KB' % (n/2**10) | |
641 elif n > 0: | |
642 size = '%d B' % (int(n)) | |
643 return size | |
644 | |
645 def makeHtml(self): | |
646 """ Create an HTML file content to list all the artifacts found in the output_dir | |
647 """ | |
648 | |
649 galhtmlprefix = """<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> | |
650 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en"> | |
651 <head> <meta http-equiv="Content-Type" content="text/html; charset=utf-8" /> | |
652 <meta name="generator" content="Galaxy %s tool output - see http://g2.trac.bx.psu.edu/" /> | |
653 <title></title> | |
654 <link rel="stylesheet" href="/static/style/base.css" type="text/css" /> | |
655 </head> | |
656 <body> | |
657 <div class="toolFormBody"> | |
658 """ | |
659 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/>""" | |
660 galhtmlpostfix = """</div></body></html>\n""" | |
661 | |
662 flist = os.listdir(self.opts.output_dir) | |
663 flist = [x for x in flist if x <> 'Rplots.pdf'] | |
664 flist.sort() | |
665 html = [] | |
666 html.append(galhtmlprefix % progname) | |
667 html.append('<div class="infomessage">Galaxy Tool "%s" run at %s</div><br/>' % (self.toolname,timenow())) | |
668 fhtml = [] | |
669 if len(flist) > 0: | |
670 logfiles = [x for x in flist if x.lower().endswith('.log')] # log file names determine sections | |
671 logfiles.sort() | |
672 logfiles = [x for x in logfiles if os.path.abspath(x) <> os.path.abspath(self.tlog)] | |
673 logfiles.append(os.path.abspath(self.tlog)) # make it the last one | |
674 pdflist = [] | |
675 npdf = len([x for x in flist if os.path.splitext(x)[-1].lower() == '.pdf']) | |
676 for rownum,fname in enumerate(flist): | |
677 dname,e = os.path.splitext(fname) | |
678 sfsize = self.getfSize(fname,self.opts.output_dir) | |
679 if e.lower() == '.pdf' : # compress and make a thumbnail | |
680 thumb = '%s.%s' % (dname,self.thumbformat) | |
681 pdff = os.path.join(self.opts.output_dir,fname) | |
682 retval = self.compressPDF(inpdf=pdff,thumbformat=self.thumbformat) | |
683 if retval == 0: | |
684 pdflist.append((fname,thumb)) | |
685 else: | |
686 pdflist.append((fname,fname)) | |
687 if (rownum+1) % 2 == 0: | |
688 fhtml.append('<tr class="odd_row"><td><a href="%s">%s</a></td><td>%s</td></tr>' % (fname,fname,sfsize)) | |
689 else: | |
690 fhtml.append('<tr><td><a href="%s">%s</a></td><td>%s</td></tr>' % (fname,fname,sfsize)) | |
691 for logfname in logfiles: # expect at least tlog - if more | |
692 if os.path.abspath(logfname) == os.path.abspath(self.tlog): # handled later | |
693 sectionname = 'All tool run' | |
694 if (len(logfiles) > 1): | |
695 sectionname = 'Other' | |
696 ourpdfs = pdflist | |
697 else: | |
698 realname = os.path.basename(logfname) | |
699 sectionname = os.path.splitext(realname)[0].split('_')[0] # break in case _ added to log | |
700 ourpdfs = [x for x in pdflist if os.path.basename(x[0]).split('_')[0] == sectionname] | |
701 pdflist = [x for x in pdflist if os.path.basename(x[0]).split('_')[0] <> sectionname] # remove | |
702 nacross = 1 | |
703 npdf = len(ourpdfs) | |
704 | |
705 if npdf > 0: | |
706 nacross = math.sqrt(npdf) ## int(round(math.log(npdf,2))) | |
707 if int(nacross)**2 != npdf: | |
708 nacross += 1 | |
709 nacross = int(nacross) | |
710 width = min(400,int(1200/nacross)) | |
711 html.append('<div class="toolFormTitle">%s images and outputs</div>' % sectionname) | |
712 html.append('(Click on a thumbnail image to download the corresponding original PDF image)<br/>') | |
713 ntogo = nacross # counter for table row padding with empty cells | |
714 html.append('<div><table class="simple" cellpadding="2" cellspacing="2">\n<tr>') | |
715 for i,paths in enumerate(ourpdfs): | |
716 fname,thumb = paths | |
717 s= """<td><a href="%s"><img src="%s" title="Click to download a PDF of %s" hspace="5" width="%d" | |
718 alt="Image called %s"/></a></td>\n""" % (fname,thumb,fname,width,fname) | |
719 if ((i+1) % nacross == 0): | |
720 s += '</tr>\n' | |
721 ntogo = 0 | |
722 if i < (npdf - 1): # more to come | |
723 s += '<tr>' | |
724 ntogo = nacross | |
725 else: | |
726 ntogo -= 1 | |
727 html.append(s) | |
728 if html[-1].strip().endswith('</tr>'): | |
729 html.append('</table></div>\n') | |
730 else: | |
731 if ntogo > 0: # pad | |
732 html.append('<td> </td>'*ntogo) | |
733 html.append('</tr></table></div>\n') | |
734 logt = open(logfname,'r').readlines() | |
735 logtext = [x for x in logt if x.strip() > ''] | |
736 html.append('<div class="toolFormTitle">%s log output</div>' % sectionname) | |
737 if len(logtext) > 1: | |
738 html.append('\n<pre>\n') | |
739 html += logtext | |
740 html.append('\n</pre>\n') | |
741 else: | |
742 html.append('%s is empty<br/>' % logfname) | |
743 if len(fhtml) > 0: | |
744 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') | |
745 fhtml.append('</table></div><br/>') | |
746 html.append('<div class="toolFormTitle">All output files available for downloading</div>\n') | |
747 html += fhtml # add all non-pdf files to the end of the display | |
748 else: | |
749 html.append('<div class="warningmessagelarge">### Error - %s returned no files - please confirm that parameters are sane</div>' % self.opts.interpreter) | |
750 html.append(galhtmlpostfix) | |
751 htmlf = file(self.opts.output_html,'w') | |
752 htmlf.write('\n'.join(html)) | |
753 htmlf.write('\n') | |
754 htmlf.close() | |
755 self.html = html | |
756 | |
757 | |
758 def run(self): | |
759 """ | |
760 scripts must be small enough not to fill the pipe! | |
761 """ | |
762 if self.treatbashSpecial and self.opts.interpreter in ['bash','sh']: | |
763 retval = self.runBash() | |
764 else: | |
765 if self.opts.output_dir: | |
766 ste = open(self.elog,'w') | |
767 sto = open(self.tlog,'w') | |
768 sto.write('## Toolfactory generated command line = %s\n' % ' '.join(self.cl)) | |
769 sto.flush() | |
770 p = subprocess.Popen(self.cl,shell=False,stdout=sto,stderr=ste,stdin=subprocess.PIPE,cwd=self.opts.output_dir) | |
771 else: | |
772 p = subprocess.Popen(self.cl,shell=False,stdin=subprocess.PIPE) | |
773 p.stdin.write(self.script) | |
774 p.stdin.close() | |
775 retval = p.wait() | |
776 if self.opts.output_dir: | |
777 sto.close() | |
778 ste.close() | |
779 err = open(self.elog,'r').readlines() | |
780 if retval <> 0 and err: # problem | |
781 print >> sys.stderr,err | |
782 if self.opts.make_HTML: | |
783 self.makeHtml() | |
784 return retval | |
785 | |
786 def runBash(self): | |
787 """ | |
788 cannot use - for bash so use self.sfile | |
789 """ | |
790 if self.opts.output_dir: | |
791 s = '## Toolfactory generated command line = %s\n' % ' '.join(self.cl) | |
792 sto = open(self.tlog,'w') | |
793 sto.write(s) | |
794 sto.flush() | |
795 p = subprocess.Popen(self.cl,shell=False,stdout=sto,stderr=sto,cwd=self.opts.output_dir) | |
796 else: | |
797 p = subprocess.Popen(self.cl,shell=False) | |
798 retval = p.wait() | |
799 if self.opts.output_dir: | |
800 sto.close() | |
801 if self.opts.make_HTML: | |
802 self.makeHtml() | |
803 return retval | |
804 | |
805 | |
806 def main(): | |
807 u = """ | |
808 This is a Galaxy wrapper. It expects to be called by a special purpose tool.xml as: | |
809 <command interpreter="python">rgBaseScriptWrapper.py --script_path "$scriptPath" --tool_name "foo" --interpreter "Rscript" | |
810 </command> | |
811 """ | |
812 op = optparse.OptionParser() | |
813 a = op.add_option | |
814 a('--script_path',default=None) | |
815 a('--tool_name',default=None) | |
816 a('--interpreter',default=None) | |
817 a('--output_dir',default='./') | |
818 a('--output_html',default=None) | |
819 a('--input_tab',default=[], action="append") # these are "galaxypath,metadataname" pairs | |
820 a("--input_formats",default="tabular") | |
821 a('--output_tab',default=None) | |
822 a('--output_format',default='tabular') | |
823 a('--user_email',default='Unknown') | |
824 a('--bad_user',default=None) | |
825 a('--make_Tool',default=None) | |
826 a('--make_HTML',default=None) | |
827 a('--help_text',default=None) | |
828 a('--tool_desc',default=None) | |
829 a('--new_tool',default=None) | |
830 a('--tool_version',default=None) | |
831 a('--include_dependencies',default="yes") | |
832 a('--citations',default=None) | |
833 a('--additional_parameters', dest='additional_parameters', action='append', default=[]) | |
834 a('--edit_additional_parameters', action="store_true", default=False) | |
835 opts, args = op.parse_args() | |
836 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) | |
837 assert opts.tool_name,'## Tool Factory expects a tool name - eg --tool_name=DESeq' | |
838 assert opts.interpreter,'## Tool Factory wrapper expects an interpreter - eg --interpreter=Rscript' | |
839 assert os.path.isfile(opts.script_path),'## Tool Factory wrapper expects a script path - eg --script_path=foo.R' | |
840 if opts.output_dir: | |
841 try: | |
842 os.makedirs(opts.output_dir) | |
843 except: | |
844 pass | |
845 opts.input_tab = [x.replace('"','').replace("'",'') for x in opts.input_tab] | |
846 for i,x in enumerate(opts.additional_parameters): # remove quotes we need to deal with spaces in CL params | |
847 opts.additional_parameters[i] = opts.additional_parameters[i].replace('"','') | |
848 r = ScriptRunner(opts) | |
849 if opts.make_Tool: | |
850 retcode = r.makeTooltar() | |
851 else: | |
852 retcode = r.run() | |
853 os.unlink(r.sfile) | |
854 if retcode: | |
855 sys.exit(retcode) # indicate failure to job runner | |
856 | |
857 | |
858 if __name__ == "__main__": | |
859 main() | |
860 | |
861 |