def sh_script(self, context, **args):
if 'echo_name' in args and args['echo_name']:
echo_part="\n# step %s:\n" % self.name
echo_part+="echo step %s 1>&2" % self.name
else:
echo_part=''
try: usage=self.usage(context)
except KeyError as ke:
raise ConfigError("missing config item '%s'" % ke)
# check for missing attrs after calling step.usage()
missing_attrs=self.missing_required_attrs()
if len(missing_attrs) > 0:
raise ConfigError("step %s: missing attributes: %s" % (self.name, ', '.join(missing_attrs)))
vars={}
vars.update(self.__dict__)
vars.update(self.pipeline.readset)
if not self.is_prov_step:
vars.update(self.pipeline[self.name])
vars['inputs']=context.inputs[self.name]
vars['outputs']=context.outputs[self.name]
vars['pipeline']=self.pipeline
vars['pipeline_run_id']=context.pipeline_run_id
#vars['step_run_id']=context.step_runs[self.name].id
#vars['next_step_run_id']=context.step_runs[self.pipeline.step_after(step.name)].id
vars['config']=RnaseqGlobals.config
vars['readset']=self.pipeline.readset
# need to add shell variables for 'set': (in cufflinks.s_?.sh scripts)
# currently root_dir, programs, reads_file, ID, format, readlen
# but really, the pipeline should specify these?
# or only things that are truly universal
vars['root_dir']=RnaseqGlobals.root_dir()
# add readset exports:
readset=self.pipeline.readset
for attr in readset.exports:
try: vars[attr]=getattr(readset, attr)
except AttributeError:
vars[attr]=''
#warn("%s.sh_script: no '%s' readset attribute!" % (self.name, attr))
# add self.exports:
try: export_list=self.exports
except: export_list=[]
for attr in export_list:
vars[attr]=getattr(self,attr)
try: script_part=evoque_template(usage, vars)
except Exception as e: raise ConfigError("step %s: %s" % (self.name, e))
script="\n".join([echo_part,script_part]) # tried using echo_part+sh_script, got weird '>' -> '>' substitutions
return script
def test_missing_value(self):
template="This has a ${missing} value"
vars={'not_missing':1}
try:
string=evoque_template(template, vars)
self.fail()
except exceptions.NameError as e:
# warn("caught %s (%s)" % (e, type(e)))
self.assertRegexpMatches(str(e), "name 'missing' is not defined")
def evoque_fields(self):
vars=self.__dict__
vars.update(RnaseqGlobals.conf_value('rnaseq'))
for a in dir(self):
if a.startswith('__'): continue
attr=getattr(self,a)
if type(attr) != type(''): continue
if not re.search('\$\{', attr): continue
try: setattr(self, a, evoque_template(attr, vars))
except NameError: pass
return self
def load_glob(self,yml,filename):
try: reads_files=yml['reads_file']
except KeyError:
try: reads_files=yml['reads_files']
except KeyError: raise ConfigError("%s: readset does not define reads_file(s)" % filename)
# get the reads directory, if present
try: reads_dir=yml['reads_dir']
except KeyError: reads_dir=os.getcwd() # fixme: hope this is right thing to do
# hack?
reads_dir=evoque_template(reads_dir, RnaseqGlobals.config['rnaseq']) # in case of ${root_dir}
#print "lg: reads_dir is %s" % reads_dir
files=[]
globlist=re.split('[\s,]+',reads_files)
missing=[]
for fglob in globlist:
if not os.path.isabs(fglob):
fglob=os.path.normpath(os.path.join(reads_dir, fglob))
filelist=glob.glob(fglob)
if len(files)==0:
missing.append("%s: no matching files for %s" % (filename, fglob))
files.extend(filelist)
if len(files)==0:
msg=", ".join(missing)
raise ConfigError(msg)
# Make the Readset object:
#print "load_glob: files is %s" % files
scalars=scalar_values(yml)
scalars['filename']=filename
scalars['reads_files']=files
return [Readset(scalars)]
def test_basic(self):
template="This is a ${adjective} template"
vars={'adjective': 'simple'}
string=evoque_template(template, vars)
self.assertEqual(string, 'This is a simple template')
def output_list_expanded(self):
l=[evoque_template(x, self, self.pipeline.readset) for x in self.output_list()]
return l
