def run(config_file, subject_list_file, p_name = None):
# Import packages
import time
# take date+time stamp for run identification purposes
unique_pipeline_id = strftime("%Y%m%d%H%M%S")
pipeline_start_stamp = strftime("%Y-%m-%d_%H:%M:%S")
try:
if not os.path.exists(config_file):
raise IOError
else:
c = Configuration(yaml.load(open(os.path.realpath(config_file), 'r')))
except IOError:
print("config file %s doesn't exist" % config_file)
raise
except Exception:
print("Error reading config file - %s" % config_file)
raise Exception
#do some validation
validate(c)
# get the pipeline name
p_name = c.pipelineName
try:
sublist = yaml.load(open(os.path.realpath(subject_list_file), 'r'))
except:
print("Subject list is not in proper YAML format. Please check your file")
raise Exception
# NOTE: strategies list is only needed in cpac_pipeline prep_workflow for
# creating symlinks
strategies = sorted(build_strategies(c))
print("strategies ---> ")
print(strategies)
sub_scan_map ={}
print("subject list: ")
print(sublist)
try:
for sub in sublist:
if sub['unique_id']:
s = sub['subject_id']+"_" + sub["unique_id"]
else:
s = sub['subject_id']
scan_ids = ['scan_anat']
for id in sub['rest']:
scan_ids.append('scan_'+ str(id))
sub_scan_map[s] = scan_ids
except:
print("\n\n" + "ERROR: Subject list file not in proper format - check if you loaded the correct file?" + "\n" + \
"Error name: cpac_runner_0001" + "\n\n")
raise Exception
create_group_log_template(sub_scan_map, os.path.join(c.outputDirectory, 'logs'))
seeds_created = []
if not (c.seedSpecificationFile is None):
try:
if os.path.exists(c.seedSpecificationFile):
seeds_created = create_seeds_(c.seedOutputLocation, c.seedSpecificationFile, c.FSLDIR)
print('seeds created %s -> ' % seeds_created)
except:
raise IOError('Problem in seedSpecificationFile')
if 1 in c.runVoxelTimeseries:
if 'roi_voxelwise' in c.useSeedInAnalysis:
c.maskSpecificationFile = append_seeds_to_file(c.workingDirectory, seeds_created, c.maskSpecificationFile)
if 1 in c.runROITimeseries:
if 'roi_average' in c.useSeedInAnalysis:
c.roiSpecificationFile = append_seeds_to_file(c.workingDirectory, seeds_created, c.roiSpecificationFile)
if 1 in c.runSCA:
if 'roi_average' in c.useSeedInAnalysis:
c.roiSpecificationFileForSCA = append_seeds_to_file(c.workingDirectory, seeds_created, c.roiSpecificationFileForSCA)
if 1 in c.runNetworkCentrality:
if 'centrality_outputs_smoothed' in c.useSeedInAnalysis:
c.templateSpecificationFile = append_seeds_to_file(c.workingDirectory, seeds_created, c.templateSpecificationFile)
pipeline_timing_info = []
pipeline_timing_info.append(unique_pipeline_id)
pipeline_timing_info.append(pipeline_start_stamp)
pipeline_timing_info.append(len(sublist))
if not c.runOnGrid:
# Import packages
from CPAC.pipeline.cpac_pipeline import prep_workflow
# Init variables
procss = [Process(target=prep_workflow,
args=(sub, c, strategies, 1,
pipeline_timing_info, p_name)) \
for sub in sublist]
pid = open(os.path.join(c.outputDirectory, 'pid.txt'), 'w')
# Init job queue
jobQueue = []
# If we're allocating more processes than are subjects, run them all
if len(sublist) <= c.numSubjectsAtOnce:
"""
Stream all the subjects as sublist is
less than or equal to the number of
subjects that need to run
"""
for p in procss:
p.start()
print(p.pid, file=pid)
# Otherwise manage resources to run processes incrementally
else:
"""
Stream the subject workflows for preprocessing.
At Any time in the pipeline c.numSubjectsAtOnce
will run, unless the number remaining is less than
the value of the parameter stated above
"""
idx = 0
while(idx < len(sublist)):
# If the job queue is empty and we haven't started indexing
if len(jobQueue) == 0 and idx == 0:
# Init subject process index
idc = idx
# Launch processes (one for each subject)
for p in procss[idc: idc + c.numSubjectsAtOnce]:
p.start()
print(p.pid, file=pid)
jobQueue.append(p)
idx += 1
# Otherwise, jobs are running - check them
else:
# Check every job in the queue's status
for job in jobQueue:
# If the job is not alive
if not job.is_alive():
# Find job and delete it from queue
print('found dead job ', job)
loc = jobQueue.index(job)
del jobQueue[loc]
# ...and start the next available process (subject)
procss[idx].start()
# Append this to job queue and increment index
jobQueue.append(procss[idx])
idx += 1
# Add sleep so while loop isn't consuming 100% of CPU
time.sleep(2)
pid.close()
else:
import subprocess
import pickle
temp_files_dir = os.path.join(os.getcwd(), 'cluster_temp_files')
print(subprocess.getoutput("mkdir -p %s" % temp_files_dir))
strategies_file = os.path.join(temp_files_dir, 'strategies.obj')
f = open(strategies_file, 'w')
pickle.dump(strategies, f)
f.close()
if 'sge' in c.resourceManager.lower():
run_sge_jobs(c, config_file, strategies_file, subject_list_file, p_name)
elif 'pbs' in c.resourceManager.lower():
run_pbs_jobs(c, config_file, strategies_file, subject_list_file, p_name)
elif 'condor' in c.resourceManager.lower():
run_condor_jobs(c, config_file, strategies_file, subject_list_file, p_name)
def prep_group_analysis_workflow(c, resource, subject_infos):
p_id, s_ids, scan_ids, s_paths = (list(tup) for tup in zip(*subject_infos))
# print "p_id -%s, s_ids -%s, scan_ids -%s, s_paths -%s" %(p_id, s_ids, scan_ids, s_paths)
def get_phenotypic_file(phenotypic_file, m_dict, m_list, mod_path):
# print "phenotypic_file, m_dict", phenotypic_file, m_dict
import csv
reader = csv.reader(open(phenotypic_file, "rU"))
columns = {}
order = {}
count = 0
headers = reader.next()
for h in headers:
columns[h] = []
order[h] = count
count += 1
for r in reader:
for h, v in zip(headers, r):
if v:
columns[h].append(str(v))
if m_dict:
for measure in m_list:
if measure in headers:
# check if 'MeanFD is present'
if len(columns[measure]) < 1:
for sub in columns["sub_id"]:
if m_dict.get(sub):
if m_dict.get(sub).get(measure):
columns[measure].append(m_dict[sub][measure])
else:
raise Exception("Couldn't find %s value for subject %s" % (measure, sub))
else:
raise Exception("Couldn't find subject %s in the parameter file" % sub)
b = zip(*([k] + columns[k] for k in sorted(columns, key=order.get)))
try:
os.makedirs(mod_path)
except:
print "%s already exist" % (mod_path)
new_phenotypic_file = os.path.join(mod_path, os.path.basename(phenotypic_file))
a = csv.writer(open(new_phenotypic_file, "w"))
for col in b:
a.writerow(list(col))
return new_phenotypic_file
threshold_val = None
measure_dict = None
measure_list = ["MeanFD", "MeanFD_Jenkinson", "MeanDVARS"]
model_sub_list = []
if re.search("(?<=/_threshold_)\d+.\d+", s_paths[0]):
threshold_val = re.search("(?<=/_threshold_)\d+.\d+", s_paths[0]).group(0)
elif len(c.scrubbingThreshold) == 1:
threshold_val = c.scrubbingThreshold[0]
else:
print ("Found Multiple threshold value ")
print "threhsold_val -->", threshold_val
if threshold_val:
try:
parameter_file = os.path.join(
c.outputDirectory, p_id[0], "%s_threshold_%s_all_params.csv" % (scan_ids[0].strip("_"), threshold_val)
)
if os.path.exists(parameter_file):
import csv
measure_dict = {}
f = csv.DictReader(open(parameter_file, "r"))
for line in f:
measure_map = {}
for m in measure_list:
if line.get(m):
measure_map[m] = line[m]
measure_dict[line["Subject"]] = measure_map
else:
print "No file name %s found" % parameter_file
except Exception:
print "Exception while extracting parameters from movement file - %s" % (parameter_file)
for config in c.modelConfigs:
import yaml
try:
conf = Configuration(yaml.load(open(os.path.realpath(config), "r")))
except:
raise Exception("Error in reading %s configuration file" % config)
subject_list = [
line.rstrip("\r\n")
for line in open(conf.subjectListFile, "r")
if not (line == "\n") and not line.startswith("#")
]
exist_paths = []
for sub in subject_list:
for path in s_paths:
if sub in path:
exist_paths.append(sub)
if len(list(set(subject_list) - set(exist_paths))) > 0:
print "list of outputs missing for subjects %s for derivative -%s at path- %s" % (
list(set(subject_list) - set(exist_paths)),
resource,
os.path.dirname(s_paths[0]).replace(s_ids[0], "*"),
)
mod_path = os.path.join(
os.path.dirname(s_paths[0]).replace(s_ids[0], "group_analysis_results/_grp_model_%s" % (conf.modelName)),
"model_files",
)
try:
os.makedirs(mod_path)
except:
print "path %s already exists" % mod_path
new_sub_file = os.path.join(mod_path, os.path.basename(conf.subjectListFile))
f = open(new_sub_file, "w")
for sub in exist_paths:
print >> f, sub
f.close()
conf.update("subjectListFile", new_sub_file)
if measure_dict != None:
conf.update(
"phenotypicFile", get_phenotypic_file(conf.phenotypicFile, measure_dict, measure_list, mod_path)
)
print "model config dictionary ->", conf.__dict__
try:
from CPAC.utils import create_fsl_model
create_fsl_model.run(conf, True)
except Exception, e:
print "Error in create_fsl_model script"
print e
model_sub_list.append((conf.outputModelFilesDirectory, conf.subjectListFile))
print "model_sub_list ->", model_sub_list
def run(config_file, subject_list_file, p_name = None):
# take date+time stamp for run identification purposes
unique_pipeline_id = strftime("%Y%m%d%H%M%S")
pipeline_start_stamp = strftime("%Y-%m-%d_%H:%M:%S")
try:
if not os.path.exists(config_file):
raise IOError
else:
c = Configuration(yaml.load(open(os.path.realpath(config_file), 'r')))
except IOError:
print "config file %s doesn't exist" % config_file
raise
except Exception:
print "Error reading config file - %s" % config_file
raise Exception
#do some validation
validate(c)
try:
sublist = yaml.load(open(os.path.realpath(subject_list_file), 'r'))
except:
print "Subject list is not in proper YAML format. Please check your file"
raise Exception
strategies = sorted(build_strategies(c))
print "strategies ---> "
print strategies
sub_scan_map ={}
print "subject list: "
print sublist
try:
for sub in sublist:
if sub['unique_id']:
s = sub['subject_id']+"_" + sub["unique_id"]
else:
s = sub['subject_id']
scan_ids = ['scan_anat']
for id in sub['rest']:
scan_ids.append('scan_'+ str(id))
sub_scan_map[s] = scan_ids
except:
print "\n\n" + "ERROR: Subject list file not in proper format - check if you loaded the correct file?" + "\n" + \
"Error name: cpac_runner_0001" + "\n\n"
raise Exception
create_group_log_template(sub_scan_map, os.path.join(c.outputDirectory, 'logs'))
seeds_created = []
if not (c.seedSpecificationFile is None):
try:
if os.path.exists(c.seedSpecificationFile):
seeds_created = create_seeds_(c.seedOutputLocation, c.seedSpecificationFile, c.FSLDIR)
print 'seeds created %s -> ' % seeds_created
except:
raise IOError('Problem in seedSpecificationFile')
if 1 in c.runVoxelTimeseries:
if 'roi_voxelwise' in c.useSeedInAnalysis:
c.maskSpecificationFile = append_seeds_to_file(c.workingDirectory, seeds_created, c.maskSpecificationFile)
if 1 in c.runROITimeseries:
if 'roi_average' in c.useSeedInAnalysis:
c.roiSpecificationFile = append_seeds_to_file(c.workingDirectory, seeds_created, c.roiSpecificationFile)
if 1 in c.runNetworkCentrality:
if 'centrality_outputs_smoothed' in c.useSeedInAnalysis:
c.templateSpecificationFile = append_seeds_to_file(c.workingDirectory, seeds_created, c.templateSpecificationFile)
pipeline_timing_info = []
pipeline_timing_info.append(unique_pipeline_id)
pipeline_timing_info.append(pipeline_start_stamp)
pipeline_timing_info.append(len(sublist))
if not c.runOnGrid:
from CPAC.pipeline.cpac_pipeline import prep_workflow
procss = [Process(target=prep_workflow, args=(sub, c, strategies, 1, pipeline_timing_info, p_name)) for sub in sublist]
pid = open(os.path.join(c.outputDirectory, 'pid.txt'), 'w')
jobQueue = []
if len(sublist) <= c.numSubjectsAtOnce:
"""
Stream all the subjects as sublist is
less than or equal to the number of
subjects that need to run
"""
for p in procss:
p.start()
print >>pid,p.pid
else:
"""
Stream the subject workflows for preprocessing.
At Any time in the pipeline c.numSubjectsAtOnce
will run, unless the number remaining is less than
the value of the parameter stated above
"""
idx = 0
while(idx < len(sublist)):
if len(jobQueue) == 0 and idx == 0:
idc = idx
for p in procss[idc: idc + c.numSubjectsAtOnce]:
p.start()
print >>pid,p.pid
jobQueue.append(p)
idx += 1
else:
for job in jobQueue:
if not job.is_alive():
print 'found dead job ', job
loc = jobQueue.index(job)
del jobQueue[loc]
procss[idx].start()
jobQueue.append(procss[idx])
idx += 1
pid.close()
else:
import commands
import pickle
temp_files_dir = os.path.join(os.getcwd(), 'cluster_temp_files')
print commands.getoutput("mkdir -p %s" % temp_files_dir)
strategies_file = os.path.join(temp_files_dir, 'strategies.obj')
f = open(strategies_file, 'w')
pickle.dump(strategies, f)
f.close()
if 'sge' in c.resourceManager.lower():
run_sge_jobs(c, config_file, strategies_file, subject_list_file, p_name)
elif 'pbs' in c.resourceManager.lower():
run_pbs_jobs(c, config_file, strategies_file, subject_list_file, p_name)
elif 'condor' in c.resourceManager.lower():
run_condor_jobs(c, config_file, strategies_file, subject_list_file, p_name)
def prep_group_analysis_workflow(c, resource, subject_infos):
p_id, s_ids, scan_ids, s_paths = (list(tup) for tup in zip(*subject_infos))
#print "p_id -%s, s_ids -%s, scan_ids -%s, s_paths -%s" %(p_id, s_ids, scan_ids, s_paths)
def get_phenotypic_file(phenotypic_file, m_dict, m_list, mod_path, sub_id):
#print "phenotypic_file, m_dict", phenotypic_file, m_dict
import csv
reader = csv.reader(open(phenotypic_file, 'rU'))
columns = {}
order = {}
count = 0
headers = reader.next()
for h in headers:
columns[h] =[]
order[h] = count
count+=1
for r in reader:
for h, v in zip(headers, r):
if v:
columns[h].append(str(v))
if m_dict:
for measure in m_list:
if measure in headers:
#check if 'MeanFD is present'
if len(columns[measure]) < 1:
for sub in columns[sub_id]:
if m_dict.get(sub):
if m_dict.get(sub).get(measure):
columns[measure].append(m_dict[sub][measure])
else:
raise Exception("Couldn't find %s value for subject %s"%(measure,sub))
else:
raise Exception("Couldn't find subject %s in the parameter file"%sub)
b = zip(*([k] + columns[k] for k in sorted(columns, key=order.get)))
try:
os.makedirs(mod_path)
except:
print "%s already exist"%(mod_path)
new_phenotypic_file = os.path.join(mod_path, os.path.basename(phenotypic_file))
a = csv.writer(open(new_phenotypic_file, 'w'))
for col in b:
a.writerow(list(col))
return new_phenotypic_file
threshold_val = None
measure_dict = None
measure_list = ['MeanFD', 'MeanFD_Jenkinson', 'MeanDVARS']
model_sub_list = []
if c.runScrubbing == 1:
#get scrubbing threshold
if re.search('(?<=/_threshold_)\d+.\d+',s_paths[0]):
threshold_val = re.search('(?<=/_threshold_)\d+.\d+',s_paths[0]).group(0)
elif len(c.scrubbingThreshold) == 1:
threshold_val = c.scrubbingThreshold[0]
else:
print ("Found Multiple threshold value ")
print "scrubbing threshold_val -->", threshold_val
else:
print "No scrubbing enabled."
print "\n"
#pick the right parameter file from the pipeline folder
#create a dictionary of subject and measures in measure_list
if c.runScrubbing == 1:
try:
parameter_file = os.path.join(c.outputDirectory, p_id[0], '%s_threshold_%s_all_params.csv'%(scan_ids[0].strip('_'),threshold_val))
if os.path.exists(parameter_file):
import csv
measure_dict = {}
f = csv.DictReader(open(parameter_file,'r'))
for line in f:
measure_map = {}
for m in measure_list:
if line.get(m):
measure_map[m] = line[m]
measure_dict[line['Subject']] = measure_map
else:
print "No file name %s found"%parameter_file
except Exception:
print "Exception while extracting parameters from movement file - %s"%(parameter_file)
for config in c.modelConfigs:
import yaml
try:
conf = Configuration(yaml.load(open(os.path.realpath(config), 'r')))
except:
raise Exception("Error in reading %s configuration file" % config)
subject_list = [line.rstrip('\r\n') for line in open(conf.subjectListFile, 'r') \
if not (line == '\n') and not line.startswith('#')]
# list of subject paths which DO exist
exist_paths = []
# check for missing subject for the derivative
for sub in subject_list :
for path in s_paths:
if sub in path:
exist_paths.append(sub)
# check to see if any derivatives of subjects are missing
if len(list(set(subject_list) - set(exist_paths))) >0:
print "-------------------------------------------"
print "List of outputs missing for subjects:"
print list(set(subject_list) - set(exist_paths))
print "\n"
print "..for derivatives:"
print resource
print "\n"
print "..at paths:"
print os.path.dirname(s_paths[0]).replace(s_ids[0], '*')
print "-------------------------------------------"
print '\n'
#import warnings
#warnings.warn(msg)
mod_path = os.path.join(os.path.dirname(s_paths[0]).replace(s_ids[0], 'group_analysis_results/_grp_model_%s'%(conf.modelName)),
'model_files')
print "basename: ", os.path.basename(conf.subjectListFile)
try:
os.makedirs(mod_path)
print "Creating directory:"
print mod_path
print "\n"
except:
print "Attempted to create directory, but path already exists:"
print mod_path
print '\n'
new_sub_file = os.path.join(mod_path, os.path.basename(conf.subjectListFile))
try:
f = open(new_sub_file, 'w')
for sub in exist_paths:
print >>f, sub
f.close()
except:
print "Error: Could not open subject list file: ", new_sub_file
print ""
raise Exception
conf.update('subjectListFile',new_sub_file)
sub_id = conf.subjectColumn
if measure_dict != None:
conf.update('phenotypicFile',get_phenotypic_file(conf.phenotypicFile, measure_dict, measure_list, mod_path, sub_id))
print "Model config dictionary ->"
print conf.__dict__
print '\n'
# Run 'create_fsl_model' script to extract phenotypic data from
# the phenotypic file for each of the subjects in the subject list
try:
from CPAC.utils import create_fsl_model
create_fsl_model.run(conf, True)
except Exception, e:
print "Error in creating models in the create_fsl_model script"
#print "Error ->", e
raise
model_sub_list.append((conf.outputModelFilesDirectory, conf.subjectListFile))
print "model_sub_list ->", model_sub_list
def prep_group_analysis_workflow(c, resource, subject_infos):
#
# this function runs once per output file during group analysis
#
# p_id = a list of pipeline IDs, i.e. the name of the output folder for
# the strat
# s_ids = a list of all the subject IDs
# scan_ids = a list of scan IDs
# s_paths = a list of all of the filepaths of this particular output
# file that prep_group_analysis_workflow is being called for
p_id, s_ids, scan_ids, s_paths = (list(tup) for tup in zip(*subject_infos))
def get_phenotypic_file(phenotypic_file, m_dict, m_list, mod_path, sub_id):
import csv
reader = csv.reader(open(phenotypic_file, 'rU'))
columns = {}
order = {}
count = 0
headers = reader.next()
for h in headers:
columns[h] =[]
order[h] = count
count+=1
for r in reader:
for h, v in zip(headers, r):
if v:
columns[h].append(str(v))
if m_dict:
for measure in m_list:
print '\n\nMeasure: ', measure, '\n\n'
if measure in headers:
#check if 'MeanFD is present'
if len(columns[measure]) < 1:
print '\n\ncolumns[sub_id]: ', columns[sub_id], '\n\n'
for sub in columns[sub_id]:
if m_dict.get(sub):
if m_dict.get(sub).get(measure):
columns[measure].append(m_dict[sub][measure])
else:
raise Exception("Couldn't find %s value for subject %s"%(measure,sub))
else:
raise Exception("Couldn't find subject %s in the parameter file"%sub)
print '\n\ncolumns[measure]: ', columns, '\n\n'
b = zip(*([k] + columns[k] for k in sorted(columns, key=order.get)))
try:
os.makedirs(mod_path)
except:
print "%s already exists"%(mod_path)
new_phenotypic_file = os.path.join(mod_path, os.path.basename(phenotypic_file))
a = csv.writer(open(new_phenotypic_file, 'w'))
for col in b:
a.writerow(list(col))
return new_phenotypic_file
# END get_phenotypic_file function
threshold_val = None
measure_dict = None
measure_list = ['MeanFD', 'MeanFD_Jenkinson', 'MeanDVARS']
model_sub_list = []
if 1 in c.runScrubbing:
#get scrubbing threshold
if re.search('(?<=/_threshold_)\d+.\d+',s_paths[0]):
threshold_val = re.search('(?<=/_threshold_)\d+.\d+',s_paths[0]).group(0)
elif len(c.scrubbingThreshold) == 1:
threshold_val = c.scrubbingThreshold[0]
else:
print "Found Multiple threshold value "
print "scrubbing threshold_val -->", threshold_val
else:
print "No scrubbing enabled."
if len(c.scrubbingThreshold) == 1:
threshold_val = c.scrubbingThreshold[0]
import yaml
for config in c.modelConfigs:
print c.modelConfigs
print config
try:
conf = Configuration(yaml.load(open(os.path.realpath(config), 'r')))
except:
raise Exception("Error in reading %s configuration file" % config)
group_sublist = open(conf.subject_list, 'r')
sublist_items = group_sublist.readlines()
subject_list = [line.rstrip('\n') for line in sublist_items \
if not (line == '\n') and not line.startswith('#')]
# list of subject paths which DO exist
exist_paths = []
print 'subject_list: ', subject_list, '\n\n'
print 's_paths: ', s_paths, '\n\n'
''' begin iteration through group subject list for processing '''
for sub in subject_list:
# let's check to make sure the subject list is formatted for
# repeated measures properly if repeated measures is enabled and
# vice versa
if (c.repeatedMeasures == True) and (',' not in sub):
print '\n\n'
print 'Whoops! The group analysis subject list is not in ' \
'the appropriate format for repeated measures.\n'
print 'Please use the appropriate format as described in ' \
'the CPAC User Guide or turn off Repeated Measures ' \
'in the CPAC pipeline configuration editor, found ' \
'in the \'Group Analysis Settings\' tab of the ' \
'pipeline configuration editor.\n'
print 'NOTE: CPAC generates a properly-formatted group ' \
'analysis subject list meant for running repeated ' \
'measures when you create your original subject ' \
'list. Look for \'subject_list_group_analysis_' \
'repeated_measures.txt\' in the directory where ' \
'you created your subject list.\n\n'
raise Exception
elif (c.repeatedMeasures == False) and (',' in sub):
print '\n\n'
print '[!] CPAC says: It looks like your group analysis ' \
'subject list is formatted for running repeated ' \
'measures, but \'Run Repeated Measures\' is not ' \
'enabled in the pipeline configuration, found in ' \
'the \'Group Analysis Settings\' tab of the ' \
'pipeline configuration editor.\n'
print 'Double-check your pipeline configuration?\n\n'
raise Exception
''' process subject ids for repeated measures, if it is on '''
# if repeated measures is being run and the subject list
# is a list of subject IDs and scan IDs concatenated
if (c.repeatedMeasures == True):
# sub.count(',') equals 1 when there is either multiple scans
# or multiple sessions but not both, for repeated measures
# sub.count(',') equals 2 when there are multiple sessions
# AND scans, for repeated measures
if sub.count(',') == 1:
sub_id = sub.split(',',1)[0]
other_id = sub.split(',',1)[1]
elif sub.count(',') == 2:
sub_id = sub.split(',',2)[0]
scan_id = sub.split(',',2)[1]
session_id = sub.split(',',2)[2]
''' drop subjects from the group subject list '''
# check the path files in path_files_here folder in the subject's
# output folder - and drop any subjects from the group analysis
# subject list which do not exist in the paths to the output files
for path in s_paths:
if (c.repeatedMeasures == True):
if sub.count(',') == 1:
if (sub_id in path) and (other_id in path):
exist_paths.append(sub)
elif sub.count(',') == 2:
if (sub_id in path) and (scan_id in path) and \
(session_id in path):
exist_paths.append(sub)
else:
if sub in path:
exist_paths.append(sub)
# check to see if any derivatives of subjects are missing
if len(list(set(subject_list) - set(exist_paths))) >0:
print "List of outputs missing for subjects:"
print list(set(subject_list) - set(exist_paths))
print "..for derivatives:"
print resource
print "..at paths:"
print os.path.dirname(s_paths[0]).replace(s_ids[0], '*')
mod_path = os.path.join(os.path.dirname(s_paths[0]).replace(s_ids[0], 'group_analysis_results/_grp_model_%s'%(conf.model_name)),
'model_files')
print "basename: ", os.path.basename(conf.subject_list)
'''
try:
os.makedirs(mod_path)
print "Creating directory:"
print mod_path
except:
print "Attempted to create directory, but path already exists:"
print mod_path
'''
if not os.path.isdir(mod_path):
os.makedirs(mod_path)
''' write the new subject list '''
new_sub_file = os.path.join(mod_path, os.path.basename(conf.subject_list))
try:
f = open(new_sub_file, 'w')
for sub in exist_paths:
print >>f, sub
f.close()
except:
print "Error: Could not open subject list file: ", new_sub_file
raise Exception
conf.update('subjectListFile',new_sub_file)
sub_id = conf.subject_id_label
if measure_dict != None:
conf.update('phenotypicFile',get_phenotypic_file(conf.pheno_file, measure_dict, measure_list, mod_path, sub_id))
print 'conf updated pheno: ', conf.pheno_file, '\n\n'
print "Model config dictionary ->"
print conf.__dict__
# Run 'create_fsl_model' script to extract phenotypic data from
# the phenotypic file for each of the subjects in the subject list
''' get the motion statistics parameter file, if present '''
# get the parameter file so it can be passed to create_fsl_model.py
# so MeanFD or other measures can be included in the design matrix
parameter_file = os.path.join(c.outputDirectory, p_id[0], '%s_threshold_%s_all_params.csv'%(scan_ids[0].strip('_'),threshold_val))
if 1 in c.runGenerateMotionStatistics:
if not os.path.exists(parameter_file):
print '\n\n[!] CPAC says: Could not open the parameter file. ' \
'If Generate Motion Statistics is enabled, this can ' \
'usually be found in the output directory of your ' \
'individual-level analysis runs.\n'
print 'Path not found: ', parameter_file, '\n\n'
raise Exception
elif (1 not in c.runGenerateMotionStatistics) and (os.path.exists(parameter_file)):
if not os.path.exists(parameter_file):
print '\n\n[!] CPAC says: Could not open the parameter file. ' \
'If Generate Motion Statistics is enabled, this can ' \
'usually be found in the output directory of your ' \
'individual-level analysis runs.\n'
print 'Path not found: ', parameter_file, '\n\n'
raise Exception
else:
def no_measures_error(measure):
print '\n\n[!] CPAC says: The measure %s was included in ' \
'your group analysis design matrix formula, but ' \
'Generate Motion Statistics was not run during ' \
'individual-level analysis.\n' % measure
print 'Please run Generate Motion Statistics if you wish ' \
'to include this measure in your model.\n'
print 'If you HAVE completed a run with this option ' \
'enabled, then you are seeing this error because ' \
'the motion parameter file normally created by this ' \
'option is missing.\n\n'
raise Exception
for measure in measure_list:
if (measure in conf.design_formula):
no_measures_error(measure)
parameter_file = None
''' run create_fsl_model.py to generate the group analysis models '''
# path to the pipeline folder to be passed to create_fsl_model.py
# so that certain files like output_means.csv can be accessed
pipeline_path = os.path.join(c.outputDirectory, p_id[0])
# the current output that cpac_group_analysis_pipeline.py and
# create_fsl_model.py is currently being run for
current_output = s_paths[0].replace(pipeline_path, '').split('/')[2]
try:
from CPAC.utils import create_fsl_model
create_fsl_model.run(conf, c.fTest, parameter_file, pipeline_path, current_output, True)
#print >>diag, "> Runs create_fsl_model."
#print >>diag, ""
except Exception, e:
print "FSL Group Analysis model not successfully created - error in create_fsl_model script"
#print "Error ->", e
raise
model_sub_list.append((conf.output_dir, conf.subject_list))
def prep_group_analysis_workflow(c, group_config_file, resource, subject_infos, threshold_val):
#
# this function runs once per output file during group analysis
#
import yaml
import commands
# p_id = a list of pipeline IDs, i.e. the name of the output folder for
# the strat
# s_ids = a list of all the subject IDs
# scan_ids = a list of scan IDs
# s_paths = a list of all of the filepaths of this particular output
# file that prep_group_analysis_workflow is being called for
p_id, s_ids, scan_ids, s_paths = (list(tup) for tup in zip(*subject_infos))
try:
group_conf = Configuration(yaml.load(open(os.path.realpath(group_config_file), 'r')))
except Exception as e:
err_string = "\n\n[!] CPAC says: Could not read group model " \
"configuration YML file. Ensure you have read access " \
"for the file and that it is formatted properly.\n\n" \
"Configuration file: %s\n\nError details: %s" \
% (group_config_file, e)
raise Exception(err_string)
group_sublist_file = open(group_conf.subject_list, 'r')
group_sublist_items = group_sublist_file.readlines()
group_sublist = [line.rstrip('\n') for line in group_sublist_items \
if not (line == '\n') and not line.startswith('#')]
# list of subjects for which paths which DO exist
exist_paths = []
# paths to the actual derivatives for those subjects
derivative_paths = []
z_threshold = float(group_conf.z_threshold[0])
p_threshold = float(group_conf.p_threshold[0])
custom_confile = group_conf.custom_contrasts
if ((custom_confile == None) or (custom_confile == '') or \
("None" in custom_confile)):
if (len(group_conf.f_tests) == 0) or (group_conf.f_tests == None):
fTest = False
else:
fTest = True
else:
if not os.path.exists(custom_confile):
errmsg = "\n[!] CPAC says: You've specified a custom contrasts " \
".CSV file for your group model, but this file cannot " \
"be found. Please double-check the filepath you have " \
"entered.\n\nFilepath: %s\n\n" % custom_confile
raise Exception(errmsg)
evs = open(custom_confile, 'r').readline()
evs = evs.rstrip('\r\n').split(',')
count_ftests = 0
fTest = False
for ev in evs:
if "f_test" in ev:
count_ftests += 1
if count_ftests > 0:
fTest = True
''' begin iteration through group subject list for processing '''
print "Sorting through subject list to check for missing outputs " \
"for %s..\n" % resource
for ga_sub in group_sublist:
# Strip out carriage-return character if it is there
if ga_sub.endswith('\r'):
ga_sub = ga_sub.rstrip('\r')
# ga_sub = subject ID taken off the group analysis subject list
# let's check to make sure the subject list is formatted for
# repeated measures properly if repeated measures is enabled
# and vice versa
if (group_conf.repeated_measures == True) and (',' not in ga_sub):
print '\n\n'
print '[!] CPAC says: The group analysis subject list ' \
'is not in the appropriate format for repeated ' \
'measures.\n'
print 'Please use the appropriate format as described in ' \
'the CPAC User Guide or turn off Repeated Measures ' \
'in the CPAC pipeline configuration editor, found ' \
'in the \'Group Analysis Settings\' tab of the ' \
'pipeline configuration editor.\n'
print 'NOTE: CPAC generates a properly-formatted group ' \
'analysis subject list meant for running repeated ' \
'measures when you create your original subject ' \
'list. Look for \'subject_list_group_analysis_' \
'repeated_measures.txt\' in the directory where ' \
'you created your subject list.\n\n'
raise Exception
elif (group_conf.repeated_measures == False) and (',' in ga_sub):
print '\n\n'
print '[!] CPAC says: It looks like your group analysis ' \
'subject list is formatted for running repeated ' \
'measures, but \'Run Repeated Measures\' is not ' \
'enabled in the pipeline configuration, found in ' \
'the \'Group Analysis Settings\' tab of the ' \
'pipeline configuration editor.\n'
print 'Double-check your pipeline configuration?\n\n'
raise Exception
''' process subject ids for repeated measures, if it is on '''
# if repeated measures is being run and the subject list
# is a list of subject IDs and scan IDs concatenated
if (group_conf.repeated_measures == True):
# sub.count(',') equals 1 when there is either multiple scans
# or multiple sessions but not both, for repeated measures
# sub.count(',') equals 2 when there are multiple sessions
# AND scans, for repeated measures
if ga_sub.count(',') == 1:
sub_id = ga_sub.split(',',1)[0]
other_id = ga_sub.split(',',1)[1]
elif ga_sub.count(',') == 2:
sub_id = ga_sub.split(',',2)[0]
scan_id = ga_sub.split(',',2)[1]
session_id = ga_sub.split(',',2)[2]
''' drop subjects from the group subject list '''
# check the path files in path_files_here folder in the
# subject's output folder - and drop any subjects from the
# group analysis subject list which do not exist in the paths
# to the output files
'''
REVISIT THIS LATER to establish a potentially better way to
pull output paths (instead of path_files_here)
'''
for path in s_paths:
if (group_conf.repeated_measures == True):
if ga_sub.count(',') == 1:
if (sub_id in path) and (other_id in path):
exist_paths.append(ga_sub)
derivative_paths.append(path)
elif ga_sub.count(',') == 2:
if (sub_id in path) and (scan_id in path) and \
(session_id in path):
exist_paths.append(ga_sub)
derivative_paths.append(path)
else:
if ga_sub in path:
exist_paths.append(ga_sub)
derivative_paths.append(path)
# END subject-dropping!
if len(derivative_paths) == 0:
print '\n\n\n[!] CPAC says: None of the subjects listed in the ' \
'group analysis subject list were found to have outputs ' \
'produced by individual-level analysis.\n\nEnsure that ' \
'the subjects listed in your group analysis subject list ' \
'are the same as the ones included in the individual-' \
'level analysis you are running group-level analysis for.' \
'\n\n\n'
raise Exception
''' END subject list iteration '''
# check to see if any derivatives of subjects are missing
if len(list(set(group_sublist) - set(exist_paths))) >0:
print "List of outputs missing for subjects:"
print list(set(group_sublist) - set(exist_paths))
print "..for derivatives:"
print resource
print "..at paths:"
print os.path.dirname(s_paths[0]).replace(s_ids[0], '*')
# create the path string for the group analysis output
out_dir = os.path.dirname(s_paths[0]).split(p_id[0] + '/')
out_dir = os.path.join(group_conf.output_dir, out_dir[1])
out_dir = out_dir.replace(s_ids[0], 'group_analysis_results_%s/_grp_model_%s'%(p_id[0],group_conf.model_name))
model_out_dir = os.path.join(group_conf.output_dir, 'group_analysis_results_%s/_grp_model_%s'%(p_id[0],group_conf.model_name))
mod_path = os.path.join(out_dir, 'model_files')
if not os.path.isdir(mod_path):
os.makedirs(mod_path)
''' write the new subject list '''
new_sub_file = os.path.join(mod_path, os.path.basename(group_conf.subject_list))
try:
f = open(new_sub_file, 'w')
for sub in exist_paths:
print >>f, sub
f.close()
except:
print "Error: Could not open subject list file: ", new_sub_file
raise Exception
group_conf.update('subject_list',new_sub_file)
sub_id_label = group_conf.subject_id_label
# Run 'create_fsl_model' script to extract phenotypic data from
# the phenotypic file for each of the subjects in the subject list
''' get the motion statistics parameter file, if present '''
# get the parameter file so it can be passed to create_fsl_model.py
# so MeanFD or other measures can be included in the design matrix
measure_list = ['MeanFD', 'MeanFD_Jenkinson', 'MeanDVARS']
for measure in measure_list:
if (measure in group_conf.design_formula):
parameter_file = os.path.join(c.outputDirectory, p_id[0], '%s%s_all_params.csv'%(scan_ids[0].strip('_'),threshold_val))
if 1 in c.runGenerateMotionStatistics:
if not os.path.exists(parameter_file):
print '\n\n[!] CPAC says: Could not find or open the motion ' \
'parameter file. This is necessary if you have included ' \
'any of the MeanFD measures in your group model.\n\n' \
'If Generate Motion Statistics is enabled, this file can ' \
'usually be found in the output directory of your ' \
'individual-level analysis runs. If it is not there, ' \
'double-check to see if individual-level analysis had ' \
'completed successfully.\n'
print 'Path not found: ', parameter_file, '\n\n'
raise Exception
else:
def no_measures_error(measure):
print '\n\n[!] CPAC says: The measure %s was included in ' \
'your group analysis design matrix formula, but ' \
'Generate Motion Statistics was not run during ' \
'individual-level analysis.\n' % measure
print 'Please run Generate Motion Statistics if you wish ' \
'to include this measure in your model.\n'
print 'If you HAVE completed a run with this option ' \
'enabled, then you are seeing this error because ' \
'the motion parameter file normally created by this ' \
'option is missing.\n\n'
raise Exception
for measure in measure_list:
if (measure in group_conf.design_formula):
no_measures_error(measure)
parameter_file = None
break
else:
parameter_file = None
# path to the pipeline folder to be passed to create_fsl_model.py
# so that certain files like output_means.csv can be accessed
pipeline_path = os.path.join(c.outputDirectory, p_id[0])
# the current output that cpac_group_analysis_pipeline.py and
# create_fsl_model.py is currently being run for
current_output = resource #s_paths[0].replace(pipeline_path, '').split('/')[2]
# generate working directory for this output's group analysis run
workDir = '%s/group_analysis/%s/%s_%s' % (c.workingDirectory, group_conf.model_name, resource, scan_ids[0])
# s_paths is a list of paths to each subject's derivative (of the current
# derivative gpa is being run on) - s_paths_dirList is a list of each directory
# in this path separated into list elements
# this makes strgy_path basically the directory path of the folders after
# the scan ID folder level
strgy_path = os.path.dirname(s_paths[0]).split(scan_ids[0])[1]
# get rid of periods in the path
for ch in ['.']:
if ch in strgy_path:
strgy_path = strgy_path.replace(ch, "")
# create nipype-workflow-name-friendly strgy_path
# (remove special characters)
strgy_path_name = strgy_path.replace('/', "_")
workDir = workDir + '/' + strgy_path_name
''' merge the remaining subjects for this current output '''
# then, take the group mask, and iterate over the list of subjects
# remaining to extract the mean of each subject using the group
# mask
merge_input = " "
merge_output_dir = workDir + "/merged_files"
if not os.path.exists(merge_output_dir):
os.makedirs(merge_output_dir)
merge_output = merge_output_dir + "/" + current_output + "_merged.nii.gz"
merge_mask_output = merge_output_dir + "/" + current_output + "_merged_mask.nii.gz"
# create a string per derivative filled with every subject's path to the
# derivative output file
for derivative_path in derivative_paths:
merge_input = merge_input + " " + derivative_path
merge_string = "fslmerge -t %s %s" % (merge_output, merge_input)
# MERGE the remaining outputs
try:
commands.getoutput(merge_string)
except Exception as e:
print "[!] CPAC says: FSL Merge failed for output: %s" % current_output
print "Error details: %s\n\n" % e
raise
merge_mask_string = "fslmaths %s -abs -Tmin -bin %s" % (merge_output, merge_mask_output)
# CREATE A MASK of the merged file
try:
commands.getoutput(merge_mask_string)
except Exception as e:
print "[!] CPAC says: FSL Mask failed for output: %s" % current_output
print "Error details: %s\n\n" % e
raise
derivative_means_dict = {}
roi_means_dict = {}
# CALCULATE THE MEANS of each remaining output using the group mask
for derivative_path in derivative_paths:
try:
if "Group Mask" in group_conf.mean_mask:
maskave_output = commands.getoutput("3dmaskave -mask %s %s" % (merge_mask_output, derivative_path))
elif "Individual Mask" in group_conf.mean_mask:
maskave_output = commands.getoutput("3dmaskave -mask %s %s" % (derivative_path, derivative_path))
except Exception as e:
print "[!] CPAC says: AFNI 3dmaskave failed for output: %s\n" \
"(Measure Mean calculation)" % current_output
print "Error details: %s\n\n" % e
raise
# get the subject ID of the current derivative path reliably
derivative_path_subID = derivative_path.replace(pipeline_path,"").strip("/").split("/")[0]
# this crazy-looking command simply extracts the mean from the
# verbose AFNI 3dmaskave output string
derivative_means_dict[derivative_path_subID] = maskave_output.split("\n")[-1].split(" ")[0]
# derivative_means_dict is now something like this:
# { 'sub001': 0.3124, 'sub002': 0.2981, .. }
# if custom ROI means are included in the model, do the same for those
if "Custom_ROI_Mean" in group_conf.design_formula:
try:
if "centrality" in derivative_path:
# resample custom roi mask to 3mm, then use that
resampled_roi_mask = merge_output_dir + "/" + current_output + "_resampled_roi_mask.nii.gz"
commands.getoutput("flirt -in %s -ref %s -o %s -applyxfm -init %s -interp nearestneighbour" % (group_conf.custom_roi_mask, derivative_path, resampled_roi_mask, c.identityMatrix))
ROIstats_output = commands.getoutput("3dROIstats -mask %s %s" % (resampled_roi_mask, derivative_path))
else:
ROIstats_output = commands.getoutput("3dROIstats -mask %s %s" % (group_conf.custom_roi_mask, derivative_path))
except Exception as e:
print "[!] CPAC says: AFNI 3dROIstats failed for output: %s" \
"\n(Custom ROI Mean calculation)" % current_output
print "Error details: %s\n\n" % e
raise
ROIstats_list = ROIstats_output.split("\t")
# calculate the number of ROIs - 3dROIstats output can be split
# into a list, and the actual ROI means begin at a certain point
num_rois = (len(ROIstats_list)-3)/2
roi_means = []
# create a list of the ROI means - each derivative of each subject
# will have N number of ROIs depending on how many ROIs were
# specified in the custom ROI mask
for num in range(num_rois+3,len(ROIstats_list)):
roi_means.append(ROIstats_list[num])
roi_means_dict[derivative_path_subID] = roi_means
else:
roi_means_dict = None
if len(derivative_means_dict.keys()) == 0:
err_string = "[!] CPAC says: Something went wrong with the " \
"calculation of the output means via the group mask.\n\n"
raise Exception(err_string)
''' run create_fsl_model.py to generate the group analysis models '''
from CPAC.utils import create_fsl_model
create_fsl_model.run(group_conf, fTest, parameter_file, derivative_means_dict, pipeline_path, current_output, model_out_dir, roi_means_dict, True)
''' begin GA workflow setup '''
if not os.path.exists(new_sub_file):
raise Exception("path to input subject list %s is invalid" % new_sub_file)
#if c.mixedScanAnalysis == True:
# wf = pe.Workflow(name = 'group_analysis/%s/grp_model_%s'%(resource, os.path.basename(model)))
#else:
wf = pe.Workflow(name = resource)
wf.base_dir = workDir
wf.config['execution'] = {'hash_method': 'timestamp', 'crashdump_dir': os.path.abspath(c.crashLogDirectory)}
log_dir = os.path.join(group_conf.output_dir, 'logs', 'group_analysis', resource, 'model_%s' % (group_conf.model_name))
if not os.path.exists(log_dir):
os.makedirs(log_dir)
else:
pass
# gp_flow
# Extracts the model files (.con, .grp, .mat, .fts) from the model
# directory and sends them to the create_group_analysis workflow gpa_wf
gp_flow = create_grp_analysis_dataflow("gp_dataflow_%s" % resource)
gp_flow.inputs.inputspec.grp_model = os.path.join(model_out_dir, "model_files", current_output)
gp_flow.inputs.inputspec.model_name = group_conf.model_name
gp_flow.inputs.inputspec.ftest = fTest
# gpa_wf
# Creates the actual group analysis workflow
gpa_wf = create_group_analysis(fTest, "gp_analysis_%s" % resource)
gpa_wf.inputs.inputspec.merged_file = merge_output
gpa_wf.inputs.inputspec.merge_mask = merge_mask_output
gpa_wf.inputs.inputspec.z_threshold = z_threshold
gpa_wf.inputs.inputspec.p_threshold = p_threshold
gpa_wf.inputs.inputspec.parameters = (c.FSLDIR, 'MNI152')
wf.connect(gp_flow, 'outputspec.mat',
gpa_wf, 'inputspec.mat_file')
wf.connect(gp_flow, 'outputspec.con',
gpa_wf, 'inputspec.con_file')
wf.connect(gp_flow, 'outputspec.grp',
gpa_wf, 'inputspec.grp_file')
if fTest:
wf.connect(gp_flow, 'outputspec.fts',
gpa_wf, 'inputspec.fts_file')
# ds
# Creates the datasink node for group analysis
ds = pe.Node(nio.DataSink(), name='gpa_sink')
if 'sca_roi' in resource:
out_dir = os.path.join(out_dir, \
re.search('sca_roi_(\d)+',os.path.splitext(os.path.splitext(os.path.basename(s_paths[0]))[0])[0]).group(0))
if 'dr_tempreg_maps_zstat_files_to_standard_smooth' in resource:
out_dir = os.path.join(out_dir, \
re.search('temp_reg_map_z_(\d)+',os.path.splitext(os.path.splitext(os.path.basename(s_paths[0]))[0])[0]).group(0))
if 'centrality' in resource:
names = ['degree_centrality_binarize', 'degree_centrality_weighted', \
'eigenvector_centrality_binarize', 'eigenvector_centrality_weighted', \
'lfcd_binarize', 'lfcd_weighted']
for name in names:
if name in os.path.basename(s_paths[0]):
out_dir = os.path.join(out_dir, name)
break
if 'tempreg_maps' in resource:
out_dir = os.path.join(out_dir, \
re.search('\w*[#]*\d+', os.path.splitext(os.path.splitext(os.path.basename(s_paths[0]))[0])[0]).group(0))
# if c.mixedScanAnalysis == True:
# out_dir = re.sub(r'(\w)*scan_(\w)*(\d)*(\w)*[/]', '', out_dir)
ds.inputs.base_directory = out_dir
ds.inputs.container = ''
ds.inputs.regexp_substitutions = [(r'(?<=rendered)(.)*[/]','/'),
(r'(?<=model_files)(.)*[/]','/'),
(r'(?<=merged)(.)*[/]','/'),
(r'(?<=stats/clusterMap)(.)*[/]','/'),
(r'(?<=stats/unthreshold)(.)*[/]','/'),
(r'(?<=stats/threshold)(.)*[/]','/'),
(r'_cluster(.)*[/]',''),
(r'_slicer(.)*[/]',''),
(r'_overlay(.)*[/]','')]
'''
if 1 in c.runSymbolicLinks:
link_node = pe.MapNode(interface=util.Function(
input_names=['in_file',
'resource'],
output_names=[],
function=prepare_gp_links),
name='link_gp_', iterfield=['in_file'])
link_node.inputs.resource = resource
wf.connect(ds, 'out_file', link_node, 'in_file')
'''
########datasink connections#########
if fTest:
wf.connect(gp_flow, 'outputspec.fts',
ds, 'model_files.@0')
wf.connect(gp_flow, 'outputspec.mat',
ds, 'model_files.@1' )
wf.connect(gp_flow, 'outputspec.con',
ds, 'model_files.@2')
wf.connect(gp_flow, 'outputspec.grp',
ds, 'model_files.@3')
wf.connect(gpa_wf, 'outputspec.merged',
ds, 'merged')
wf.connect(gpa_wf, 'outputspec.zstats',
ds, 'stats.unthreshold')
wf.connect(gpa_wf, 'outputspec.zfstats',
ds,'stats.unthreshold.@01')
wf.connect(gpa_wf, 'outputspec.fstats',
ds,'stats.unthreshold.@02')
wf.connect(gpa_wf, 'outputspec.cluster_threshold_zf',
ds, 'stats.threshold')
wf.connect(gpa_wf, 'outputspec.cluster_index_zf',
ds,'stats.clusterMap')
wf.connect(gpa_wf, 'outputspec.cluster_localmax_txt_zf',
ds, 'stats.clusterMap.@01')
wf.connect(gpa_wf, 'outputspec.overlay_threshold_zf',
ds, 'rendered')
wf.connect(gpa_wf, 'outputspec.rendered_image_zf',
ds, 'rendered.@01')
wf.connect(gpa_wf, 'outputspec.cluster_threshold',
ds, 'stats.threshold.@01')
wf.connect(gpa_wf, 'outputspec.cluster_index',
ds, 'stats.clusterMap.@02')
wf.connect(gpa_wf, 'outputspec.cluster_localmax_txt',
ds, 'stats.clusterMap.@03')
wf.connect(gpa_wf, 'outputspec.overlay_threshold',
ds, 'rendered.@02')
wf.connect(gpa_wf, 'outputspec.rendered_image',
ds, 'rendered.@03')
######################################
# Run the actual group analysis workflow
wf.run()
'''
except:
print "Error: Group analysis workflow run command did not complete successfully."
print "subcount: ", subcount
print "pathcount: ", pathcount
print "sublist: ", sublist_items
print "input subject list: "
print "conf: ", conf.subjectListFile
raise Exception
'''
print "**Workflow finished for model %s and resource %s"%(os.path.basename(group_conf.output_dir), resource)
def testConfig(self, event):
'''
This function runs when the user clicks the "Test Configuration"
button in the pipeline configuration window.
It prompts the user for a sample subject list (i.e. one that they will
be using with the config they are building). Then it builds the
pipeline but does not run it. It then reports whether or not the
config will run or not depending on if the pipeline gets built
successfully.
'''
import os
import yaml
from CPAC.utils import Configuration
from CPAC.pipeline.cpac_pipeline import prep_workflow
from CPAC.pipeline.cpac_runner import build_strategies
def display(win, msg, changeBg=True):
wx.MessageBox(msg, "Error")
if changeBg:
win.SetBackgroundColour("pink")
win.SetFocus()
win.Refresh()
# Collect a sample subject list and parse it in
testDlg0 = wx.MessageDialog(
self, 'This tool will run a quick check on the current pipeline configuration.' \
' Click OK to provide a subject list you will be using with this setup.',
'Subject List',
wx.OK | wx.ICON_INFORMATION)
testDlg0.ShowModal()
testDlg0.Destroy()
dlg = wx.FileDialog(self,
message="Choose the CPAC Subject list file",
defaultDir=os.getcwd(),
defaultFile="CPAC_subject_list.yml",
wildcard="YAML files(*.yaml, *.yml)|*.yaml;*.yml",
style=wx.OPEN | wx.CHANGE_DIR)
if dlg.ShowModal() == wx.ID_OK:
subListPath = dlg.GetPath()
sublist = yaml.load(open(os.path.realpath(subListPath), 'r'))
# Check to ensure the user is providing an actual subject
# list and not some other kind of file
try:
subInfo = sublist[0]
except:
errDlg4 = wx.MessageDialog(
self, 'ERROR: Subject list file not in proper format - check if you' \
' loaded the correct file? \n\n' \
'Error name: config_window_0001',
'Subject List Error',
wx.OK | wx.ICON_ERROR)
errDlg4.ShowModal()
errDlg4.Destroy()
raise Exception
# Another check to ensure the actual subject list was generated
# properly and that it will work
if 'subject_id' not in subInfo:
errDlg3 = wx.MessageDialog(
self, 'ERROR: Subject list file not in proper format - check if you' \
' loaded the correct file? \n\n' \
'Error name: config_window_0002',
'Subject List Error',
wx.OK | wx.ICON_ERROR)
errDlg3.ShowModal()
errDlg3.Destroy()
raise Exception
# Following code reads in the parameters and selections from the
# pipeline configuration window and populate the config_list
config_list = []
wf_counter = []
for page in self.nb.get_page_list():
switch = page.page.get_switch()
ctrl_list = page.page.get_ctrl_list()
validate = False
if switch:
switch_val = str(switch.get_selection()).lower()
if switch_val == 'on' or switch_val == 'true' or switch_val == '1':
validate = True
wf_counter.append(page.get_counter())
for ctrl in ctrl_list:
# option_name will be the selection name as it is written
# as the dictionary key of the config.yml dictionary
option_name = ctrl.get_name()
#validating
if (switch == None or validate) and ctrl.get_validation() \
and (option_name != 'derivativeList') and (option_name != 'modelConfigs'):
win = ctrl.get_ctrl()
if isinstance(ctrl.get_selection(), list):
value = ctrl.get_selection()
if not value:
display(
win,
"%s field is empty or the items are not checked!"
% ctrl.get_name(), False)
return
else:
value = str(ctrl.get_selection())
if len(value) == 0:
display(win, "%s field is empty!" % ctrl.get_name())
return
if '/' in value and '$' not in value and not isinstance(
value, list):
if not os.path.exists(
ctrl.get_selection()) and value != 'On/Off':
display(
win,
"%s field contains incorrect path. Please update the path!"
% ctrl.get_name())
return
config_list.append(ctrl)
# Get the user's CPAC output directory for use in this script
for config in config_list:
if config.get_name() == 'outputDirectory':
outDir = config.get_selection()
# Write out a pipeline_config file, read it in and then delete it
# (Will revise the data structure of the config files later so this
# can just pass the data structure instead of doing it this way)
try:
self.write(outDir + 'testConfig.yml', config_list)
c = Configuration(
yaml.load(
open(os.path.realpath(outDir + 'testConfig.yml'), 'r')))
os.remove(outDir + 'testConfig.yml')
except:
errDlg2 = wx.MessageDialog(
self, 'A problem occurred with preparing the pipeline test run. \n\n' \
'Please ensure you have rights access to the directories you' \
' have chosen for the CPAC working, crash, and output folders.',
'Test Configuration Error',
wx.OK | wx.ICON_ERROR)
errDlg2.ShowModal()
errDlg2.Destroy()
if (1 in c.runNuisance) or (c.Corrections != None):
strategies = sorted(build_strategies(c))
else:
strategies = None
# Run the actual pipeline building prep and see if it works or not
testDlg1 = wx.MessageDialog(
self,
'Click OK to run the test. This should take only a few seconds.',
'Running Test', wx.OK | wx.ICON_INFORMATION)
testDlg1.ShowModal()
# Check file paths first
# Just getting proper names of config file parameters
try:
params_file = open(
p.resource_filename('CPAC',
'GUI/resources/config_parameters.txt'),
"r")
except:
print "Error: Could not open configuration parameter file.", "\n"
raise Exception
paramInfo = params_file.read().split('\n')
paramList = []
for param in paramInfo:
if param != '':
paramList.append(param.split(','))
# function for file path checking
def testFile(filepath, paramName):
try:
if filepath != None:
fileTest = open(filepath)
fileTest.close()
except:
testDlg1.Destroy()
for param in paramList:
if param[0] == paramName:
paramTitle = param[1]
paramGroup = param[2]
break
errDlgFileTest = wx.MessageDialog(
self, 'Error reading file - either it does not exist or you' \
' do not have read access. \n\n' \
'Parameter: %s \n' \
'In tab: %s \n\n' \
'Path: %s' % (paramTitle, paramGroup, filepath),
'Pipeline Not Ready',
wx.OK | wx.ICON_ERROR)
errDlgFileTest.ShowModal()
errDlgFileTest.Destroy()
testFile(c.template_brain_only_for_anat,
'template_brain_only_for_anat')
testFile(c.template_skull_for_anat, 'template_skull_for_anat')
testFile(c.PRIORS_WHITE, 'PRIORS_WHITE')
testFile(c.PRIORS_GRAY, 'PRIORS_GRAY')
testFile(c.PRIORS_CSF, 'PRIORS_CSF')
testFile(c.template_brain_only_for_func,
'template_brain_only_for_func')
testFile(c.template_skull_for_func, 'template_skull_for_func')
testFile(c.identityMatrix, 'identityMatrix')
testFile(c.boundaryBasedRegistrationSchedule,
'boundaryBasedRegistrationSchedule')
testFile(c.lateral_ventricles_mask, 'lateral_ventricles_mask')
testFile(c.seedSpecificationFile, 'seedSpecificationFile')
testFile(c.roiSpecificationFile, 'roiSpecificationFile')
testFile(c.roiSpecificationFileForSCA, 'roiSpecificationFileForSCA')
testFile(c.maskSpecificationFile, 'maskSpecificationFile')
testFile(c.maskSpecificationFileForSCA, 'maskSpecificationFileForSCA')
testFile(c.spatialPatternMaps, 'spatialPatternMaps')
testFile(c.template_symmetric_brain_only,
'template_symmetric_brain_only')
testFile(c.template_symmetric_skull, 'template_symmetric_skull')
testFile(c.dilated_symmetric_brain_mask,
'dilated_symmetric_brain_mask')
testFile(c.configFileTwomm, 'configFileTwomm')
testFile(c.templateSpecificationFile, 'templateSpecificationFile')
testFile(c.bascAffinityThresholdFile, 'bascAffinityThresholdFile')
testFile(c.cwasROIFile, 'cwasROIFile')
testFile(c.cwasRegressorFile, 'cwasRegressorFile')
try:
# Run the pipeline building
prep_workflow(sublist[0], c, strategies, 0)
except:
testDlg1.Destroy()
errDlg1 = wx.MessageDialog(
self, 'There are issues with the current configuration which need to be' \
' resolved - please check to make sure the options you are running' \
' have the proper pre-requisites selected.',
'Pipeline Not Ready',
wx.OK | wx.ICON_ERROR)
errDlg1.ShowModal()
errDlg1.Destroy()
else:
testDlg1.Destroy()
okDlg1 = wx.MessageDialog(
self, 'The current configuration will run successfully. You can safely' \
' save and run this setup!',
'Pipeline Ready',
wx.OK | wx.ICON_INFORMATION)
okDlg1.ShowModal()
okDlg1.Destroy()
def prep_group_analysis_workflow(c, resource, subject_infos):
p_id, s_ids, scan_ids, s_paths = (list(tup) for tup in zip(*subject_infos))
#print "p_id -%s, s_ids -%s, scan_ids -%s, s_paths -%s" %(p_id, s_ids, scan_ids, s_paths)
def get_phenotypic_file(phenotypic_file, m_dict, m_list, mod_path, sub_id):
#print "phenotypic_file, m_dict", phenotypic_file, m_dict
import csv
reader = csv.reader(open(phenotypic_file, 'rU'))
columns = {}
order = {}
count = 0
headers = reader.next()
for h in headers:
columns[h] = []
order[h] = count
count += 1
for r in reader:
for h, v in zip(headers, r):
if v:
columns[h].append(str(v))
if m_dict:
for measure in m_list:
if measure in headers:
#check if 'MeanFD is present'
if len(columns[measure]) < 1:
for sub in columns[sub_id]:
if m_dict.get(sub):
if m_dict.get(sub).get(measure):
columns[measure].append(
m_dict[sub][measure])
else:
raise Exception(
"Couldn't find %s value for subject %s"
% (measure, sub))
else:
raise Exception(
"Couldn't find subject %s in the parameter file"
% sub)
b = zip(*([k] + columns[k] for k in sorted(columns, key=order.get)))
try:
os.makedirs(mod_path)
except:
print "%s already exist" % (mod_path)
new_phenotypic_file = os.path.join(mod_path,
os.path.basename(phenotypic_file))
a = csv.writer(open(new_phenotypic_file, 'w'))
for col in b:
a.writerow(list(col))
return new_phenotypic_file
threshold_val = None
measure_dict = None
measure_list = ['MeanFD', 'MeanFD_Jenkinson', 'MeanDVARS']
model_sub_list = []
if c.runScrubbing == 1:
#get scrubbing threshold
if re.search('(?<=/_threshold_)\d+.\d+', s_paths[0]):
threshold_val = re.search('(?<=/_threshold_)\d+.\d+',
s_paths[0]).group(0)
elif len(c.scrubbingThreshold) == 1:
threshold_val = c.scrubbingThreshold[0]
else:
print("Found Multiple threshold value ")
print "scrubbing threshold_val -->", threshold_val
else:
print "No scrubbing enabled."
print "\n"
#pick the right parameter file from the pipeline folder
#create a dictionary of subject and measures in measure_list
if c.runScrubbing == 1:
try:
parameter_file = os.path.join(
c.outputDirectory, p_id[0], '%s_threshold_%s_all_params.csv' %
(scan_ids[0].strip('_'), threshold_val))
if os.path.exists(parameter_file):
import csv
measure_dict = {}
f = csv.DictReader(open(parameter_file, 'r'))
for line in f:
measure_map = {}
for m in measure_list:
if line.get(m):
measure_map[m] = line[m]
measure_dict[line['Subject']] = measure_map
else:
print "No file name %s found" % parameter_file
except Exception:
print "Exception while extracting parameters from movement file - %s" % (
parameter_file)
for config in c.modelConfigs:
import yaml
try:
conf = Configuration(yaml.load(open(os.path.realpath(config),
'r')))
except:
raise Exception("Error in reading %s configuration file" % config)
subject_list = [line.rstrip('\r\n') for line in open(conf.subjectListFile, 'r') \
if not (line == '\n') and not line.startswith('#')]
# list of subject paths which DO exist
exist_paths = []
# check for missing subject for the derivative
for sub in subject_list:
for path in s_paths:
if sub in path:
exist_paths.append(sub)
# check to see if any derivatives of subjects are missing
if len(list(set(subject_list) - set(exist_paths))) > 0:
print "-------------------------------------------"
print "List of outputs missing for subjects:"
print list(set(subject_list) - set(exist_paths))
print "\n"
print "..for derivatives:"
print resource
print "\n"
print "..at paths:"
print os.path.dirname(s_paths[0]).replace(s_ids[0], '*')
print "-------------------------------------------"
print '\n'
#import warnings
#warnings.warn(msg)
mod_path = os.path.join(
os.path.dirname(s_paths[0]).replace(
s_ids[0],
'group_analysis_results/_grp_model_%s' % (conf.modelName)),
'model_files')
print "basename: ", os.path.basename(conf.subjectListFile)
try:
os.makedirs(mod_path)
print "Creating directory:"
print mod_path
print "\n"
except:
print "Attempted to create directory, but path already exists:"
print mod_path
print '\n'
new_sub_file = os.path.join(mod_path,
os.path.basename(conf.subjectListFile))
try:
f = open(new_sub_file, 'w')
for sub in exist_paths:
print >> f, sub
f.close()
except:
print "Error: Could not open subject list file: ", new_sub_file
print ""
raise Exception
conf.update('subjectListFile', new_sub_file)
sub_id = conf.subjectColumn
if measure_dict != None:
conf.update(
'phenotypicFile',
get_phenotypic_file(conf.phenotypicFile, measure_dict,
measure_list, mod_path, sub_id))
print "Model config dictionary ->"
print conf.__dict__
print '\n'
# Run 'create_fsl_model' script to extract phenotypic data from
# the phenotypic file for each of the subjects in the subject list
try:
from CPAC.utils import create_fsl_model
create_fsl_model.run(conf, True)
except Exception, e:
print "Error in creating models in the create_fsl_model script"
#print "Error ->", e
raise
model_sub_list.append(
(conf.outputModelFilesDirectory, conf.subjectListFile))
print "model_sub_list ->", model_sub_list
def run(config_file, subject_list_file, p_name = None):
# take date+time stamp for run identification purposes
unique_pipeline_id = strftime("%Y%m%d%H%M%S")
pipeline_start_stamp = strftime("%Y-%m-%d_%H:%M:%S")
try:
if not os.path.exists(config_file):
raise IOError
else:
c = Configuration(yaml.load(open(os.path.realpath(config_file), 'r')))
except IOError:
print "config file %s doesn't exist" % config_file
raise
except Exception:
print "Error reading config file - %s" % config_file
raise Exception
#do some validation
validate(c)
try:
sublist = yaml.load(open(os.path.realpath(subject_list_file), 'r'))
except:
print "Subject list is not in proper YAML format. Please check your file"
raise Exception
# NOTE: strategies list is only needed in cpac_pipeline prep_workflow for
# creating symlinks
strategies = sorted(build_strategies(c))
print "strategies ---> "
print strategies
sub_scan_map ={}
print "subject list: "
print sublist
try:
for sub in sublist:
if sub['unique_id']:
s = sub['subject_id']+"_" + sub["unique_id"]
else:
s = sub['subject_id']
scan_ids = ['scan_anat']
for id in sub['rest']:
scan_ids.append('scan_'+ str(id))
sub_scan_map[s] = scan_ids
except:
print "\n\n" + "ERROR: Subject list file not in proper format - check if you loaded the correct file?" + "\n" + \
"Error name: cpac_runner_0001" + "\n\n"
raise Exception
create_group_log_template(sub_scan_map, os.path.join(c.outputDirectory, 'logs'))
seeds_created = []
if not (c.seedSpecificationFile is None):
try:
if os.path.exists(c.seedSpecificationFile):
seeds_created = create_seeds_(c.seedOutputLocation, c.seedSpecificationFile, c.FSLDIR)
print 'seeds created %s -> ' % seeds_created
except:
raise IOError('Problem in seedSpecificationFile')
if 1 in c.runVoxelTimeseries:
if 'roi_voxelwise' in c.useSeedInAnalysis:
c.maskSpecificationFile = append_seeds_to_file(c.workingDirectory, seeds_created, c.maskSpecificationFile)
if 1 in c.runROITimeseries:
if 'roi_average' in c.useSeedInAnalysis:
c.roiSpecificationFile = append_seeds_to_file(c.workingDirectory, seeds_created, c.roiSpecificationFile)
if 1 in c.runNetworkCentrality:
if 'centrality_outputs_smoothed' in c.useSeedInAnalysis:
c.templateSpecificationFile = append_seeds_to_file(c.workingDirectory, seeds_created, c.templateSpecificationFile)
pipeline_timing_info = []
pipeline_timing_info.append(unique_pipeline_id)
pipeline_timing_info.append(pipeline_start_stamp)
pipeline_timing_info.append(len(sublist))
if not c.runOnGrid:
from CPAC.pipeline.cpac_pipeline import prep_workflow
procss = [Process(target=prep_workflow, args=(sub, c, strategies, 1, pipeline_timing_info, p_name)) for sub in sublist]
pid = open(os.path.join(c.outputDirectory, 'pid.txt'), 'w')
jobQueue = []
if len(sublist) <= c.numSubjectsAtOnce:
"""
Stream all the subjects as sublist is
less than or equal to the number of
subjects that need to run
"""
for p in procss:
p.start()
print >>pid,p.pid
else:
"""
Stream the subject workflows for preprocessing.
At Any time in the pipeline c.numSubjectsAtOnce
will run, unless the number remaining is less than
the value of the parameter stated above
"""
idx = 0
while(idx < len(sublist)):
if len(jobQueue) == 0 and idx == 0:
idc = idx
for p in procss[idc: idc + c.numSubjectsAtOnce]:
p.start()
print >>pid,p.pid
jobQueue.append(p)
idx += 1
else:
for job in jobQueue:
if not job.is_alive():
print 'found dead job ', job
loc = jobQueue.index(job)
del jobQueue[loc]
procss[idx].start()
jobQueue.append(procss[idx])
idx += 1
pid.close()
else:
import commands
import pickle
temp_files_dir = os.path.join(os.getcwd(), 'cluster_temp_files')
print commands.getoutput("mkdir -p %s" % temp_files_dir)
strategies_file = os.path.join(temp_files_dir, 'strategies.obj')
f = open(strategies_file, 'w')
pickle.dump(strategies, f)
f.close()
if 'sge' in c.resourceManager.lower():
run_sge_jobs(c, config_file, strategies_file, subject_list_file, p_name)
elif 'pbs' in c.resourceManager.lower():
run_pbs_jobs(c, config_file, strategies_file, subject_list_file, p_name)
elif 'condor' in c.resourceManager.lower():
run_condor_jobs(c, config_file, strategies_file, subject_list_file, p_name)
def AddConfig(self, event):
'''
docstring
'''
# Gets called when you click 'Load' for pipeline config in the GUI
dlg = wx.FileDialog(self,
message="Choose the CPAC Configuration file",
defaultDir=os.getcwd(),
defaultFile="",
wildcard="YAML files(*.yaml, *.yml)|*.yaml;*.yml",
style=wx.OPEN | wx.CHANGE_DIR)
# User clicks "OK"
if dlg.ShowModal() == wx.ID_OK:
# Load config file into memory and verify its not a subject list
path = dlg.GetPath()
# Check for path existence
if os.path.exists(path):
path = os.path.realpath(path)
try:
f_cfg = yaml.load(open(path, 'r'))
except Exception as e:
print '\n\nUnable to load the specified file: %s' % path
print "The YAML file may not be formatted properly."
print 'Error:\n%s\n\n' % e
raise Exception
if type(f_cfg) == dict:
if not f_cfg.has_key('pipelineName'):
err_msg = 'File is not a pipeline configuration '\
'file. It might be a data configuration file.'
raise Exception(err_msg)
else:
err_msg = 'File is not a pipeline configuration '\
'file. It might be a subject list file.'
raise Exception(err_msg)
# Otherwise, report error
else:
err_msg = 'File %s does not exist. Check and try again.' % path
raise Exception(err_msg)
# If config file is ok, proceed to load
if self.check_config(path) > 0:
while True:
try:
c = Configuration(f_cfg)
except Exception as e:
if "object has no attribute" in e:
err = "%s\n\nIs this attribute linked (using " \
"'${}') in any of your configuration " \
"parameters? (Standard template paths, " \
"for example). If this is a pipeline " \
"configuration file from an older version "\
"of CPAC, this parameter may be obsolete. "\
"Double-check your selections.\n\n" % e
print err
else:
print '\n\nERROR: Configuration file could not ' \
'be loaded properly - the file might be '\
'access-protected or you might have ' \
'chosen the wrong file.\n'
print 'Error name: main_window_0001\n\n'
print 'Exception: %s' % e
# Valid pipeline name
if c.pipelineName != None:
if self.pipeline_map.get(c.pipelineName) == None:
# this runs if you click 'Load' on the main
# CPAC window, enter a path, and the pipeline
# name attribute of the pipeline config file
# you are loading does NOT already exist in
# the listbox, i.e., the proper condition
self.pipeline_map[str(c.pipelineName)] = path
self.listbox.Append(str(c.pipelineName))
dlg.Destroy()
break
else:
# this runs if you click 'Load' on the main
# CPAC window, enter a path, and the pipeline
# name attribute of the pipeline config file
# you are loading DOES already exist in
# the listbox, which is a conflict
dlg3 = wx.MessageDialog(self, 'The \'' \
'Pipeline Name\' attribute of the ' \
'configuration file you are loading' \
' already exists in one of the' \
' configuration files listed under' \
' \'Pipelines\'.\n\nPlease change' \
' the pipeline name attribute (not' \
' the filename) from within the' \
' pipeline editor (under the' \
' \'Output Settings\' tab in' \
' \'Environment Setup\'), or load a' \
' new configuration file.\n\n' \
'Pipeline configuration with' \
' conflicting name:\n%s' \
% c.pipelineName,
'Conflicting Pipeline Names',
wx.OK | wx.ICON_ERROR)
dlg3.ShowModal()
dlg3.Destroy()
break
# Pipeline name is None
else:
dlg4 = wx.MessageDialog(self, 'Warning: Pipeline name is blank.\n\nPlease edit' \
' the pipeline_config.yml file in a text editor and' \
' restore the pipelineName field.',
'Warning',
wx.OK | wx.ICON_ERROR)
dlg4.ShowModal()
dlg4.Destroy()
dlg.Destroy
break
def testConfig(self, event):
'''
This function runs when the user clicks the "Test Configuration"
button in the pipeline configuration window.
It prompts the user for a sample subject list (i.e. one that they will
be using with the config they are building). Then it builds the
pipeline but does not run it. It then reports whether or not the
config will run or not depending on if the pipeline gets built
successfully.
'''
# Import packages
import os
import yaml
from CPAC.utils import Configuration
from CPAC.pipeline.cpac_pipeline import prep_workflow
from CPAC.pipeline.cpac_runner import build_strategies
def display(win, msg, changeBg=True):
wx.MessageBox(msg, "Error")
if changeBg:
win.SetBackgroundColour("pink")
win.SetFocus()
win.Refresh()
# Collect a sample subject list and parse it in
testDlg0 = wx.MessageDialog(
self, 'This tool will run a quick check on the current pipeline '\
'configuration. Click OK to provide a subject list you ' \
'will be using with this setup.',
'Subject List',
wx.OK | wx.ICON_INFORMATION)
testDlg0.ShowModal()
testDlg0.Destroy()
dlg = wx.FileDialog(self,
message="Choose the CPAC Subject list file",
defaultDir=os.getcwd(),
defaultFile="CPAC_subject_list.yml",
wildcard="YAML files(*.yaml, *.yml)|*.yaml;*.yml",
style=wx.OPEN | wx.CHANGE_DIR)
if dlg.ShowModal() == wx.ID_OK:
subListPath = dlg.GetPath()
# Load and test the subject list
print 'Checking subject list: %s...' % subListPath
sublist = yaml.load(open(os.path.realpath(subListPath), 'r'))
sub_flg = self.test_sublist(sublist)
if not sub_flg:
raise Exception
print 'Subject list looks good!'
# Following code reads in the parameters and selections from the
# pipeline configuration window and populate the config_list
config_list = []
wf_counter = []
for page in self.nb.get_page_list():
switch = page.page.get_switch()
ctrl_list = page.page.get_ctrl_list()
validate = False
if switch:
switch_val = str(switch.get_selection()).lower()
if switch_val == 'on' or switch_val == 'true' or \
switch_val == '1':
validate = True
wf_counter.append(page.get_counter())
for ctrl in ctrl_list:
# option_name will be the selection name as it is written
# as the dictionary key of the config.yml dictionary
option_name = ctrl.get_name()
#validating
if (switch == None or validate) and ctrl.get_validation() \
and (option_name != 'derivativeList') and \
(option_name != 'modelConfigs'):
win = ctrl.get_ctrl()
if isinstance(ctrl.get_selection(), list):
value = ctrl.get_selection()
if not value:
display(
win, "%s field is empty or the items are " \
"not checked!" % ctrl.get_name(), False)
return
elif (option_name == "tsa_roi_paths") or \
(option_name == "sca_roi_paths"):
# fires if the control is the checkbox grid for
# multiple paths assigned to multiple options
# (i.e. timeseries analysis)
config_list.append(ctrl)
continue
else:
value = str(ctrl.get_selection())
if len(value) == 0:
display(win, "%s field is empty!" % ctrl.get_name())
return
if '/' in value and '$' not in value and not \
isinstance(value, list):
if not os.path.exists(ctrl.get_selection()) and \
value != 'On/Off':
display(
win, "%s field contains incorrect path. " \
"Please update the path!" % ctrl.get_name())
return
config_list.append(ctrl)
# Write out a pipeline_config file, read it in and then delete it
# (Will revise the data structure of the config files later so this
# can just pass the data structure instead of doing it this way)
try:
test_cfg_yml = '/tmp/test_config.yml'
self.write(test_cfg_yml, config_list)
c = Configuration(
yaml.load(open(os.path.realpath(test_cfg_yml), 'r')))
os.remove(test_cfg_yml)
except:
errDlg2 = wx.MessageDialog(
self, 'A problem occurred with preparing the pipeline test run. \n\n' \
'Please ensure you have rights access to the directories you' \
' have chosen for the CPAC working, crash, and output folders.',
'Test Configuration Error',
wx.OK | wx.ICON_ERROR)
errDlg2.ShowModal()
errDlg2.Destroy()
if (1 in c.runNuisance) or (c.Regressors != None):
strategies = sorted(build_strategies(c))
else:
strategies = None
# Run the actual pipeline building prep and see if it works or not
testDlg1 = wx.MessageDialog(
self,
'Click OK to run the test. This should take only a few seconds.',
'Running Test', wx.OK | wx.ICON_INFORMATION)
testDlg1.ShowModal()
# Check file paths first
# Just getting proper names of config file parameters
try:
params_file = open(
p.resource_filename('CPAC',
'GUI/resources/config_parameters.txt'),
"r")
except:
print "Error: Could not open configuration parameter file.", "\n"
raise Exception
paramInfo = params_file.read().split('\n')
paramList = []
for param in paramInfo:
if param != '':
paramList.append(param.split(','))
# function for file path checking
def testFile(filepath, paramName, switch):
try:
if (1 in switch) and (filepath != None):
fileTest = open(filepath)
fileTest.close()
except:
testDlg1.Destroy()
for param in paramList:
if param[0] == paramName:
paramTitle = param[1]
paramGroup = param[2]
break
errDlgFileTest = wx.MessageDialog(
self, 'Error reading file - either it does not exist or '\
'you do not have read access. \n\n' \
'Parameter: %s \n' \
'In tab: %s \n\n' \
'Path: %s' % (paramTitle, paramGroup, filepath),
'Pipeline Not Ready',
wx.OK | wx.ICON_ERROR)
errDlgFileTest.ShowModal()
errDlgFileTest.Destroy()
# Check S3 output bucket access if writing to S3
output_dir = c.outputDirectory
s3_str = 's3://'
if output_dir.lower().startswith(s3_str):
output_dir_sp = output_dir.split('/')
output_dir_sp[0] = output_dir_sp[0].lower()
output_dir = '/'.join(output_dir_sp)
if type(output_dir) is str and output_dir.lower().startswith(s3_str):
from indi_aws import fetch_creds
creds_path = c.awsOutputBucketCredentials
bucket_name = output_dir.split(s3_str)[1].split('/')[0]
try:
bucket = fetch_creds.return_bucket(creds_path, bucket_name)
print 'Connection with output bucket "%s" successful!' % bucket_name
except Exception as exc:
err_msg = 'Unable to access output S3 bucket: "%s" with '\
'credentials in: "%s". Check bucket name '\
'and credentials file and try again'\
% (bucket_name, creds_path)
testDlg1.Destroy()
errDlg1 = wx.MessageDialog(self, err_msg, 'Pipeline Not Ready',
wx.OK | wx.ICON_ERROR)
errDlg1.ShowModal()
errDlg1.Destroy()
return
testFile(c.template_brain_only_for_anat, \
'template_brain_only_for_anat',[1])
testFile(c.template_skull_for_anat, 'template_skull_for_anat', [1])
testFile(c.PRIORS_WHITE, 'PRIORS_WHITE',
c.runSegmentationPreprocessing)
testFile(c.PRIORS_GRAY, 'PRIORS_GRAY', c.runSegmentationPreprocessing)
testFile(c.PRIORS_CSF, 'PRIORS_CSF', c.runSegmentationPreprocessing)
testFile(c.template_brain_only_for_func, \
'template_brain_only_for_func',c.runRegisterFuncToMNI)
testFile(c.template_skull_for_func,'template_skull_for_func', \
c.runRegisterFuncToMNI)
testFile(c.identityMatrix, 'identityMatrix', c.runRegisterFuncToMNI)
testFile(c.boundaryBasedRegistrationSchedule, \
'boundaryBasedRegistrationSchedule', \
c.runRegisterFuncToAnat)
testFile(c.lateral_ventricles_mask,'lateral_ventricles_mask', \
c.runNuisance)
testFile(c.template_symmetric_brain_only, \
'template_symmetric_brain_only',c.runVMHC)
testFile(c.template_symmetric_skull,'template_symmetric_skull', \
c.runVMHC)
testFile(c.dilated_symmetric_brain_mask, \
'dilated_symmetric_brain_mask',c.runVMHC)
testFile(c.configFileTwomm, 'configFileTwomm', c.runVMHC)
testFile(c.templateSpecificationFile,'templateSpecificationFile', \
c.runNetworkCentrality)
if c.tsa_roi_paths and type(c.tsa_roi_paths[0]) == dict:
for roi_path in c.tsa_roi_paths[0].keys():
testFile(roi_path, "tsa_roi_paths", c.runROITimeseries)
if c.sca_roi_paths and type(c.sca_roi_paths[0]) == dict:
for roi_path in c.sca_roi_paths[0].keys():
testFile(roi_path, "sca_roi_paths", c.runSCA)
try:
# Run the pipeline building
prep_workflow(sublist[0], c, strategies, 0)
except Exception as xxx:
print xxx
print "an exception occured"
testDlg1.Destroy()
errDlg1 = wx.MessageDialog(
self, 'There are issues with the current configuration ' \
'which need to be resolved - please check to make ' \
'sure the options you are running have the proper ' \
'pre-requisites selected.\n\nIssue Info:\n%s' % xxx,
'Pipeline Not Ready',
wx.OK | wx.ICON_ERROR)
errDlg1.ShowModal()
errDlg1.Destroy()
else:
testDlg1.Destroy()
okDlg1 = wx.MessageDialog(
self, 'The current configuration will run successfully. You '\
'can safely save and run this setup!',
'Pipeline Ready',
wx.OK | wx.ICON_INFORMATION)
okDlg1.ShowModal()
okDlg1.Destroy()
def prep_group_analysis_workflow(c, resource, subject_infos):
#
# this function runs once per output file during group analysis
#
# p_id = a list of pipeline IDs, i.e. the name of the output folder for
# the strat
# s_ids = a list of all the subject IDs
# scan_ids = a list of scan IDs
# s_paths = a list of all of the filepaths of this particular output
# file that prep_group_analysis_workflow is being called for
p_id, s_ids, scan_ids, s_paths = (list(tup) for tup in zip(*subject_infos))
# set this to False for now
fTest = False
def get_phenotypic_file(phenotypic_file, m_dict, m_list, mod_path, sub_id):
import csv
reader = csv.reader(open(phenotypic_file, 'rU'))
columns = {}
order = {}
count = 0
headers = next(reader)
for h in headers:
columns[h] =[]
order[h] = count
count+=1
for r in reader:
for h, v in zip(headers, r):
if v:
columns[h].append(str(v))
if m_dict:
for measure in m_list:
print('\n\nMeasure: ', measure, '\n\n')
if measure in headers:
#check if 'MeanFD is present'
if len(columns[measure]) < 1:
print('\n\ncolumns[sub_id]: ', columns[sub_id], '\n\n')
for sub in columns[sub_id]:
if m_dict.get(sub):
if m_dict.get(sub).get(measure):
columns[measure].append(m_dict[sub][measure])
else:
raise Exception("Couldn't find %s value for subject %s"%(measure,sub))
else:
raise Exception("Couldn't find subject %s in the parameter file"%sub)
print('\n\ncolumns[measure]: ', columns, '\n\n')
b = list(zip(*([k] + columns[k] for k in sorted(columns, key=order.get))))
try:
os.makedirs(mod_path)
except:
print("%s already exists"%(mod_path))
new_phenotypic_file = os.path.join(mod_path, os.path.basename(phenotypic_file))
a = csv.writer(open(new_phenotypic_file, 'w'))
for col in b:
a.writerow(list(col))
return new_phenotypic_file
# END get_phenotypic_file function
threshold_val = None
measure_dict = None
measure_list = ['MeanFD', 'MeanFD_Jenkinson', 'MeanDVARS']
model_sub_list = []
if 1 in c.runScrubbing:
#get scrubbing threshold
if re.search('(?<=/_threshold_)\d+.\d+',s_paths[0]):
threshold_val = re.search('(?<=/_threshold_)\d+.\d+',s_paths[0]).group(0)
elif len(c.scrubbingThreshold) == 1:
threshold_val = c.scrubbingThreshold[0]
else:
print("Found Multiple threshold value ")
print("scrubbing threshold_val -->", threshold_val)
else:
print("No scrubbing enabled.")
if len(c.scrubbingThreshold) == 1:
threshold_val = c.scrubbingThreshold[0]
import yaml
for config in c.modelConfigs:
print(c.modelConfigs)
print(config)
try:
conf = Configuration(yaml.load(open(os.path.realpath(config), 'r')))
except:
raise Exception("Error in reading %s configuration file" % config)
group_sublist = open(conf.subject_list, 'r')
sublist_items = group_sublist.readlines()
subject_list = [line.rstrip('\n') for line in sublist_items \
if not (line == '\n') and not line.startswith('#')]
# list of subject paths which DO exist
exist_paths = []
''' begin iteration through group subject list for processing '''
for sub in subject_list:
# let's check to make sure the subject list is formatted for
# repeated measures properly if repeated measures is enabled and
# vice versa
if (c.repeatedMeasures == True) and (',' not in sub):
print('\n\n')
print('[!] CPAC says: The group analysis subject list is ' \
'not inthe appropriate format for repeated ' \
'measures.\n')
print('Please use the appropriate format as described in ' \
'the CPAC User Guide or turn off Repeated Measures ' \
'in the CPAC pipeline configuration editor, found ' \
'in the \'Group Analysis Settings\' tab of the ' \
'pipeline configuration editor.\n')
print('NOTE: CPAC generates a properly-formatted group ' \
'analysis subject list meant for running repeated ' \
'measures when you create your original subject ' \
'list. Look for \'subject_list_group_analysis_' \
'repeated_measures.txt\' in the directory where ' \
'you created your subject list.\n\n')
raise Exception
elif (c.repeatedMeasures == False) and (',' in sub):
print('\n\n')
print('[!] CPAC says: It looks like your group analysis ' \
'subject list is formatted for running repeated ' \
'measures, but \'Run Repeated Measures\' is not ' \
'enabled in the pipeline configuration, found in ' \
'the \'Group Analysis Settings\' tab of the ' \
'pipeline configuration editor.\n')
print('Double-check your pipeline configuration?\n\n')
raise Exception
''' process subject ids for repeated measures, if it is on '''
# if repeated measures is being run and the subject list
# is a list of subject IDs and scan IDs concatenated
if (c.repeatedMeasures == True):
# sub.count(',') equals 1 when there is either multiple scans
# or multiple sessions but not both, for repeated measures
# sub.count(',') equals 2 when there are multiple sessions
# AND scans, for repeated measures
if sub.count(',') == 1:
sub_id = sub.split(',',1)[0]
other_id = sub.split(',',1)[1]
elif sub.count(',') == 2:
sub_id = sub.split(',',2)[0]
scan_id = sub.split(',',2)[1]
session_id = sub.split(',',2)[2]
''' drop subjects from the group subject list '''
# check the path files in path_files_here folder in the subject's
# output folder - and drop any subjects from the group analysis
# subject list which do not exist in the paths to the output files
for path in s_paths:
if (c.repeatedMeasures == True):
if sub.count(',') == 1:
if (sub_id in path) and (other_id in path):
exist_paths.append(sub)
elif sub.count(',') == 2:
if (sub_id in path) and (scan_id in path) and \
(session_id in path):
exist_paths.append(sub)
else:
if sub in path:
exist_paths.append(sub)
# check to see if any derivatives of subjects are missing
if len(list(set(subject_list) - set(exist_paths))) >0:
print("List of outputs missing for subjects:")
print(list(set(subject_list) - set(exist_paths)))
print("..for derivatives:")
print(resource)
print("..at paths:")
print(os.path.dirname(s_paths[0]).replace(s_ids[0], '*'))
# create the path string for the group analysis output
out_dir = os.path.dirname(s_paths[0]).split(p_id[0] + '/')
out_dir = os.path.join(conf.output_dir, out_dir[1])
out_dir = out_dir.replace(s_ids[0], 'group_analysis_results_%s/_grp_model_%s'%(p_id[0],conf.model_name))
mod_path = os.path.join(out_dir, 'model_files')
if not os.path.isdir(mod_path):
os.makedirs(mod_path)
''' write the new subject list '''
new_sub_file = os.path.join(mod_path, os.path.basename(conf.subject_list))
try:
f = open(new_sub_file, 'w')
for sub in exist_paths:
print(sub, file=f)
f.close()
except:
print("Error: Could not open subject list file: ", new_sub_file)
raise Exception
conf.update('subject_list',new_sub_file)
sub_id = conf.subject_id_label
if measure_dict != None:
conf.update('pheno_file',get_phenotypic_file(conf.pheno_file, measure_dict, measure_list, mod_path, sub_id))
print('conf updated pheno: ', conf.pheno_file, '\n\n')
print("Model config dictionary ->")
print(conf.__dict__)
# Run 'create_fsl_model' script to extract phenotypic data from
# the phenotypic file for each of the subjects in the subject list
''' get the motion statistics parameter file, if present '''
# get the parameter file so it can be passed to create_fsl_model.py
# so MeanFD or other measures can be included in the design matrix
parameter_file = os.path.join(c.outputDirectory, p_id[0], '%s_threshold_%s_all_params.csv'%(scan_ids[0].strip('_'),threshold_val))
if 1 in c.runGenerateMotionStatistics:
if not os.path.exists(parameter_file):
print('\n\n[!] CPAC says: Could not open the parameter file. ' \
'If Generate Motion Statistics is enabled, this can ' \
'usually be found in the output directory of your ' \
'individual-level analysis runs.\n')
print('Path not found: ', parameter_file, '\n\n')
raise Exception
elif (1 not in c.runGenerateMotionStatistics) and (os.path.exists(parameter_file)):
if not os.path.exists(parameter_file):
print('\n\n[!] CPAC says: Could not open the parameter file. ' \
'If Generate Motion Statistics is enabled, this can ' \
'usually be found in the output directory of your ' \
'individual-level analysis runs.\n')
print('Path not found: ', parameter_file, '\n\n')
raise Exception
else:
def no_measures_error(measure):
print('\n\n[!] CPAC says: The measure %s was included in ' \
'your group analysis design matrix formula, but ' \
'Generate Motion Statistics was not run during ' \
'individual-level analysis.\n' % measure)
print('Please run Generate Motion Statistics if you wish ' \
'to include this measure in your model.\n')
print('If you HAVE completed a run with this option ' \
'enabled, then you are seeing this error because ' \
'the motion parameter file normally created by this ' \
'option is missing.\n\n')
raise Exception
for measure in measure_list:
if (measure in conf.design_formula):
no_measures_error(measure)
parameter_file = None
''' run create_fsl_model.py to generate the group analysis models '''
# path to the pipeline folder to be passed to create_fsl_model.py
# so that certain files like output_means.csv can be accessed
pipeline_path = os.path.join(c.outputDirectory, p_id[0])
# the current output that cpac_group_analysis_pipeline.py and
# create_fsl_model.py is currently being run for
current_output = s_paths[0].replace(pipeline_path, '').split('/')[2]
try:
from CPAC.utils import create_fsl_model
create_fsl_model.run(conf, fTest, parameter_file, pipeline_path, current_output, True)
#print >>diag, "> Runs create_fsl_model."
#print >>diag, ""
except Exception as e:
print("FSL Group Analysis model not successfully created - error in create_fsl_model script")
#print "Error ->", e
raise
model_sub_list.append((conf.output_dir, conf.subject_list))
if len(model_sub_list) == 0:
raise Exception("no model found")
''' start group analysis '''
print('\n\nPreparing the group analysis workflow..\n\n')
for model_sub in model_sub_list:
#print >>diag, "Current model_sub: ", model_sub
#print >>diag, ""
model, subject_list = model_sub
if not os.path.exists(model):
raise Exception("path to model %s doesn't exist"%model)
if not os.path.exists(subject_list):
raise Exception("path to input subject list %s is invalid" % subject_list)
#if c.mixedScanAnalysis == True:
# wf = pe.Workflow(name = 'group_analysis/%s/grp_model_%s'%(resource, os.path.basename(model)))
#else:
# s_paths is a list of paths to each subject's derivative (of the current
# derivative gpa is being run on) - s_paths_dirList is a list of each directory
# in this path separated into list elements
s_paths_dirList = s_paths[0].split('/')
currentDerivativeFile = s_paths_dirList[-1]
currentDerivative = currentDerivativeFile.split('.')[0]
currentDerivative = currentDerivative.replace('#', '_')
strgy_path = os.path.dirname(s_paths[0]).split(scan_ids[0])[1]
for ch in ['.']:
if ch in strgy_path:
strgy_path = strgy_path.replace(ch, '_')
# create nipype-workflow-name-friendly strgy_path
# (remove special characters)
strgy_path_name = strgy_path.replace('/', '__')
wf = pe.Workflow(name = currentDerivative)
workDir = c.workingDirectory + '/group_analysis__%s__grp_model_%s__%s' % (resource, conf.model_name, scan_ids[0])
workDir = workDir + '/' + strgy_path_name
wf.base_dir = workDir
wf.config['execution'] = {'hash_method': 'timestamp', 'crashdump_dir': os.path.abspath(c.crashLogDirectory)}
log_dir = os.path.join(conf.output_dir, 'logs', 'group_analysis', resource, 'model_%s' % (conf.model_name))
if not os.path.exists(log_dir):
os.makedirs(log_dir)
else:
print("log_dir already exist")
# enable logging
from nipype import config
from nipype import logging
config.update_config({'logging': {'log_directory': log_dir,
'log_to_file': True}})
# Temporarily disable until solved
#logging.update_logging(config)
iflogger = logging.getLogger('interface')
''' create the list of paths to all output files to go to model '''
# create the 'ordered_paths' list, which is a list of all of the
# output paths of the output files being included in the current
# group-level analysis model
# 'ordered_paths' is later connected to the 'zmap_files' input
# of the group analysis workflow - the files listed in this list
# are merged into the merged 4D file that goes into group analysis
group_sublist = open(subject_list, 'r')
sublist_items = group_sublist.readlines()
input_subject_list = [line.rstrip('\n') for line in sublist_items \
if not (line == '\n') and not line.startswith('#')]
ordered_paths = []
pathcount = 0
subcount = 0
for sub in input_subject_list:
subcount += 1
if (c.repeatedMeasures == True):
# sub.count(',') equals 1 when there is either multiple scans
# or multiple sessions but not both, for repeated measures
# sub.count(',') equals 2 when there are multiple sessions
# AND scans, for repeated measures
if sub.count(',') == 1:
sub_id = sub.split(',',1)[0]
other_id = sub.split(',',1)[1]
elif sub.count(',') == 2:
sub_id = sub.split(',',2)[0]
scan_id = sub.split(',',2)[1]
session_id = sub.split(',',2)[2]
for path in s_paths:
if (c.repeatedMeasures == True):
# if repeated measures is enabled, make sure all of the
# relevant indicators are in the path before adding it
# to 'ordered_paths', i.e. the session and/or scan IDs
if sub.count(',') == 1:
if (sub_id in path) and (other_id in path):
pathcount += 1
ordered_paths.append(path)
elif sub.count(',') == 2:
if (sub_id in path) and (scan_id in path) and \
(session_id in path):
pathcount += 1
ordered_paths.append(path)
else:
if sub in path:
pathcount += 1
ordered_paths.append(path)
print('S_paths length: ', len(s_paths))
print("Ordered paths length (number of subjects): ", len(ordered_paths))
print("input_subject_list -> %s" % input_subject_list)
print("strgy_path: ", strgy_path)
if len(ordered_paths) == 0:
print('\n\n\n[!] CPAC says: None of the subjects listed in the ' \
'group analysis subject list were found to have outputs ' \
'produced by individual-level analysis.\n\nEnsure that ' \
'the subjects listed in your group analysis subject list ' \
'are the same as the ones included in the individual-' \
'level analysis you are running group-level analysis for.' \
'\n\n\n')
raise Exception
# gp_flow
# Extracts the model files (.con, .grp, .mat, .fts) from the model
# directory and sends them to the create_group_analysis workflow gpa_wf
gp_flow = create_grp_analysis_dataflow("gp_dataflow_%s" % currentDerivative)
gp_flow.inputs.inputspec.grp_model = model
gp_flow.inputs.inputspec.fTest = fTest
# gpa_wf
# Creates the actual group analysis workflow
gpa_wf = create_group_analysis(fTest, "gp_analysis_%s" % currentDerivative)
gpa_wf.inputs.inputspec.zmap_files = ordered_paths
gpa_wf.inputs.inputspec.z_threshold = c.zThreshold
gpa_wf.inputs.inputspec.p_threshold = c.pThreshold
gpa_wf.inputs.inputspec.parameters = (c.FSLDIR, 'MNI152')
print("group model: ", model)
print("f test: ", fTest)
print("z threshold: ", c.zThreshold)
print("p threshold: ", c.pThreshold)
print("parameters: ", (c.FSLDIR, 'MNI152'))
wf.connect(gp_flow, 'outputspec.mat',
gpa_wf, 'inputspec.mat_file')
wf.connect(gp_flow, 'outputspec.con',
gpa_wf, 'inputspec.con_file')
wf.connect(gp_flow, 'outputspec.grp',
gpa_wf, 'inputspec.grp_file')
if fTest:
wf.connect(gp_flow, 'outputspec.fts',
gpa_wf, 'inputspec.fts_file')
# ds
# Creates the datasink node for group analysis
ds = pe.Node(nio.DataSink(), name='gpa_sink')
if 'sca_roi' in resource:
out_dir = os.path.join(out_dir, \
re.search('ROI_number_(\d)+',os.path.splitext(os.path.splitext(os.path.basename(s_paths[0]))[0])[0]).group(0))
if 'centrality' in resource:
names = ['degree_centrality_binarize', 'degree_centrality_weighted', \
'eigenvector_centrality_binarize', 'eigenvector_centrality_weighted', \
'lfcd_binarize', 'lfcd_weighted']
for name in names:
if name in os.path.basename(s_paths[0]):
out_dir = os.path.join(out_dir, name)
break
if 'tempreg_maps_z_files' in resource:
out_dir = os.path.join(out_dir, \
re.search('\w*[#]*\d+', os.path.splitext(os.path.splitext(os.path.basename(s_paths[0]))[0])[0]).group(0))
# if c.mixedScanAnalysis == True:
# out_dir = re.sub(r'(\w)*scan_(\w)*(\d)*(\w)*[/]', '', out_dir)
ds.inputs.base_directory = out_dir
ds.inputs.container = ''
ds.inputs.regexp_substitutions = [(r'(?<=rendered)(.)*[/]','/'),
(r'(?<=model_files)(.)*[/]','/'),
(r'(?<=merged)(.)*[/]','/'),
(r'(?<=stats/clusterMap)(.)*[/]','/'),
(r'(?<=stats/unthreshold)(.)*[/]','/'),
(r'(?<=stats/threshold)(.)*[/]','/'),
(r'_cluster(.)*[/]',''),
(r'_slicer(.)*[/]',''),
(r'_overlay(.)*[/]','')]
'''
if 1 in c.runSymbolicLinks:
link_node = pe.MapNode(interface=util.Function(
input_names=['in_file',
'resource'],
output_names=[],
function=prepare_gp_links),
name='link_gp_', iterfield=['in_file'])
link_node.inputs.resource = resource
wf.connect(ds, 'out_file', link_node, 'in_file')
'''
########datasink connections#########
if fTest:
wf.connect(gp_flow, 'outputspec.fts',
ds, 'model_files.@0')
wf.connect(gp_flow, 'outputspec.mat',
ds, 'model_files.@1' )
wf.connect(gp_flow, 'outputspec.con',
ds, 'model_files.@2')
wf.connect(gp_flow, 'outputspec.grp',
ds, 'model_files.@3')
wf.connect(gpa_wf, 'outputspec.merged',
ds, 'merged')
wf.connect(gpa_wf, 'outputspec.zstats',
ds, 'stats.unthreshold')
wf.connect(gpa_wf, 'outputspec.zfstats',
ds,'stats.unthreshold.@01')
wf.connect(gpa_wf, 'outputspec.fstats',
ds,'stats.unthreshold.@02')
wf.connect(gpa_wf, 'outputspec.cluster_threshold_zf',
ds, 'stats.threshold')
wf.connect(gpa_wf, 'outputspec.cluster_index_zf',
ds,'stats.clusterMap')
wf.connect(gpa_wf, 'outputspec.cluster_localmax_txt_zf',
ds, 'stats.clusterMap.@01')
wf.connect(gpa_wf, 'outputspec.overlay_threshold_zf',
ds, 'rendered')
wf.connect(gpa_wf, 'outputspec.rendered_image_zf',
ds, 'rendered.@01')
wf.connect(gpa_wf, 'outputspec.cluster_threshold',
ds, 'stats.threshold.@01')
wf.connect(gpa_wf, 'outputspec.cluster_index',
ds, 'stats.clusterMap.@02')
wf.connect(gpa_wf, 'outputspec.cluster_localmax_txt',
ds, 'stats.clusterMap.@03')
wf.connect(gpa_wf, 'outputspec.overlay_threshold',
ds, 'rendered.@02')
wf.connect(gpa_wf, 'outputspec.rendered_image',
ds, 'rendered.@03')
######################################
# Run the actual group analysis workflow
wf.run()
'''
except:
print "Error: Group analysis workflow run command did not complete successfully."
print "subcount: ", subcount
print "pathcount: ", pathcount
print "sublist: ", sublist_items
print "input subject list: "
print "conf: ", conf.subjectListFile
raise Exception
'''
print("**Workflow finished for model %s and resource %s"%(os.path.basename(model), resource))
class ListBox(wx.Frame):
def __init__(self, parent, id, title):
wx.Frame.__init__(self,
parent,
id,
title,
size=(700, 650),
style=wx.SYSTEM_MENU | wx.CAPTION | wx.CLOSE_BOX)
# Import packages
import CPAC
self.CreateStatusBar()
self.SetStatusText("The Configurable Pipeline for the Analysis of "
"Connectomes (C-PAC) v" + CPAC.__version__)
self.pipeline_map = {}
self.sublist_map = {}
self.pids = []
mainPanel = wx.Panel(self)
mainPanel.SetBackgroundColour('#E9E3DB')
mainSizer = wx.BoxSizer(wx.VERTICAL)
outerPanel1 = wx.Panel(mainPanel)
outerSizer1 = wx.BoxSizer(wx.HORIZONTAL)
outerPanel2 = wx.Panel(mainPanel)
outerSizer2 = wx.BoxSizer(wx.HORIZONTAL)
outerPanel3 = wx.Panel(mainPanel)
outerSizer3 = wx.BoxSizer(wx.HORIZONTAL)
innerPanel1 = wx.Panel(outerPanel1)
innerSizer1 = wx.BoxSizer(wx.HORIZONTAL)
innerPanel2 = wx.Panel(outerPanel1, )
innerSizer2 = wx.BoxSizer(wx.HORIZONTAL)
lboxPanel1 = wx.Panel(innerPanel1)
lboxSizer1 = wx.BoxSizer(wx.VERTICAL)
btnPanel1 = wx.Panel(innerPanel1, -1)
btnSizer1 = wx.BoxSizer(wx.VERTICAL)
label = wx.StaticText(lboxPanel1, -1, "Pipelines")
if 'linux' in sys.platform:
label.SetFont(wx.Font(12, wx.SWISS, wx.NORMAL, wx.BOLD))
else:
label.SetFont(wx.Font(16, wx.SWISS, wx.NORMAL, wx.BOLD))
self.listbox = wx.CheckListBox(lboxPanel1, -1, size=(160, 400))
lboxSizer1.Add(label, 0, wx.ALIGN_CENTER)
lboxSizer1.Add(self.listbox, 1, wx.EXPAND | wx.ALL, 10)
lboxPanel1.SetSizer(lboxSizer1)
lboxPanel1.SetBackgroundColour('#E9E3DB')
new = wx.Button(btnPanel1, ID_NEW, 'New', size=(90, 30))
group = wx.Button(btnPanel1, ID_GROUP, 'New Group', size=(90, 30))
ren = wx.Button(btnPanel1, ID_RENAME, 'Rename', size=(90, 30))
dlt = wx.Button(btnPanel1, ID_DELETE, 'Delete', size=(90, 30))
load = wx.Button(btnPanel1, ID_LOAD, 'Load', size=(90, 30))
edit = wx.Button(btnPanel1, ID_EDIT, 'Edit', size=(90, 30))
shw = wx.Button(btnPanel1, ID_DISPLAY, 'View', size=(90, 30))
clr = wx.Button(btnPanel1, ID_CLEAR, 'Clear', size=(90, 30))
self.Bind(wx.EVT_BUTTON, self.NewItem, id=ID_NEW)
self.Bind(wx.EVT_BUTTON, self.NewGroup, id=ID_GROUP)
self.Bind(wx.EVT_BUTTON, self.OnRename, id=ID_RENAME)
self.Bind(wx.EVT_BUTTON, self.OnDelete, id=ID_DELETE)
self.Bind(wx.EVT_BUTTON, self.AddConfig, id=ID_LOAD)
self.Bind(wx.EVT_BUTTON, self.OnEdit, id=ID_EDIT)
self.Bind(wx.EVT_BUTTON, self.OnDisplay, id=ID_DISPLAY)
self.Bind(wx.EVT_BUTTON,
lambda event: self.OnClear(event, 1),
id=ID_CLEAR)
self.Bind(wx.EVT_LISTBOX_DCLICK, self.OnDisplay)
if 'linux' in sys.platform:
btnSizer1.Add((-1, 30))
else:
btnSizer1.Add((-1, 27))
btnSizer1.Add(new, 0, wx.TOP)
btnSizer1.Add(group, 0, wx.TOP)
btnSizer1.Add(load, 0, wx.TOP)
btnSizer1.Add(edit, 0, wx.TOP)
btnSizer1.Add(shw, 0, wx.TOP)
btnSizer1.Add(ren, 0, wx.TOP)
btnSizer1.Add(dlt, 0, wx.TOP)
btnSizer1.Add(clr, 0, wx.TOP)
btnPanel1.SetSizer(btnSizer1)
btnPanel1.SetBackgroundColour('#E9E3DB')
innerSizer1.Add(lboxPanel1, 1, wx.EXPAND | wx.ALL)
if 'linux' in sys.platform:
innerSizer1.Add(btnPanel1, 1, wx.EXPAND | wx.ALL, 5)
else:
innerSizer1.Add(btnPanel1, 1, wx.EXPAND | wx.ALL)
innerPanel1.SetSizer(innerSizer1)
innerPanel1.SetBackgroundColour('#E9E3DB')
lboxPanel2 = wx.Panel(innerPanel2)
lboxSizer2 = wx.BoxSizer(wx.VERTICAL)
btnPanel2 = wx.Panel(innerPanel2, -1)
btnSizer2 = wx.BoxSizer(wx.VERTICAL)
label2 = wx.StaticText(lboxPanel2, -1, "Data Configurations")
if 'linux' in sys.platform:
label2.SetFont(wx.Font(12, wx.SWISS, wx.NORMAL, wx.BOLD))
else:
label2.SetFont(wx.Font(16, wx.SWISS, wx.NORMAL, wx.BOLD))
self.listbox2 = wx.CheckListBox(lboxPanel2, -1, size=(160, 400))
self.listbox2.Bind(wx.EVT_LISTBOX_DCLICK, self.OnShow)
lboxSizer2.Add(label2, 0, wx.ALIGN_CENTER)
lboxSizer2.Add(self.listbox2, 1, wx.EXPAND | wx.ALL, 10)
lboxPanel2.SetSizer(lboxSizer2)
lboxPanel2.SetBackgroundColour('#E9E3DB')
create = wx.Button(btnPanel2, ID_CREATE, 'New', size=(90, 30))
add = wx.Button(btnPanel2, ID_ADD, 'Load', size=(90, 30))
show = wx.Button(btnPanel2, ID_SHOW, 'View', size=(90, 30))
clr2 = wx.Button(btnPanel2, ID_CLEARALL, 'Clear', size=(90, 30))
self.Bind(wx.EVT_BUTTON, self.CreateItem, id=ID_CREATE)
self.Bind(wx.EVT_BUTTON, self.AddItem, id=ID_ADD)
self.Bind(wx.EVT_BUTTON, self.OnShow, id=ID_SHOW)
self.Bind(wx.EVT_BUTTON,
lambda event: self.OnClear(event, 2),
id=ID_CLEARALL)
if 'linux' in sys.platform:
btnSizer2.Add((-1, 30))
else:
btnSizer2.Add((-1, 27))
# Add buttons to button sizer
btnSizer2.Add(create, 0, wx.TOP)
btnSizer2.Add(add, 0, wx.TOP)
btnSizer2.Add(show, 0, wx.TOP)
btnSizer2.Add(clr2, 0, wx.TOP)
btnPanel2.SetSizer(btnSizer2)
btnPanel2.SetBackgroundColour('#E9E3DB')
innerSizer2.Add(lboxPanel2, 1, wx.EXPAND | wx.ALL)
if 'linux' in sys.platform:
innerSizer2.Add(btnPanel2, 1, wx.EXPAND | wx.ALL, 5)
else:
innerSizer2.Add(btnPanel2, 1, wx.EXPAND | wx.ALL)
innerPanel2.SetSizer(innerSizer2)
innerPanel2.SetBackgroundColour('#E9E3DB')
outerSizer1.Add(innerPanel2, 1, wx.EXPAND | wx.ALL)
outerSizer1.Add(innerPanel1, 1, wx.EXPAND | wx.ALL)
outerPanel1.SetSizer(outerSizer1)
outerPanel1.SetBackgroundColour('#E9E3DB')
self.runCPAC1 = wx.Button(outerPanel2, -1,
'Run Individual Level Analysis')
self.runCPAC1.Bind(wx.EVT_BUTTON, self.runIndividualAnalysis)
self.runCPAC2 = wx.Button(outerPanel2, -1, 'Run Group Level Analysis')
self.runCPAC2.Bind(wx.EVT_BUTTON, self.runGroupLevelAnalysis)
self.openPresets = wx.Button(outerPanel3, -1,
'Generate FSL-FEAT Presets')
self.openPresets.Bind(wx.EVT_BUTTON, self.openFSLPresets)
self.buildModels = wx.Button(outerPanel3, -1, 'Build FSL-FEAT Models')
self.buildModels.Bind(wx.EVT_BUTTON, self.buildFSLModels)
outerSizer2.Add(self.runCPAC1, 1, wx.RIGHT, 12)
outerSizer2.Add(self.runCPAC2, 1, wx.LEFT, 12)
outerSizer3.Add(self.openPresets, 1, wx.RIGHT, 12)
outerSizer3.Add(self.buildModels, 1, wx.LEFT, 12)
#outerSizer3.Add(self.stopCPAC1, 1, wx.RIGHT, 20)
#outerSizer3.Add(self.stopCPAC2, 1, wx.LEFT, 20)
outerPanel2.SetSizer(outerSizer2)
outerPanel2.SetBackgroundColour('#E9E3DB')
outerPanel3.SetSizer(outerSizer3)
outerPanel3.SetBackgroundColour('#E9E3DB')
hbox = wx.BoxSizer(wx.HORIZONTAL)
text1 = wx.StaticText(mainPanel, -1, "Configure CPAC")
if 'linux' in sys.platform:
text1.SetFont(wx.Font(14, wx.SWISS, wx.NORMAL, wx.BOLD))
else:
text1.SetFont(wx.Font(18, wx.SWISS, wx.NORMAL, wx.BOLD))
img = wx.Image(
p.resource_filename('CPAC',
'GUI/resources/images/cpac_new_logo.png'),
wx.BITMAP_TYPE_PNG).ConvertToBitmap()
logo = wx.StaticBitmap(mainPanel, -1, img)
hbox.Add(text1, 1, wx.TOP | wx.EXPAND, 15)
hbox.Add(logo, 0, wx.ALIGN_RIGHT | wx.RIGHT)
text2 = wx.StaticText(mainPanel, -1, "Run CPAC")
if 'linux' in sys.platform:
text2.SetFont(wx.Font(14, wx.SWISS, wx.NORMAL, wx.BOLD))
else:
text2.SetFont(wx.Font(18, wx.SWISS, wx.NORMAL, wx.BOLD))
mainSizer.Add(hbox, 0, wx.EXPAND | wx.ALL, 10)
mainSizer.Add(outerPanel1, 1, wx.EXPAND | wx.ALL, 20)
mainSizer.Add(wx.StaticLine(mainPanel), 0,
wx.EXPAND | wx.TOP | wx.BOTTOM, 10)
mainSizer.Add(text2, 0, wx.EXPAND | wx.ALL, 5)
mainSizer.Add(outerPanel2, 0, wx.EXPAND | wx.ALL, 5)
mainSizer.Add(outerPanel3, 0, wx.EXPAND | wx.ALL, 5)
mainPanel.SetSizer(mainSizer)
self.Centre()
self.Show(True)
def runAnalysis1(self, pipeline, sublist, p):
try:
import CPAC
from CPAC.utils import Configuration
except ImportError, e:
wx.MessageBox("Error importing CPAC. %s" % e, "Error")
print "Error importing CPAC"
print e
c = Configuration(yaml.load(open(os.path.realpath(pipeline), 'r')))
plugin_args = {
'n_procs': c.maxCoresPerParticipant,
'memory_gb': c.maximumMemoryPerParticipant
}
# TODO: make this work
if self.pids:
#print "THERE'S SOMETHING RUNNING!"
pass
CPAC.pipeline.cpac_runner.run(pipeline,
sublist,
p,
plugin='MultiProc',
plugin_args=plugin_args)
def prep_group_analysis_workflow(model_df, pipeline_config_obj, \
model_name, group_config_obj, resource_id, preproc_strat, \
series_or_repeated_label):
#
# this function runs once per derivative type and preproc strat combo
# during group analysis
#
import os
import nipype.pipeline.engine as pe
import nipype.interfaces.utility as util
import nipype.interfaces.io as nio
pipeline_ID = pipeline_config_obj.pipeline_name
# get thresholds
z_threshold = float(group_config_obj.z_threshold[0])
p_threshold = float(group_config_obj.p_threshold[0])
sub_id_label = group_config_obj.subject_id_label
# determine if f-tests are included or not
custom_confile = group_config_obj.custom_contrasts
if ((custom_confile == None) or (custom_confile == '') or \
("None" in custom_confile) or ("none" in custom_confile)):
if (len(group_config_obj.f_tests) == 0) or \
(group_config_obj.f_tests == None):
fTest = False
else:
fTest = True
else:
if not os.path.exists(custom_confile):
errmsg = "\n[!] CPAC says: You've specified a custom contrasts " \
".CSV file for your group model, but this file cannot " \
"be found. Please double-check the filepath you have " \
"entered.\n\nFilepath: %s\n\n" % custom_confile
raise Exception(errmsg)
with open(custom_confile,"r") as f:
evs = f.readline()
evs = evs.rstrip('\r\n').split(',')
count_ftests = 0
fTest = False
for ev in evs:
if "f_test" in ev:
count_ftests += 1
if count_ftests > 0:
fTest = True
# create path for output directory
out_dir = os.path.join(group_config_obj.output_dir, \
"group_analysis_results_%s" % pipeline_ID, \
"group_model_%s" % model_name, resource_id, \
series_or_repeated_label, preproc_strat)
model_path = os.path.join(out_dir, 'model_files')
# generate working directory for this output's group analysis run
work_dir = os.path.join(c.workingDirectory, "group_analysis", model_name,\
resource_id, series_or_repeated_label, preproc_strat)
log_dir = os.path.join(out_dir, 'logs', resource_id, \
'model_%s' % model_name)
# create the actual directories
if not os.path.isdir(model_path):
try:
os.makedirs(model_path)
except Exception as e:
err = "\n\n[!] Could not create the group analysis output " \
"directories.\n\nAttempted directory creation: %s\n\n" \
"Error details: %s\n\n" % (model_path, e)
raise Exception(err)
if not os.path.isdir(work_dir):
try:
os.makedirs(work_dir)
except Exception as e:
err = "\n\n[!] Could not create the group analysis working " \
"directories.\n\nAttempted directory creation: %s\n\n" \
"Error details: %s\n\n" % (model_path, e)
raise Exception(err)
if not os.path.isdir(log_dir):
try:
os.makedirs(log_dir)
except Exception as e:
err = "\n\n[!] Could not create the group analysis logfile " \
"directories.\n\nAttempted directory creation: %s\n\n" \
"Error details: %s\n\n" % (model_path, e)
raise Exception(err)
# create new subject list based on which subjects are left after checking
# for missing outputs
new_participant_list = []
for part in list(model_df["Participant"]):
# do this instead of using "set" just in case, to preserve order
# only reason there may be duplicates is because of multiple-series
# repeated measures runs
if part not in new_participant_list:
new_participant_list.append(part)
new_sub_file = write_new_sub_file(model_path, \
group_config_obj.participant_list, \
new_participant_list)
group_conf.update('participant_list',new_sub_file)
# start processing the dataframe further
design_formula = group_config_obj.design_formula
# demean the motion params
if ("MeanFD" in design_formula) or ("MeanDVARS" in design_formula):
params = ["MeanFD_Power", "MeanFD_Jenkinson", "MeanDVARS"]
for param in params:
model_df[param] = model_df[param].astype(float)
model_df[param] = model_df[param].sub(model_df[param].mean())
# create 4D merged copefile, in the correct order, identical to design
# matrix
merge_outfile = model_name + "_" + resource_id + "_merged.nii.gz"
merge_outfile = os.path.join(model_path, merge_outfile)
merge_file = create_merged_copefile(list(model_df["Filepath"]), \
merge_outfile)
# create merged group mask
if group_config_obj.mean_mask[0] == "Group Mask":
merge_mask_outfile = os.path.basename(merge_file) + "_mask.nii.gz"
merge_mask = create_merged_mask(merge_file, merge_mask_outfile)
# calculate measure means, and demean
if "Measure_Mean" in design_formula:
model_df = calculate_measure_mean_in_df(model_df, merge_mask)
# calculate custom ROIs, and demean (in workflow?)
if "Custom_ROI_Mean" in design_formula:
custom_roi_mask = group_config_obj.custom_roi_mask
if (custom_roi_mask == None) or (custom_roi_mask == "None") or \
(custom_roi_mask == "none") or (custom_roi_mask == ""):
err = "\n\n[!] You included 'Custom_ROI_Mean' in your design " \
"formula, but you didn't supply a custom ROI mask file." \
"\n\nDesign formula: %s\n\n" % design_formula
raise Exception(err)
# make sure the custom ROI mask file is the same resolution as the
# output files - if not, resample and warn the user
roi_mask = check_mask_file_resolution(list(model_df["Raw_Filepath"])[0], \
custom_roi_mask, model_path, \
resource_id)
# if using group merged mask, trim the custom ROI mask to be within
# its constraints
if merge_mask:
output_mask = os.path.join(model_path, "group_masked_%s" \
% os.path.basename(input_mask))
roi_mask = trim_mask(roi_mask, merge_mask, output_mask)
# calculate
model_df = calculate_custom_roi_mean_in_df(model_df, roi_mask)
# modeling group variances separately
# add repeated measures 1's matrices
# patsify model DF, drop columns not in design formula
# process contrasts
wf = pe.Workflow(name=resource_id)
wf.base_dir = work_dir
crash_dir = os.path.join(pipeline_config_obj.crashLogDirectory, \
"group_analysis", model_name)
wf.config['execution'] = {'hash_method': 'timestamp', \
'crashdump_dir': crash_dir}
if "Measure_Mean" in design_formula:
measure_mean = pe.Node(util.Function(input_names=['model_df',
'merge_mask'],
output_names=['model_df'],
function=calculate_measure_mean_in_df),
name='measure_mean')
measure_mean.inputs.model_df = model_df
wf.connect(merge_mask, "out_file", measure_mean, "merge_mask")
if "Custom_ROI_Mean" in design_formula:
roi_mean = pe.Node(util.Function())
group_config_obj.custom_roi_mask
#----------------
import yaml
import pandas as pd
# load group analysis model configuration file
try:
with open(os.path.realpath(group_config_file),"r") as f:
group_conf = Configuration(yaml.load(f))
except Exception as e:
err_string = "\n\n[!] CPAC says: Could not read group model " \
"configuration YML file. Ensure you have read access " \
"for the file and that it is formatted properly.\n\n" \
"Configuration file: %s\n\nError details: %s" \
% (group_config_file, e)
raise Exception(err_string)
# gather all of the information
# - lists of all the participant unique IDs (participant_site_session) and
# of all of the series IDs present in output_file_list
# - also returns the pipeline ID
new_participant_list, all_series_names, pipeline_ID = \
gather_new_participant_list(output_path_file, output_file_list)
# create the path string for the group analysis output
# replicate the directory path of one of the participant's output
# folder path to the derivative's file, but replace the participant ID
# with the group model name
# this is to ensure nothing gets overwritten between strategies
# or thresholds, etc.
out_dir = os.path.dirname(output_file_list[0]).split(pipeline_ID + '/')
out_dir = out_dir[1].split(out_dir[1].split("/")[-1])[0]
out_dir = os.path.join(group_conf.output_dir, out_dir)
out_dir = out_dir.replace(new_participant_list[0], \
'group_analysis_results_%s/_grp_model_%s' \
% (pipeline_ID, group_conf.model_name))
# !!!!!!!!!!
if (group_conf.repeated_measures == True) and (series_ids[0] != None):
out_dir = out_dir.replace(series_ids[0] + "/", "multiple_series")
# create model file output directories
model_out_dir = os.path.join(group_conf.output_dir, \
'group_analysis_results_%s/_grp_model_%s' \
%(pipeline_ID, group_conf.model_name))
mod_path = os.path.join(model_out_dir, 'model_files')
if not os.path.isdir(mod_path):
os.makedirs(mod_path)
# current_mod_path = folder under
# "/gpa_output/_grp_model_{model name}/model_files/{current derivative}"
current_mod_path = os.path.join(mod_path, resource)
if not os.path.isdir(current_mod_path):
os.makedirs(current_mod_path)
# create new subject list based on which subjects are left after checking
# for missing outputs
new_sub_file = write_new_sub_file(current_mod_path, \
group_conf.subject_list, new_participant_list)
group_conf.update('subject_list',new_sub_file)
# create new design matrix with only the subjects that are left
# Run 'create_fsl_model' script to extract phenotypic data from
# the phenotypic file for each of the subjects in the subject list
# get the motion statistics parameter file, if present
# get the parameter file so it can be passed to create_fsl_model.py
# so MeanFD or other measures can be included in the design matrix
''' okay, here we go... how are we handling series? because here it needs to take in '''
''' the appropriate series to get the appropriate parameter file ! ! ! '''
''' MAY HAVE TO GO BACK ON THIS, and just have one series sent in per this function...'''
power_params_files = {}
measure_list = ['MeanFD_Power', 'MeanFD_Jenkinson', 'MeanDVARS']
for measure in measure_list:
if measure in group_conf.design_formula:
for series_id in all_series_names:
parameter_file = os.path.join(c.outputDirectory, \
pipeline_ID, \
'%s%s_all_params.csv' % \
(series_id.strip('_'), \
threshold_val))
if not os.path.exists(parameter_file):
err = "\n\n[!] CPAC says: Could not find or open the motion "\
"parameter file. This is necessary if you have " \
"included any of the MeanFD measures in your group " \
"model.\n\nThis file can usually be found in the " \
"output directory of your individual-level analysis " \
"runs. If it is not there, double-check to see if " \
"individual-level analysis had completed successfully."\
"\n\nPath not found: %s\n\n" % parameter_file
raise Exception(err)
power_params_files[series_id] = parameter_file
break
else:
power_params_files = None
# path to the pipeline folder to be passed to create_fsl_model.py
# so that certain files like output_means.csv can be accessed
pipeline_path = os.path.join(c.outputDirectory, pipeline_ID)
# generate working directory for this output's group analysis run
workDir = '%s/group_analysis/%s/%s' % (c.workingDirectory, \
group_conf.model_name, \
resource)
# this makes strgy_path basically the directory path of the folders after
# the resource/derivative folder level
strgy_path = os.path.dirname(output_file_list[0]).split(resource)[1]
# get rid of periods in the path
for ch in ['.']:
if ch in strgy_path:
strgy_path = strgy_path.replace(ch, "")
# create nipype-workflow-name-friendly strgy_path
# (remove special characters)
strgy_path_name = strgy_path.replace('/', "_")
workDir = workDir + '/' + strgy_path_name
# merge the subjects for this current output
# then, take the group mask, and iterate over the list of subjects
# to extract the mean of each subject using the group mask
merge_output, merge_mask_output, merge_output_dir = \
create_merged_files(workDir, resource, output_file_list)
# CALCULATE THE MEANS of each output using the group mask
derivative_means_dict, roi_means_dict = \
calculate_output_means(resource, output_file_list, \
group_conf.mean_mask, \
group_conf.design_formula, \
group_conf.custom_roi_mask, pipeline_path, \
merge_output_dir, c.identityMatrix)
measure_dict = {}
# extract motion measures from CPAC-generated power params file
if power_params_files != None:
for param_file in power_params_files.values():
new_measure_dict = get_measure_dict(param_file)
measure_dict.update(new_measure_dict)
# combine the motion measures dictionary with the measure_mean
# dictionary (if it exists)
if derivative_means_dict:
measure_dict["Measure_Mean"] = derivative_means_dict
# run create_fsl_model.py to generate the group analysis models
from CPAC.utils import create_fsl_model, kill_me
create_fsl_model.run(group_conf, resource, parameter_file, \
derivative_means_dict, roi_means_dict, \
current_mod_path, True)
# begin GA workflow setup
if not os.path.exists(new_sub_file):
raise Exception("path to input subject list %s is invalid" % new_sub_file)
#if c.mixedScanAnalysis == True:
# wf = pe.Workflow(name = 'group_analysis/%s/grp_model_%s'%(resource, os.path.basename(model)))
#else:
wf = pe.Workflow(name = resource)
wf.base_dir = workDir
wf.config['execution'] = {'hash_method': 'timestamp', 'crashdump_dir': os.path.abspath(c.crashLogDirectory)}
log_dir = os.path.join(group_conf.output_dir, 'logs', 'group_analysis', resource, 'model_%s' % (group_conf.model_name))
if not os.path.exists(log_dir):
os.makedirs(log_dir)
else:
pass
# gp_flow
# Extracts the model files (.con, .grp, .mat, .fts) from the model
# directory and sends them to the create_group_analysis workflow gpa_wf
gp_flow = create_grp_analysis_dataflow("gp_dataflow_%s" % resource)
gp_flow.inputs.inputspec.grp_model = os.path.join(mod_path, resource)
gp_flow.inputs.inputspec.model_name = group_conf.model_name
gp_flow.inputs.inputspec.ftest = fTest
# gpa_wf
# Creates the actual group analysis workflow
gpa_wf = create_group_analysis(fTest, "gp_analysis_%s" % resource)
gpa_wf.inputs.inputspec.merged_file = merge_output
gpa_wf.inputs.inputspec.merge_mask = merge_mask_output
gpa_wf.inputs.inputspec.z_threshold = z_threshold
gpa_wf.inputs.inputspec.p_threshold = p_threshold
gpa_wf.inputs.inputspec.parameters = (c.FSLDIR, 'MNI152')
wf.connect(gp_flow, 'outputspec.mat',
gpa_wf, 'inputspec.mat_file')
wf.connect(gp_flow, 'outputspec.con',
gpa_wf, 'inputspec.con_file')
wf.connect(gp_flow, 'outputspec.grp',
gpa_wf, 'inputspec.grp_file')
if fTest:
wf.connect(gp_flow, 'outputspec.fts',
gpa_wf, 'inputspec.fts_file')
# ds
# Creates the datasink node for group analysis
ds = pe.Node(nio.DataSink(), name='gpa_sink')
if 'sca_roi' in resource:
out_dir = os.path.join(out_dir, \
re.search('sca_roi_(\d)+',os.path.splitext(os.path.splitext(os.path.basename(output_file_list[0]))[0])[0]).group(0))
if 'dr_tempreg_maps_zstat_files_to_standard_smooth' in resource:
out_dir = os.path.join(out_dir, \
re.search('temp_reg_map_z_(\d)+',os.path.splitext(os.path.splitext(os.path.basename(output_file_list[0]))[0])[0]).group(0))
if 'centrality' in resource:
names = ['degree_centrality_binarize', 'degree_centrality_weighted', \
'eigenvector_centrality_binarize', 'eigenvector_centrality_weighted', \
'lfcd_binarize', 'lfcd_weighted']
for name in names:
if name in os.path.basename(output_file_list[0]):
out_dir = os.path.join(out_dir, name)
break
if 'tempreg_maps' in resource:
out_dir = os.path.join(out_dir, \
re.search('\w*[#]*\d+', os.path.splitext(os.path.splitext(os.path.basename(output_file_list[0]))[0])[0]).group(0))
# if c.mixedScanAnalysis == True:
# out_dir = re.sub(r'(\w)*scan_(\w)*(\d)*(\w)*[/]', '', out_dir)
ds.inputs.base_directory = out_dir
ds.inputs.container = ''
ds.inputs.regexp_substitutions = [(r'(?<=rendered)(.)*[/]','/'),
(r'(?<=model_files)(.)*[/]','/'),
(r'(?<=merged)(.)*[/]','/'),
(r'(?<=stats/clusterMap)(.)*[/]','/'),
(r'(?<=stats/unthreshold)(.)*[/]','/'),
(r'(?<=stats/threshold)(.)*[/]','/'),
(r'_cluster(.)*[/]',''),
(r'_slicer(.)*[/]',''),
(r'_overlay(.)*[/]','')]
########datasink connections#########
if fTest:
wf.connect(gp_flow, 'outputspec.fts',
ds, 'model_files.@0')
wf.connect(gp_flow, 'outputspec.mat',
ds, 'model_files.@1' )
wf.connect(gp_flow, 'outputspec.con',
ds, 'model_files.@2')
wf.connect(gp_flow, 'outputspec.grp',
ds, 'model_files.@3')
wf.connect(gpa_wf, 'outputspec.merged',
ds, 'merged')
wf.connect(gpa_wf, 'outputspec.zstats',
ds, 'stats.unthreshold')
wf.connect(gpa_wf, 'outputspec.zfstats',
ds,'stats.unthreshold.@01')
wf.connect(gpa_wf, 'outputspec.fstats',
ds,'stats.unthreshold.@02')
wf.connect(gpa_wf, 'outputspec.cluster_threshold_zf',
ds, 'stats.threshold')
wf.connect(gpa_wf, 'outputspec.cluster_index_zf',
ds,'stats.clusterMap')
wf.connect(gpa_wf, 'outputspec.cluster_localmax_txt_zf',
ds, 'stats.clusterMap.@01')
wf.connect(gpa_wf, 'outputspec.overlay_threshold_zf',
ds, 'rendered')
wf.connect(gpa_wf, 'outputspec.rendered_image_zf',
ds, 'rendered.@01')
wf.connect(gpa_wf, 'outputspec.cluster_threshold',
ds, 'stats.threshold.@01')
wf.connect(gpa_wf, 'outputspec.cluster_index',
ds, 'stats.clusterMap.@02')
wf.connect(gpa_wf, 'outputspec.cluster_localmax_txt',
ds, 'stats.clusterMap.@03')
wf.connect(gpa_wf, 'outputspec.overlay_threshold',
ds, 'rendered.@02')
wf.connect(gpa_wf, 'outputspec.rendered_image',
ds, 'rendered.@03')
######################################
# Run the actual group analysis workflow
wf.run()
print "\n\nWorkflow finished for model %s and resource %s\n\n" \
% (os.path.basename(group_conf.output_dir), resource)
def run(config_file, subject_list_file, p_name = None):
try:
if not os.path.exists(config_file):
raise IOError
else:
c = Configuration(yaml.load(open(os.path.realpath(config_file), 'r')))
except IOError:
print "config file %s doesn't exist" %config_file
raise
except Exception:
raise Exception("Error reading config file - %s"%config_file)
#do some validation
validate(c)
try:
sublist = yaml.load(open(os.path.realpath(subject_list_file), 'r'))
except:
raise Exception ("Subject list is not in proper YAML format. Please check your file")
strategies = sorted(build_strategies(c))
print "strategies ---> ", strategies
sub_ids =[]
for sub in sublist:
if sub['unique_id']:
sub_ids.append(sub['subject_id']+"_" + sub["unique_id"])
else:
sub_ids.append(sub['subject_id'])
create_group_log_template(sub_ids, os.path.join(c.outputDirectory, 'logs'))
seeds_created = []
if not (c.seedSpecificationFile is None):
try:
if os.path.exists(c.seedSpecificationFile):
seeds_created = create_seeds_(c.seedOutputLocation, c.seedSpecificationFile, c.FSLDIR)
print 'seeds created %s -> ' % seeds_created
except:
raise IOError('Problem in seedSpecificationFile')
if 1 in c.runVoxelTimeseries:
if 2 in c.useSeedInAnalysis:
c.maskSpecificationFile = append_seeds_to_file(c.workingDirectory, seeds_created, c.maskSpecificationFile)
if 1 in c.runROITimeseries:
if 1 in c.useSeedInAnalysis:
c.roiSpecificationFile = append_seeds_to_file(c.workingDirectory, seeds_created, c.roiSpecificationFile)
if 1 in c.runNetworkCentrality:
if 3 in c.useSeedInAnalysis:
c.templateSpecificationFile = append_seeds_to_file(c.workingDirectory, seeds_created, c.templateSpecificationFile)
if not c.runOnGrid:
from CPAC.pipeline.cpac_pipeline import prep_workflow
procss = [Process(target=prep_workflow, args=(sub, c, strategies, p_name)) for sub in sublist]
pid = open(os.path.join(c.outputDirectory, 'pid.txt'), 'w')
import subprocess
jobQueue = []
if len(sublist) <= c.numSubjectsAtOnce:
"""
Stream all the subjects as sublist is
less than or equal to the number of
subjects that need to run
"""
for p in procss:
p.start()
print >>pid,p.pid
else:
"""
Stream the subject worlflows for preprocessing.
At Any time in the pipeline c.numSubjectsAtOnce
will run, unless the number remaining is less than
the value of the parameter stated above
"""
idx = 0
while(idx < len(sublist)):
if len(jobQueue) == 0 and idx == 0:
idc = idx
for p in procss[idc: idc + c.numSubjectsAtOnce]:
p.start()
print >>pid,p.pid
jobQueue.append(p)
idx += 1
else:
for job in jobQueue:
if not job.is_alive():
print 'found dead job ', job
loc = jobQueue.index(job)
del jobQueue[loc]
procss[idx].start()
jobQueue.append(procss[idx])
idx += 1
pid.close()
else:
import commands
import pickle
from time import strftime
temp_files_dir = os.path.join(os.getcwd(), 'cluster_temp_files')
print commands.getoutput("mkdir -p %s" % temp_files_dir)
strategies_file = os.path.join(temp_files_dir, 'strategies.obj')
f = open(strategies_file, 'w')
pickle.dump(strategies, f)
f.close()
if 'sge' in c.resourceManager.lower():
run_sge_jobs(c, config_file, strategies_file, subject_list_file, p_name)
elif 'pbs' in c.resourceManager.lower():
run_pbs_jobs(c, config_file, strategies_file, subject_list_file, p_name)
elif 'condor' in c.resourceManager.lower():
run_condor_jobs(c, config_file, strategies_file, subject_list_file, p_name)
return 1
def run_cpac_on_cluster(config_file, subject_list_file, strategies_file,
cluster_files_dir):
'''
Function to build a SLURM batch job submission script and
submit it to the scheduler via 'sbatch'
'''
# Import packages
import commands
import getpass
import re
from time import strftime
from CPAC.utils import Configuration
from indi_schedulers import cluster_templates
# Load in pipeline config
try:
pipeline_dict = yaml.load(open(os.path.realpath(config_file), 'r'))
pipeline_config = Configuration(pipeline_dict)
except:
raise Exception('Pipeline config is not in proper YAML format. '\
'Please check your file')
# Load in the subject list
try:
sublist = yaml.load(open(os.path.realpath(subject_list_file), 'r'))
except:
raise Exception('Subject list is not in proper YAML format. '\
'Please check your file')
# Init variables
timestamp = str(strftime("%Y_%m_%d_%H_%M_%S"))
job_scheduler = pipeline_config.resourceManager.lower()
# For SLURM time limit constraints only, hh:mm:ss
hrs_limit = 8 * len(sublist)
time_limit = '%d:00:00' % hrs_limit
# Batch file variables
shell = commands.getoutput('echo $SHELL')
user_account = getpass.getuser()
num_subs = len(sublist)
# Run CPAC via python -c command
python_cpac_str = 'python -c "from CPAC.pipeline.cpac_pipeline import run; '\
'run(\'%(config_file)s\', \'%(subject_list_file)s\', '\
'%(env_arr_idx)s, \'%(strategies_file)s\', '\
'\'%(pipeline_name)s\', plugin=\'MultiProc\', '\
'plugin_args=%(plugin_args)s)"'
# Init plugin arguments
plugin_args = {
'n_procs': pipeline_config.numCoresPerSubject,
'memory_gb': pipeline_config.memoryAllocatedPerSubject
}
# Set up run command dictionary
run_cmd_dict = {
'config_file': config_file,
'subject_list_file': subject_list_file,
'strategies_file': strategies_file,
'pipeline_name': pipeline_config.pipelineName,
'plugin_args': plugin_args
}
# Set up config dictionary
config_dict = {
'timestamp': timestamp,
'shell': shell,
'job_name': 'CPAC_' + pipeline_config.pipelineName,
'num_tasks': num_subs,
'queue': pipeline_config.queue,
'par_env': pipeline_config.parallelEnvironment,
'cores_per_task': pipeline_config.numCoresPerSubject,
'user': user_account,
'work_dir': cluster_files_dir,
'time_limit': time_limit
}
# Get string template for job scheduler
if job_scheduler == 'pbs':
env_arr_idx = '$PBS_ARRAYID'
batch_file_contents = cluster_templates.pbs_template
confirm_str = '(?<=Your job-array )\d+'
exec_cmd = 'qsub'
elif job_scheduler == 'sge':
env_arr_idx = '$SGE_TASK_ID'
batch_file_contents = cluster_templates.sge_template
confirm_str = '(?<=Your job-array )\d+'
exec_cmd = 'qsub'
elif job_scheduler == 'slurm':
env_arr_idx = '$SLURM_ARRAY_TASK_ID'
batch_file_contents = cluster_templates.slurm_template
confirm_str = '(?<=Submitted batch job )\d+'
exec_cmd = 'sbatch'
# Populate rest of dictionary
config_dict['env_arr_idx'] = env_arr_idx
run_cmd_dict['env_arr_idx'] = env_arr_idx
config_dict['run_cmd'] = python_cpac_str % run_cmd_dict
# Populate string from config dict values
batch_file_contents = batch_file_contents % config_dict
# Write file
batch_filepath = os.path.join(cluster_files_dir, 'cpac_submit_%s.%s' \
% (timestamp, job_scheduler))
with open(batch_filepath, 'w') as f:
f.write(batch_file_contents)
# Get output response from job submission
out = commands.getoutput('%s %s' % (exec_cmd, batch_filepath))
# Check for successful qsub submission
if re.search(confirm_str, out) == None:
err_msg = 'Error submitting C-PAC pipeline run to %s queue' \
% job_scheduler
raise Exception(err_msg)
# Get pid and send to pid file
pid = re.search(confirm_str, out).group(0)
pid_file = os.path.join(cluster_files_dir, 'pid.txt')
with open(pid_file, 'w') as f:
f.write(pid)
def prep_group_analysis_workflow(c, resource, subject_infos):
p_id, s_ids, scan_ids, s_paths = (list(tup) for tup in zip(*subject_infos))
#print "p_id -%s, s_ids -%s, scan_ids -%s, s_paths -%s" %(p_id, s_ids, scan_ids, s_paths)
'''
#diag = open(os.path.join('/home/data/Projects/CPAC_Regression_Test/2013-08-19-20_v0-3-1/fsl-model/2013-09-03', 'group_analysis_diagnostic.txt'), 'wt')
#for tup in subject_infos:
# print >>diag, list(tup)
#print >>diag, ""
#for tup in zip(*subject_infos):
# print >>diag, list(tup)
#print >>diag, ""
print >>diag, "Working variables passed from cpac_group_runner: "
print >>diag, ""
print >>diag, "Pipeline ID (p_id): ", p_id
print >>diag, "Subject IDs (s_ids): ", s_ids
print >>diag, "Scan IDs (scan_ids): ", scan_ids
print >>diag, "(s_paths): ", s_paths
print >>diag, ""
'''
def get_phenotypic_file(phenotypic_file, m_dict, m_list, mod_path, sub_id):
#print "phenotypic_file, m_dict", phenotypic_file, m_dict
import csv
reader = csv.reader(open(phenotypic_file, 'rU'))
columns = {}
order = {}
count = 0
headers = reader.next()
for h in headers:
columns[h] =[]
order[h] = count
count+=1
for r in reader:
for h, v in zip(headers, r):
if v:
columns[h].append(str(v))
if m_dict:
for measure in m_list:
if measure in headers:
#check if 'MeanFD is present'
if len(columns[measure]) < 1:
for sub in columns[sub_id]:
if m_dict.get(sub):
if m_dict.get(sub).get(measure):
columns[measure].append(m_dict[sub][measure])
else:
raise Exception("Couldn't find %s value for subject %s"%(measure,sub))
else:
raise Exception("Couldn't find subject %s in the parameter file"%sub)
b = zip(*([k] + columns[k] for k in sorted(columns, key=order.get)))
try:
os.makedirs(mod_path)
except:
print "%s already exists"%(mod_path)
new_phenotypic_file = os.path.join(mod_path, os.path.basename(phenotypic_file))
a = csv.writer(open(new_phenotypic_file, 'w'))
for col in b:
a.writerow(list(col))
return new_phenotypic_file
threshold_val = None
measure_dict = None
measure_list = ['MeanFD', 'MeanFD_Jenkinson', 'MeanDVARS']
model_sub_list = []
if c.runScrubbing == 1:
#get scrubbing threshold
if re.search('(?<=/_threshold_)\d+.\d+',s_paths[0]):
threshold_val = re.search('(?<=/_threshold_)\d+.\d+',s_paths[0]).group(0)
elif len(c.scrubbingThreshold) == 1:
threshold_val = c.scrubbingThreshold[0]
else:
print "Found Multiple threshold value "
print "scrubbing threshold_val -->", threshold_val
else:
print "No scrubbing enabled."
#pick the right parameter file from the pipeline folder
#create a dictionary of subject and measures in measure_list
if c.runScrubbing == 1:
try:
parameter_file = os.path.join(c.outputDirectory, p_id[0], '%s_threshold_%s_all_params.csv'%(scan_ids[0].strip('_'),threshold_val))
if os.path.exists(parameter_file):
import csv
measure_dict = {}
f = csv.DictReader(open(parameter_file,'r'))
for line in f:
measure_map = {}
for m in measure_list:
if line.get(m):
measure_map[m] = line[m]
measure_dict[line['Subject']] = measure_map
else:
print "No file name %s found"%parameter_file
except Exception:
print "Exception while extracting parameters from movement file - %s"%(parameter_file)
#print >>diag, "Begins to iterate over each config file listed here: ", c.modelConfigs
#print >>diag, ""
for config in c.modelConfigs:
import yaml
try:
conf = Configuration(yaml.load(open(os.path.realpath(config), 'r')))
except:
raise Exception("Error in reading %s configuration file" % config)
#print >>diag, "Starting iteration for config: ", config
#print >>diag, ""
group_sublist = open(conf.subjectListFile, 'r')
sublist_items = group_sublist.readlines()
subject_list = [line.rstrip('\n') for line in sublist_items \
if not (line == '\n') and not line.startswith('#')]
#print >>diag, "Subject list run-through #1: ", subject_list
#print >>diag, ""
#subject_list = [line.rstrip('\r\n') for line in open(conf.subjectListFile, 'r') \
# if not (line == '\n') and not line.startswith('#')]
# list of subject paths which DO exist
exist_paths = []
# check for missing subject for the derivative
#print >>diag, "> Iterates over subject_list - for each subject in the subject list, it iterates over the paths in s_paths."
#print >>diag, "> For each path, it checks if the current subject exists in this path, and then appends this subject to 'exist_paths' list."
#print >>diag, ""
for sub in subject_list :
for path in s_paths:
if sub in path:
exist_paths.append(sub)
#print >>diag, "Current status of exist_paths list: "
#print >>diag, exist_paths
#print >>diag, ""
# check to see if any derivatives of subjects are missing
if len(list(set(subject_list) - set(exist_paths))) >0:
print "List of outputs missing for subjects:"
print list(set(subject_list) - set(exist_paths))
print "..for derivatives:"
print resource
print "..at paths:"
print os.path.dirname(s_paths[0]).replace(s_ids[0], '*')
#import warnings
#warnings.warn(msg)
mod_path = os.path.join(os.path.dirname(s_paths[0]).replace(s_ids[0], 'group_analysis_results/_grp_model_%s'%(conf.modelName)),
'model_files')
#print >>diag, "> Created mod_path variable: ", mod_path
#print >>diag, ""
print "basename: ", os.path.basename(conf.subjectListFile)
try:
os.makedirs(mod_path)
print "Creating directory:"
print mod_path
except:
print "Attempted to create directory, but path already exists:"
print mod_path
new_sub_file = os.path.join(mod_path, os.path.basename(conf.subjectListFile))
try:
f = open(new_sub_file, 'w')
for sub in exist_paths:
print >>f, sub
f.close()
#print >>diag, "> Created new subject list file: ", new_sub_file
#print >>diag, ""
#print >>diag, "> ..which is filled with the subjects from exist_paths"
#print >>diag, ""
except:
print "Error: Could not open subject list file: ", new_sub_file
raise Exception
#print >>diag, "> Updates the FSL model config's subject list file parameter from: ", conf.subjectListFile
conf.update('subjectListFile',new_sub_file)
#print >>diag, "> ..to new subject list file: ", conf.subjectListFile
#print >>diag, ""
sub_id = conf.subjectColumn
'''
print >>diag, "> If measure_dict is not empty, it updates the phenotypic file with these parameters: "
print >>diag, ""
print >>diag, "measure_dict: ", measure_dict
print >>diag, "measure_list: ", measure_list
print >>diag, "mod_path: ", mod_path
print >>diag, "sub_id: ", sub_id
print >>diag, ""
'''
if measure_dict != None:
conf.update('phenotypicFile',get_phenotypic_file(conf.phenotypicFile, measure_dict, measure_list, mod_path, sub_id))
print "Model config dictionary ->"
print conf.__dict__
# Run 'create_fsl_model' script to extract phenotypic data from
# the phenotypic file for each of the subjects in the subject list
try:
from CPAC.utils import create_fsl_model
create_fsl_model.run(conf, c.fTest, True)
#print >>diag, "> Runs create_fsl_model."
#print >>diag, ""
except Exception, e:
print "FSL Group Analysis model not successfully created - error in create_fsl_model script"
#print "Error ->", e
raise
model_sub_list.append((conf.outputModelFilesDirectory, conf.subjectListFile))
print "model_sub_list ->", model_sub_list
'''
def run(config_file,
subject_list_file,
p_name=None,
plugin=None,
plugin_args=None):
'''
'''
# Import packages
import commands
import os
import pickle
import time
from CPAC.pipeline.cpac_pipeline import prep_workflow
# Init variables
config_file = os.path.realpath(config_file)
subject_list_file = os.path.realpath(subject_list_file)
# take date+time stamp for run identification purposes
unique_pipeline_id = strftime("%Y%m%d%H%M%S")
pipeline_start_stamp = strftime("%Y-%m-%d_%H:%M:%S")
# Load in pipeline config file
try:
if not os.path.exists(config_file):
raise IOError
else:
c = Configuration(yaml.load(open(config_file, 'r')))
except IOError:
print "config file %s doesn't exist" % config_file
raise
except Exception:
print "Error reading config file - %s" % config_file
raise Exception
# Do some validation
validate(c)
# Get the pipeline name
p_name = c.pipelineName
# Load in subject list
try:
sublist = yaml.load(open(subject_list_file, 'r'))
except:
print "Subject list is not in proper YAML format. Please check your file"
raise Exception
# NOTE: strategies list is only needed in cpac_pipeline prep_workflow for
# creating symlinks
strategies = sorted(build_strategies(c))
# Populate subject scan map
sub_scan_map = {}
try:
for sub in sublist:
if sub['unique_id']:
s = sub['subject_id'] + "_" + sub["unique_id"]
else:
s = sub['subject_id']
scan_ids = ['scan_anat']
try:
for id in sub['func']:
scan_ids.append('scan_' + str(id))
except KeyError:
for id in sub['rest']:
scan_ids.append('scan_' + str(id))
sub_scan_map[s] = scan_ids
except:
print "\n\n" + "ERROR: Subject list file not in proper format - " \
"check if you loaded the correct file?" + "\n" + \
"Error name: cpac_runner_0001" + "\n\n"
raise Exception
create_group_log_template(sub_scan_map, c.logDirectory)
pipeline_timing_info = []
pipeline_timing_info.append(unique_pipeline_id)
pipeline_timing_info.append(pipeline_start_stamp)
pipeline_timing_info.append(len(sublist))
# If we're running on cluster, execute job scheduler
if c.runOnGrid:
# Create cluster log dir
cluster_files_dir = os.path.join(c.logDirectory, 'cluster_files')
if not os.path.exists(cluster_files_dir):
os.makedirs(cluster_files_dir)
# Create strategies file
strategies_file = os.path.join(cluster_files_dir, 'strategies.obj')
with open(strategies_file, 'w') as f:
pickle.dump(strategies, f)
# Check if its a condor job, and run that
if 'condor' in c.resourceManager.lower():
run_condor_jobs(c, config_file, strategies_file, subject_list_file,
p_name)
# All other schedulers are supported
else:
run_cpac_on_cluster(config_file, subject_list_file,
strategies_file, cluster_files_dir)
# Run on one computer
else:
# Init variables
procss = [Process(target=prep_workflow,
args=(sub, c, strategies, 1,
pipeline_timing_info, p_name, plugin, plugin_args)) \
for sub in sublist]
if not os.path.exists(c.workingDirectory):
try:
os.makedirs(c.workingDirectory)
except:
err = "\n\n[!] CPAC says: Could not create the working " \
"directory: %s\n\nMake sure you have permissions " \
"to write to this directory.\n\n" % c.workingDirectory
raise Exception(err)
pid = open(os.path.join(c.workingDirectory, 'pid.txt'), 'w')
# Init job queue
jobQueue = []
# If we're allocating more processes than are subjects, run them all
if len(sublist) <= c.numParticipantsAtOnce:
for p in procss:
p.start()
print >> pid, p.pid
# Otherwise manage resources to run processes incrementally
else:
idx = 0
while (idx < len(sublist)):
# If the job queue is empty and we haven't started indexing
if len(jobQueue) == 0 and idx == 0:
# Init subject process index
idc = idx
# Launch processes (one for each subject)
for p in procss[idc:idc + c.numParticipantsAtOnce]:
p.start()
print >> pid, p.pid
jobQueue.append(p)
idx += 1
# Otherwise, jobs are running - check them
else:
# Check every job in the queue's status
for job in jobQueue:
# If the job is not alive
if not job.is_alive():
# Find job and delete it from queue
print 'found dead job ', job
loc = jobQueue.index(job)
del jobQueue[loc]
# ...and start the next available process (subject)
procss[idx].start()
# Append this to job queue and increment index
jobQueue.append(procss[idx])
idx += 1
# Add sleep so while loop isn't consuming 100% of CPU
time.sleep(2)
# Close PID txt file to indicate finish
pid.close()
def run(config_file, subject_list_file, p_name=None, plugin=None,
plugin_args=None, tracking=True, num_subs_at_once=None, debug=False):
'''
'''
# Import packages
import commands
import os
import pickle
import time
from CPAC.pipeline.cpac_pipeline import prep_workflow
# Init variables
config_file = os.path.realpath(config_file)
subject_list_file = os.path.realpath(subject_list_file)
# take date+time stamp for run identification purposes
unique_pipeline_id = strftime("%Y%m%d%H%M%S")
pipeline_start_stamp = strftime("%Y-%m-%d_%H:%M:%S")
# Load in pipeline config file
try:
if not os.path.exists(config_file):
raise IOError
else:
c = Configuration(yaml.load(open(config_file, 'r')))
except IOError:
print "config file %s doesn't exist" % config_file
raise
except Exception as e:
raise Exception("Error reading config file - {0}\n\nError details:"
"\n{1}\n\n".format(config_file, e))
c.logDirectory = os.path.abspath(c.logDirectory)
c.workingDirectory = os.path.abspath(c.workingDirectory)
if 's3://' not in c.outputDirectory:
c.outputDirectory = os.path.abspath(c.outputDirectory)
c.crashLogDirectory = os.path.abspath(c.crashLogDirectory)
if debug:
c.write_debugging_outputs = "[1]"
if num_subs_at_once:
if not str(num_subs_at_once).isdigit():
raise Exception('[!] Value entered for --num_cores not a digit.')
c.numParticipantsAtOnce = int(num_subs_at_once)
# Do some validation
validate(c)
# Get the pipeline name
p_name = p_name or c.pipelineName
# Load in subject list
try:
with open(subject_list_file, 'r') as sf:
sublist = yaml.load(sf)
except:
print "Subject list is not in proper YAML format. Please check " \
"your file"
raise Exception
# Populate subject scan map
sub_scan_map = {}
try:
for sub in sublist:
if sub['unique_id']:
s = sub['subject_id'] + "_" + sub["unique_id"]
else:
s = sub['subject_id']
scan_ids = ['scan_anat']
if 'func' in sub:
for id in sub['func']:
scan_ids.append('scan_'+ str(id))
if 'rest' in sub:
for id in sub['rest']:
scan_ids.append('scan_'+ str(id))
sub_scan_map[s] = scan_ids
except:
print "\n\n" + "ERROR: Subject list file not in proper format - " \
"check if you loaded the correct file?" + "\n" + \
"Error name: cpac_runner_0001" + "\n\n"
raise Exception
pipeline_timing_info = []
pipeline_timing_info.append(unique_pipeline_id)
pipeline_timing_info.append(pipeline_start_stamp)
pipeline_timing_info.append(len(sublist))
if tracking:
track_run(level='participant', participants=len(sublist))
# If we're running on cluster, execute job scheduler
if c.runOnGrid:
# Create cluster log dir
cluster_files_dir = os.path.join(c.logDirectory, 'cluster_files')
if not os.path.exists(cluster_files_dir):
os.makedirs(cluster_files_dir)
# Check if its a condor job, and run that
if 'condor' in c.resourceManager.lower():
run_condor_jobs(c, config_file, subject_list_file, p_name)
# All other schedulers are supported
else:
run_cpac_on_cluster(config_file, subject_list_file, cluster_files_dir)
# Run on one computer
else:
if not os.path.exists(c.workingDirectory):
try:
os.makedirs(c.workingDirectory)
except:
err = "\n\n[!] CPAC says: Could not create the working " \
"directory: %s\n\nMake sure you have permissions " \
"to write to this directory.\n\n" % c.workingDirectory
raise Exception(err)
# If it only allows one, run it linearly
if c.numParticipantsAtOnce == 1:
for sub in sublist:
prep_workflow(sub, c, True, pipeline_timing_info,
p_name, plugin, plugin_args)
return
pid = open(os.path.join(c.workingDirectory, 'pid.txt'), 'w')
# Init job queue
job_queue = []
# Allocate processes
processes = [Process(target=prep_workflow,
args=(sub, c, True, pipeline_timing_info,
p_name, plugin, plugin_args))
for sub in sublist]
# If we're allocating more processes than are subjects, run them all
if len(sublist) <= c.numParticipantsAtOnce:
for p in processes:
p.start()
print >>pid, p.pid
# Otherwise manage resources to run processes incrementally
else:
idx = 0
while idx < len(sublist):
# If the job queue is empty and we haven't started indexing
if len(job_queue) == 0 and idx == 0:
# Init subject process index
idc = idx
# Launch processes (one for each subject)
for p in processes[idc: idc+c.numParticipantsAtOnce]:
p.start()
print >>pid, p.pid
job_queue.append(p)
idx += 1
# Otherwise, jobs are running - check them
else:
# Check every job in the queue's status
for job in job_queue:
# If the job is not alive
if not job.is_alive():
# Find job and delete it from queue
print 'found dead job ', job
loc = job_queue.index(job)
del job_queue[loc]
# ...and start the next available process
# (subject)
processes[idx].start()
# Append this to job queue and increment index
job_queue.append(processes[idx])
idx += 1
# Add sleep so while loop isn't consuming 100% of CPU
time.sleep(2)
# Close PID txt file to indicate finish
pid.close()
def configuration_strategy_mock(method='FSL'):
# mock the config dictionary
c = Configuration({
"num_ants_threads": 4,
"workingDirectory": "/scratch/pipeline_tests",
"crashLogDirectory": "/scratch",
"outputDirectory":
"/output/output/pipeline_analysis_nuisance/sub-M10978008_ses-NFB3",
"resolution_for_func_preproc": "3mm",
"resolution_for_func_derivative": "3mm",
"template_for_resample":
"/usr/share/fsl/5.0/data/standard/MNI152_T1_1mm_brain.nii.gz",
"template_brain_only_for_func":
"/usr/share/fsl/5.0/data/standard/MNI152_T1_${resolution_for_func_preproc}_brain.nii.gz",
"template_skull_for_func":
"/usr/share/fsl/5.0/data/standard/MNI152_T1_${resolution_for_func_preproc}.nii.gz",
"identityMatrix": "/usr/share/fsl/5.0/etc/flirtsch/ident.mat",
"funcRegFSLinterpolation": "sinc",
"funcRegANTSinterpolation": "LanczosWindowedSinc"
})
if method == 'ANTS':
c.update('regOption', 'ANTS')
else:
c.update('regOption', 'FSL')
# mock the strategy
strat = Strategy()
resource_dict = {
"mean_functional":
os.path.join(
c.outputDirectory,
"mean_functional/sub-M10978008_ses-NFB3_task-test_bold_calc_tshift_resample_volreg_calc_tstat.nii.gz"
),
"motion_correct":
os.path.join(
c.outputDirectory,
"motion_correct/_scan_test/sub-M10978008_ses-NFB3_task-test_bold_calc_tshift_resample_volreg.nii.gz"
),
"anatomical_brain":
os.path.join(
c.outputDirectory,
"anatomical_brain/sub-M10978008_ses-NFB3_acq-ao_brain_resample.nii.gz"
),
"ants_initial_xfm":
os.path.join(
c.outputDirectory,
"ants_initial_xfm/transform0DerivedInitialMovingTranslation.mat"),
"ants_affine_xfm":
os.path.join(c.outputDirectory,
"ants_affine_xfm/transform2Affine.mat"),
"ants_rigid_xfm":
os.path.join(c.outputDirectory, "ants_rigid_xfm/transform1Rigid.mat"),
"anatomical_to_mni_linear_xfm":
os.path.join(
c.outputDirectory,
"anatomical_to_mni_linear_xfm/sub-M10978008_ses-NFB3_T1w_resample_calc_flirt.mat"
),
"functional_to_anat_linear_xfm":
os.path.join(
c.outputDirectory,
"functional_to_anat_linear_xfm/_scan_test/sub-M10978008_ses-NFB3_task-test_bold_calc_tshift_resample_volreg_calc_tstat_flirt.mat"
),
'ants_symm_warp_field':
os.path.join(
c.outputDirectory,
"anatomical_to_symmetric_mni_nonlinear_xfm/transform3Warp.nii.gz"),
'ants_symm_affine_xfm':
os.path.join(c.outputDirectory,
"ants_symmetric_affine_xfm/transform2Affine.mat"),
'ants_symm_rigid_xfm':
os.path.join(c.outputDirectory,
"ants_symmetric_rigid_xfm/transform1Rigid.mat"),
'ants_symm_initial_xfm':
os.path.join(
c.outputDirectory,
"ants_symmetric_initial_xfm/transform0DerivedInitialMovingTranslation.mat"
),
"dr_tempreg_maps_files": [
os.path.join(
'/scratch', 'resting_preproc_sub-M10978008_ses-NFB3_cpac105',
'temporal_dual_regression_0/_scan_test/_selector_CSF-2mmE-M_aC-WM-2mmE-DPC5_G-M_M-SDB_P-2/_spatial_map_PNAS_Smith09_rsn10_spatial_map_file_..cpac_templates..PNAS_Smith09_rsn10.nii.gz/split_raw_volumes/temp_reg_map_000{0}.nii.gz'
.format(n)) for n in range(10)
]
}
if method == 'ANTS':
resource_dict["anatomical_to_mni_nonlinear_xfm"] = os.path.join(
c.outputDirectory,
"anatomical_to_mni_nonlinear_xfm/transform3Warp.nii.gz")
else:
resource_dict["anatomical_to_mni_nonlinear_xfm"] = os.path.join(
c.outputDirectory,
"anatomical_to_mni_nonlinear_xfm/sub-M10978008_ses-NFB3_T1w_resample_fieldwarp.nii.gz"
)
file_node_num = 0
for resource, filepath in resource_dict.items():
strat.update_resource_pool(
{resource: file_node(filepath, file_node_num)})
strat.append_name(resource + '_0')
file_node_num += 1
templates_for_resampling = [
(c.resolution_for_func_preproc, c.template_brain_only_for_func,
'template_brain_for_func_preproc', 'resolution_for_func_preproc'),
(c.resolution_for_func_preproc, c.template_brain_only_for_func,
'template_skull_for_func_preproc', 'resolution_for_func_preproc')
]
for resolution, template, template_name, tag in templates_for_resampling:
resampled_template = pe.Node(Function(
input_names=['resolution', 'template', 'template_name', 'tag'],
output_names=['resampled_template'],
function=resolve_resolution,
as_module=True),
name='resampled_' + template_name)
resampled_template.inputs.resolution = resolution
resampled_template.inputs.template = template
resampled_template.inputs.template_name = template_name
resampled_template.inputs.tag = tag
strat.update_resource_pool(
{template_name: (resampled_template, 'resampled_template')})
strat.append_name('resampled_template_0')
return c, strat
