def read_subject_info_list(file_name, separator=Hcp3TSubjectInfo.DEFAULT_SEPARATOR()): """Reads a subject information list from the specified file. :param file_name: name of file from which to read :type file_name: str """ subject_info_list = [] input_file = open(file_name, "r") for line in input_file: # remove new line characters line = str_utils.remove_ending_new_lines(line) # remove leading and trailing spaces line = line.strip() # ignore blank lines and comment lines - starting with # if line != "" and line[0] != "#": (project, subject_id, extra) = line.split(separator) # Make the string 'None' in the file translate to a None type instead of just the # string itself if extra == "None": extra = None subject_info = Hcp3TSubjectInfo(project, subject_id, extra) subject_info_list.append(subject_info) return subject_info_list
def submit_process_data_jobs(self, stage, prior_job=None): module_logger.debug(debug_utils.get_name()) # go ahead and submit the standard process data job and then # submit an additional freesurfer assessor job standard_process_data_jobno, all_process_data_jobs = super().submit_process_data_jobs(stage, prior_job) if OneSubjectJobSubmitter._SUPPRESS_FREESURFER_ASSESSOR_JOB: module_logger.info("freesufer assessor job not submitted because freesurfer assessor creation has been suppressed") return standard_process_data_jobno, all_process_data_jobs if stage >= ccf_processing_stage.ProcessingStage.PROCESS_DATA: if standard_process_data_jobno: fs_submit_cmd = 'qsub -W depend=afterok:' + standard_process_data_jobno + ' ' + self.freesurfer_assessor_script_name else: fs_submit_cmd = 'qsub ' + self.freesurfer_assessor_script_name completed_submit_process = subprocess.run( fs_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) fs_job_no = str_utils.remove_ending_new_lines(completed_submit_process.stdout) all_process_data_jobs.append(fs_job_no) return fs_job_no, all_process_data_jobs else: module_logger.info("freesurfer assessor job not submitted because of requested processing stage") return standard_process_data_jobno, all_process_data_jobs
def read_subject_info_list(file_name, separator=SubjectInfo.DEFAULT_SEPARATOR()): """ Reads a subject information list from the specified file. """ subject_info_list = [] input_file = open(file_name, 'r') for line in input_file: # remove new line characters line = str_utils.remove_ending_new_lines(line) # remove leading and trailing spaces line = line.strip() # ignore blank lines and comment lines - starting with # if line != '' and line[0] != '#': (project, subject_id, classifier, extra) = line.split(separator) # Make the string 'None' in the file translate to a None type instead of # just the string itself if extra == 'None': extra = None subject_info = SubjectInfo(project, subject_id, classifier, extra) subject_info_list.append(subject_info) input_file.close() return subject_info_list
def main(): # create a parser object for getting the command line options parser = my_argparse.MyArgumentParser() # mandatory arguments parser.add_argument('-u', '--user', dest='user', required=True, type=str) parser.add_argument(dest='input_file') # optional arguments parser.add_argument('-ser', '--server', dest='server', required=False, default='https://' + os_utils.getenv_required('XNAT_PBS_JOBS_XNAT_SERVER'), type=str) parser.add_argument('-p', '--password', dest='password', required=False, type=str) # parse the command line arguments args = parser.parse_args() if args.password: password = args.password else: password = getpass.getpass("Password: "******"Parsed arguments:") _inform(" Username: "******" Password: "******"*** password mask ***") _inform(" Server: " + args.server) _inform(" Input File: " + args.input_file) _inform("") input_file = open(args.input_file, 'r') for line in input_file: line = str_utils.remove_ending_new_lines(line) line = line.strip() if line != '' and line[0] != '#': (project, subject, session, resource) = line.split('\t') _inform("") _inform(" Project: " + project) _inform(" Subject: " + subject) _inform(" Session: " + session) _inform(" Resource: " + resource) _inform("") delete_resource.delete_resource(args.user, password, args.server, project, subject, session, resource)
def submit_no_longer_running_jobs(self, stage, prior_job=None): module_logger.debug(debug_utils.get_name()) if prior_job: cmd = 'qsub -W depend=afterany:' + prior_job + ' ' + self.mark_no_longer_running_script_name else: cmd = 'qsub ' + self.mark_no_longer_running_script_name completed_submit_process = subprocess.run( cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) job_no = str_utils.remove_ending_new_lines(completed_submit_process.stdout) return job_no, [job_no]
def make_all_links_into_copies_ext(full_path): xnat_pbs_jobs = os_utils.getenv_required('XNAT_PBS_JOBS') command_str = xnat_pbs_jobs + os.sep + 'lib' + os.sep + 'utils' + os.sep + 'make_symlinks_into_copies.sh' + ' ' + full_path completed_subprocess = subprocess.run(command_str, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) output = str_utils.remove_ending_new_lines(completed_subprocess.stdout) print(output)
def read_subject_info_list(file_name, separator=Hcp7TSubjectInfo.DEFAULT_SEPARATOR()): """Reads a subject information list from the specified file. :param file_name: name of file from which to read :type file_name: str """ subject_info_list = [] input_file = open(file_name, 'r') for line in input_file: # remove new line characters line = str_utils.remove_ending_new_lines(line) # remove leading and trailing spaces line = line.strip() # ignore blank lines and comment lines - starting with # if line != '' and line[0] != '#': try: (project, structural_ref_project, subject_id, extra) = line.split(separator) except ValueError as e: if str(e) == 'not enough values to unpack (expected 4, got 3)': (project, structural_ref_project, subject_id) = line.split(separator) extra = None else: raise # Make the string 'None' in the file translate to a None type instead of just the # string itself if project == 'None': project = None if structural_ref_project == 'None': structural_ref_project = None if subject_id == 'None': subject_id = None if extra == 'None': extra = None subject_info = Hcp7TSubjectInfo(project, structural_ref_project, subject_id, extra) subject_info_list.append(subject_info) input_file.close() return subject_info_list
def submit_check_jobs(self, stage, prior_job=None): module_logger.debug(debug_utils.get_name()) if stage >= ccf_processing_stage.ProcessingStage.CHECK_DATA: if prior_job: check_submit_cmd = 'qsub -W depend=afterok:' + prior_job + ' ' + self.check_data_job_script_name else: check_submit_cmd = 'qsub ' + self.check_data_job_script_name completed_submit_process = subprocess.run( check_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) check_job_no = str_utils.remove_ending_new_lines(completed_submit_process.stdout) return check_job_no, [check_job_no] else: module_logger.info("Check data job not submitted") return None, None
def main(): # create a parser object for getting the command line options parser = my_argparse.MyArgumentParser() # mandatory arguments parser.add_argument("-u", "--user", dest="user", required=True, type=str) parser.add_argument(dest="input_file") # optional arguments parser.add_argument( "-ser", "--server", dest="server", required=False, default="https://db.humanconnectome.org", type=str ) parser.add_argument("-p", "--password", dest="password", required=False, type=str) # parse the command line arguments args = parser.parse_args() if args.password: password = args.password else: password = getpass.getpass("Password: "******"Parsed arguments:") _inform(" Username: "******" Password: "******"*** password mask ***") _inform(" Server: " + args.server) _inform(" Input File: " + args.input_file) _inform("") input_file = open(args.input_file, "r") for line in input_file: line = str_utils.remove_ending_new_lines(line) line = line.strip() if line != "" and line[0] != "#": (project, subject, session, resource) = line.split("\t") _inform("") _inform(" Project: " + project) _inform(" Subject: " + subject) _inform(" Session: " + session) _inform(" Resource: " + resource) _inform("") delete_resource.delete_resource(args.user, password, args.server, project, subject, session, resource)
def _get_volume_count(self, file_name): if not os.path.isfile(file_name): return 0 cmd = 'fslinfo ' + file_name + ' | grep dim4 | grep -v pix' completed_process = subprocess.run(cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) # remove new line characters and leading and trailing spaces output = str_utils.remove_ending_new_lines(completed_process.stdout).strip() # collapse whitespace to single spaces output = re.sub('\W+', ' ', output) log.debug("output: " + output) (name, value) = output.split(' ') log.debug("name: " + name) log.debug("value: " + value) return int(value)
def read_subject_id_list(file_name): """Reads a subject id list from the specified file.""" subject_id_list = [] input_file = open(file_name, 'r') for line in input_file: # remove new line characters line = str_utils.remove_ending_new_lines(line) # remove leading and trailing spaces line = line.strip() # ignore blank lines and comment lines - starting with # if line != '' and line[0] != '#': subject_id = line subject_id_list.append(subject_id) input_file.close() return subject_id_list
def _get_expected_volume_count(self, file_name): sum_col1 = 0 sum_col2 = 0 series_file = open(file_name, 'r') for line in series_file: # remove new line characters and leading and trailing spaces line = str_utils.remove_ending_new_lines(line).strip() log.debug("line: " + line) # get 2 values from the line (num1, num2) = line.split(' ') log.debug("num1: " + num1) log.debug("num2: " + num2) sum_col1 += int(num1) sum_col2 += int(num2) series_file.close() return min(sum_col1, sum_col2)
def read_subject_info_list(file_name, separator=Hcp7TSubjectInfo.DEFAULT_SEPARATOR()): """Reads a subject information list from the specified file. :param file_name: name of file from which to read :type file_name: str """ subject_info_list = [] input_file = open(file_name, 'r') for line in input_file: # remove new line characters line = str_utils.remove_ending_new_lines(line) # remove leading and trailing spaces line = line.strip() # ignore blank lines and comment lines - starting with # if line != '' and line[0] != '#': try: (project, structural_ref_project, subject_id, extra) = line.split(separator) except ValueError as e: if str(e) == 'not enough values to unpack (expected 4, got 3)': (project, structural_ref_project, subject_id) = line.split(separator) extra = None else: raise # Make the string 'None' in the file translate to a None type instead of just the # string itself if extra == 'None': extra = None subject_info = Hcp7TSubjectInfo(project, structural_ref_project, subject_id, extra) subject_info_list.append(subject_info) input_file.close() return subject_info_list
def submit_jobs(self, processing_stage=ProcessingStage.PUT_DATA): logger.debug("submit_jobs: processing_stage: " + str(processing_stage)) logger.info("----------") logger.info("Submitting " + self.PIPELINE_NAME + " jobs for") logger.info(" Project: " + self.project) logger.info(" Subject: " + self.subject) logger.info(" Session: " + self.session) logger.info(" Stage: " + str(processing_stage)) logger.info("----------") # make sure working directories do not have the same name based on # the same start time by sleeping a few seconds time.sleep(5) # build the working directory name self._working_directory_name = \ self.build_working_directory_name(self.project, self.PIPELINE_NAME, self.subject) logger.info("Making working directory: " + self._working_directory_name) os.makedirs(name=self._working_directory_name) # # get JSESSION ID # jsession_id = xnat_access.get_jsession_id( # server=os_utils.getenv_required('XNAT_PBS_JOBS_XNAT_SERVER'), # username=self.username, # password=self.password) # logger.info("jsession_id: " + jsession_id) # # get XNAT Session ID (a.k.a. the experiment ID, e.g. ConnectomeDB_E1234) # xnat_session_id = xnat_access.get_session_id( # server=os_utils.getenv_required('XNAT_PBS_JOBS_XNAT_SERVER'), # username=self.username, # password=self.password, # project=self.project, # subject=self.subject, # session=self.session) # logger.info("xnat_session_id: " + xnat_session_id) # # get XNAT Workflow ID # workflow_obj = xnat_access.Workflow(self.username, self.password, # os_utils.getenv_required('XNAT_PBS_JOBS_XNAT_SERVER'), # jsession_id) # self._workflow_id = workflow_obj.create_workflow(xnat_session_id, # self.project, # self.PIPELINE_NAME, # 'Queued') # logger.info("workflow_id: " + self._workflow_id) # determine output resource name self._output_resource_name = "Diffusion_bedpostx" # clean output resource if requested if self.clean_output_resource_first: logger.info("Deleting resource: " + self._output_resource_name + " for:") logger.info(" project: " + self.project) logger.info(" subject: " + self.subject) logger.info(" session: " + self.session) delete_resource.delete_resource( self.username, self.password, str_utils.get_server_name(self.server), self.project, self.subject, self.session, self._output_resource_name) # create scripts for various stages of processing if processing_stage >= ProcessingStage.PREPARE_SCRIPTS: # create script to get data self._create_get_data_script() self._create_process_script() self._create_clean_data_script() put_script_name = self._put_data_script_name() self.create_put_script(put_script_name, self.username, self.password, self.put_server, self.project, self.subject, self.session, self._working_directory_name, self._output_resource_name, self.PIPELINE_NAME) # run the script to get the data if processing_stage >= ProcessingStage.GET_DATA: stdout_file = open(self._get_data_script_name() + '.stdout', 'w') stderr_file = open(self._get_data_script_name() + '.stderr', 'w') logger.info("Running get data script") logger.info(" stdout: " + stdout_file.name) logger.info(" stderr: " + stderr_file.name) proc = subprocess.Popen( ['bash', self._get_data_script_name()], stdout=stdout_file, stderr=stderr_file) proc.communicate() logger.info(" return code: " + str(proc.returncode)) stdout_file.close() stderr_file.close() if proc.returncode != 0: raise RuntimeError( "get data script ended with non-zero return code") else: logger.info("Get data script not run") # run the script to submit processing jobs if processing_stage >= ProcessingStage.PROCESS_DATA: stdout_file = open(self._process_script_name() + '.stdout', 'w') stderr_file = open(self._process_script_name() + '.stderr', 'w') logger.info("Running script to submit processing jobs") logger.info(" stdout: " + stdout_file.name) logger.info(" stderr: " + stderr_file.name) proc = subprocess.Popen( ['bash', self._process_script_name()], stdout=stdout_file, stderr=stderr_file) proc.communicate() stdout_file.close() stderr_file.close() logger.info(" return code: " + str(proc.returncode)) if proc.returncode != 0: raise RuntimeError( "script to submit processing jobs ended with non-zero return code" ) else: logger.info("process data job not submitted") # submit the job to clean the data if processing_stage >= ProcessingStage.CLEAN_DATA: # figure out what the job id number is for the bedpostx postprocessing job postproc_file_name = self._working_directory_name + os.sep postproc_file_name += self.subject + os.sep postproc_file_name += 'T1w' + os.sep postproc_file_name += 'Diffusion.bedpostX' + os.sep postproc_file_name += 'logs' + os.sep postproc_file_name += 'postproc_ID' logger.info("Post-processing job ID file name: " + postproc_file_name) f = open(postproc_file_name, 'r') id_str = f.readline().rstrip() logger.info("Post-processing job ID: " + id_str) f.close() clean_submit_cmd = 'qsub -W depend=afterok:' + id_str + ' ' + self._clean_data_script_name( ) logger.info("clean_submit_cmd: " + clean_submit_cmd) completed_clean_submit_process = subprocess.run( clean_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) clean_job_no = str_utils.remove_ending_new_lines( completed_clean_submit_process.stdout) logger.info("clean_job_no: " + clean_job_no) else: logger.info("Clean data job not submitted") # submit the job to put the resulting data in the DB if processing_stage >= ProcessingStage.PUT_DATA: put_submit_cmd = 'qsub -W depend=afterok:' + clean_job_no + ' ' + put_script_name logger.info("put_submit_cmd: " + put_submit_cmd) completed_put_submit_process = subprocess.run( put_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) put_job_no = str_utils.remove_ending_new_lines( completed_put_submit_process.stdout) logger.info("put_job_no: " + put_job_no) else: logger.info("Put data job not submitted")
def submit_jobs(self, packaging_stage=PackagingStage.CREATE_PACKAGE): logger.debug("submit_jobs: packaging_stage: " + str(packaging_stage)) logger.info("----------") logger.info("Submitting " + self.WORK_DESC + " jobs for") logger.info(" Project: " + self.project) logger.info(" Subject: " + self.subject) logger.info(" Session: " + self.session) logger.info(" Stage: " + str(packaging_stage)) logger.info("----------") # make sure working directories do not have the same name based on # the same starting time by sleeping a few seconds time.sleep(5) # build the working directory logger.info("Making working directory: " + self.working_directory_name) os.makedirs(name=self.working_directory_name) if packaging_stage >= PackagingStage.PREPARE_SCRIPTS: self._write_get_data_script() self._write_create_package_script() self._write_clean_build_space_script() # submit the script to get the data if packaging_stage >= PackagingStage.GET_DATA: get_data_submit_cmd = 'qsub ' + self.get_data_script_name logger.info("get_data_submit_cmd: " + get_data_submit_cmd) completed_process = subprocess.run(get_data_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) get_data_job_no = str_utils.remove_ending_new_lines( completed_process.stdout) logger.info("get_data_job_no: " + get_data_job_no) else: logger.info("Get data job not submitted") # submit the script to create the package if packaging_stage >= PackagingStage.CREATE_PACKAGE: create_package_submit_cmd = 'qsub -W depend=afterok:' + get_data_job_no + ' ' create_package_submit_cmd += self.create_package_script_name logger.info("create_package_submit_cmd: " + create_package_submit_cmd) completed_process = subprocess.run(create_package_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) create_package_job_no = str_utils.remove_ending_new_lines( completed_process.stdout) logger.info("create_package_job_no: " + create_package_job_no) else: logger.info("Create package job not submitted") # submit the script to clean the build space if packaging_stage >= PackagingStage.CLEAN_BUILD_SPACE: clean_build_space_submit_cmd = 'qsub -W depend=afterok:' + create_package_job_no + ' ' clean_build_space_submit_cmd += self.clean_build_space_script_name logger.info("clean_build_space_submit_cmd: " + clean_build_space_submit_cmd) completed_process = subprocess.run(clean_build_space_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) clean_build_space_job_no = str_utils.remove_ending_new_lines( completed_process.stdout) logger.info("clean_build_space_job_no: " + clean_build_space_job_no) else: logger.info("Clean build space job not submitted")
def submit_jobs(self): _debug("submit_jobs") if self.validate_parameters(): _inform("") _inform("--------------------------------------------------") _inform("Submitting " + self.PIPELINE_NAME + " jobs for") _inform(" Project: " + self.project) _inform(" Subject: " + self.subject) _inform(" Session: " + self.session) _inform("--------------------------------------------------") # make sure working directories don't have the same name based on # the same start time by sleeping a few seconds time.sleep(5) current_seconds_since_epoch = int(time.time()) # build the working directory name self._working_directory_name = self.build_home self._working_directory_name += os.sep + self.project self._working_directory_name += os.sep + self.PIPELINE_NAME self._working_directory_name += '.' + self.subject self._working_directory_name += '.' + str(current_seconds_since_epoch) # make the working directory _inform("making working directory: " + self._working_directory_name) os.makedirs(name=self._working_directory_name) # get JSESSION ID jsession_id = xnat_access.get_jsession_id( server='db.humanconnectome.org', username=self.username, password=self.password) _inform("jsession_id: " + jsession_id) # get XNAT Session ID (a.k.a. the experiment ID, e.g. ConnectomeDB_E1234) xnat_session_id = xnat_access.get_session_id( server='db.humanconnectome.org', username=self.username, password=self.password, project=kself.project, subject=self.subject, session=self.session) _inform("xnat_session_id: " + xnat_session_id) # get XNAT Workflow ID workflow_obj = xnat_access.Workflow(self.username, self.password, 'https://db.humanconnectome.org', jsession_id) self._workflow_id = workflow_obj.create_workflow(xnat_session_id, self.project, self.PIPELINE_NAME, 'Queued') _inform("workflow_id: " + self._workflow_id) # determine output resource name self._output_resource_name = 'Diffusion_preproc' # clean the output resource if requested if self.clean_output_resource_first: _inform("Deleting resource: " + self._output_resource_name + " for:") _inform(" project: " + self.project) _inform(" subject: " + self.subject) _inform(" session: " + self.session) delete_resource.delete_resource( self.username, self.password, str_utils.get_server_name(self.server), self.project, self.subject, self.session, self._output_resource_name) # create script to do the PreEddy work self._create_pre_eddy_script() # create script to do the Eddy work self._create_eddy_script() # create script to do the PostEddy work self._create_post_eddy_script() # create script to put the results into the DB put_script_name = self._get_scripts_start_name() + '.XNAT_PBS_PUT_job.sh' self.create_put_script(put_script_name, self.username, self.password, self.put_server, self.project, self.subject, self.session, self._working_directory_name, self._output_resource_name, self.PIPELINE_NAME) # Submit the job to do the Pre-Eddy work pre_eddy_submit_cmd = 'qsub ' + self._pre_eddy_script_name _inform("pre_eddy_submit_cmd: " + pre_eddy_submit_cmd) completed_pre_eddy_submit_process = subprocess.run( pre_eddy_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) pre_eddy_job_no = str_utils.remove_ending_new_lines(completed_pre_eddy_submit_process.stdout) _inform("pre_eddy_job_no: " + pre_eddy_job_no) # Submit the job to do the Eddy work eddy_submit_cmd = 'qsub -W depend=afterok:' + pre_eddy_job_no + ' ' + self._eddy_script_name _inform("eddy_submit_cmd: " + eddy_submit_cmd) completed_eddy_submit_process = subprocess.run( eddy_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) eddy_job_no = str_utils.remove_ending_new_lines(completed_eddy_submit_process.stdout) _inform("eddy_job_no: " + eddy_job_no) # Submit the job to do the Post-Eddy work post_eddy_submit_cmd = 'qsub -W depend=afterok:' + eddy_job_no + ' ' + self._post_eddy_script_name _inform("post_eddy_submit_cmd: " + post_eddy_submit_cmd) completed_post_eddy_submit_process = subprocess.run( post_eddy_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) post_eddy_job_no = str_utils.remove_ending_new_lines(completed_post_eddy_submit_process.stdout) _inform("post_eddy_job_no: " + post_eddy_job_no) # Submit the job to put the results in the DB put_submit_cmd = 'qsub -W depend=afterok:' + post_eddy_job_no + ' ' + put_script_name _inform("put_submit_cmd: " + put_submit_cmd) completed_put_submit_process = subprocess.run( put_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) put_job_no = str_utils.remove_ending_new_lines(completed_put_submit_process.stdout) _inform("put_job_no: " + put_job_no) else: _inform("Unable to submit jobs")
def submit_jobs( self, username, password, server, project, subject, session, structural_reference_project, structural_reference_session, put_server, clean_output_resource_first, setup_script, incomplete_only, scan, walltime_limit_hours, mem_limit_gbs, # UNUSED vmem_limit_gbs): subject_info = hcp7t_subject.Hcp7TSubjectInfo( project, structural_reference_project, subject) # determine names of preprocessed resting state scans that are # available for the subject preproc_resting_state_scan_names = self.archive.available_resting_state_preproc_names( subject_info) inform("Preprocessed resting state scans available for subject: " + str(preproc_resting_state_scan_names)) # determine names of the preprocessed MOVIE task scans that are available for the subject preproc_movie_scan_names = self.archive.available_movie_preproc_names( subject_info) inform("Preprocessed movie scans available for subject " + str(preproc_movie_scan_names)) # determine names of the FIX processed scans that are available for the subject fix_processed_scan_names = self.archive.available_FIX_processed_names( subject_info) inform("FIX processed scans available for subject " + str(fix_processed_scan_names)) # build list of scans to process scan_list = [] if scan is None: scan_list = fix_processed_scan_names else: scan_list.append(scan) # process specified scans for scan_name in scan_list: if incomplete_only: completion_checker = PostFixHCP7T_OneSubjectCompletionChecker.PostFixHCP7T_OneSubjectCompletionChecker( ) if completion_checker.is_processing_complete( self.archive, subject_info, scan_name): inform("scan: " + scan_name + " has already completed PostFixHCP7T processing") inform( "Only submitting jobs for incomplete scans - skipping " + scan_name) continue inform("scan_name: " + scan_name) long_scan_name = self.archive.functional_scan_long_name(scan_name) output_resource_name = self.archive.PostFix_processed_resource_name( scan_name) inform("") inform("-------------------------------------------------") inform("Submitting jobs for scan: " + long_scan_name) inform("Output resource name: " + output_resource_name) inform("-------------------------------------------------") inform("") # make sure working directories don't have the same name based on the # same start time by sleeping a few seconds time.sleep(5) current_seconds_since_epoch = int(time.time()) working_directory_name = self.build_home working_directory_name += os.sep + project working_directory_name += os.sep + self.PIPELINE_NAME working_directory_name += '.' + subject working_directory_name += '.' + long_scan_name working_directory_name += '.' + str(current_seconds_since_epoch) # make the working directory inform("Making working directory: " + working_directory_name) os.makedirs(name=working_directory_name) # get JSESSION ID jsession_id = xnat_access.get_jsession_id( server=os_utils.getenv_required('XNAT_PBS_JOBS_XNAT_SERVER'), username=username, password=password) inform("jsession_id: " + jsession_id) # get XNAT Session ID (a.k.a. the experiment ID, e.g. ConnectomeDB_E1234) xnat_session_id = xnat_access.get_session_id( server=os_utils.getenv_required('XNAT_PBS_JOBS_XNAT_SERVER'), username=username, password=password, project=project, subject=subject, session=session) inform("xnat_session_id: " + xnat_session_id) # get XNAT Workflow ID workflow_obj = xnat_access.Workflow(username, password, server, jsession_id) workflow_id = workflow_obj.create_workflow(xnat_session_id, project, self.PIPELINE_NAME, 'Queued') inform("workflow_id: " + workflow_id) # Clean the output resource if requested if clean_output_resource_first: inform("Deleting resource: " + output_resource_name + " for:") inform(" project: " + project) inform(" subject: " + subject) inform(" session: " + session) delete_resource.delete_resource( username, password, str_utils.get_server_name(server), project, subject, session, output_resource_name) script_file_start_name = working_directory_name script_file_start_name += os.sep + subject script_file_start_name += '.' + long_scan_name script_file_start_name += '.' + self.PIPELINE_NAME script_file_start_name += '.' + project script_file_start_name += '.' + session # Create script to submit to do the actual work work_script_name = script_file_start_name + '.XNAT_PBS_job.sh' with contextlib.suppress(FileNotFoundError): os.remove(work_script_name) work_script = open(work_script_name, 'w') nodes_spec = 'nodes=1:ppn=1' walltime_spec = 'walltime=' + str(walltime_limit_hours) + ':00:00' vmem_spec = 'vmem=' + str(vmem_limit_gbs) + 'gb' work_script.write('#PBS -l ' + nodes_spec + ',' + walltime_spec + ',' + vmem_spec + os.linesep) work_script.write('#PBS -o ' + working_directory_name + os.linesep) work_script.write('#PBS -e ' + working_directory_name + os.linesep) work_script.write(os.linesep) work_script.write(self.xnat_pbs_jobs_home + os.sep + '7T' + os.sep + 'PostFixHCP7T' + os.sep + 'PostFixHCP7T.XNAT.sh \\' + os.linesep) work_script.write(' --user="******" \\' + os.linesep) work_script.write(' --password="******" \\' + os.linesep) work_script.write(' --server="' + str_utils.get_server_name(server) + '" \\' + os.linesep) work_script.write(' --project="' + project + '" \\' + os.linesep) work_script.write(' --subject="' + subject + '" \\' + os.linesep) work_script.write(' --session="' + session + '" \\' + os.linesep) work_script.write(' --scan="' + long_scan_name + '" \\' + os.linesep) work_script.write(' --working-dir="' + working_directory_name + '" \\' + os.linesep) work_script.write(' --workflow-id="' + workflow_id + '" \\' + os.linesep) work_script.write(' --setup-script=' + setup_script + os.linesep) work_script.close() os.chmod(work_script_name, stat.S_IRWXU | stat.S_IRWXG) # Create script to put the results into the DB put_script_name = script_file_start_name + '.XNAT_PBS_PUT_job.sh' self.create_put_script(put_script_name, username, password, put_server, project, subject, session, working_directory_name, output_resource_name, scan_name + '_' + self.PIPELINE_NAME) # Submit the job to do the work work_submit_cmd = 'qsub ' + work_script_name inform("work_submit_cmd: " + work_submit_cmd) completed_work_submit_process = subprocess.run( work_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) work_job_no = str_utils.remove_ending_new_lines( completed_work_submit_process.stdout) inform("work_job_no: " + work_job_no) # Submit the job to put the results in the DB put_submit_cmd = 'qsub -W depend=afterok:' + work_job_no + ' ' + put_script_name inform("put_submit_cmd: " + put_submit_cmd) completed_put_submit_process = subprocess.run( put_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) put_job_no = str_utils.remove_ending_new_lines( completed_put_submit_process.stdout) inform("put_job_no: " + put_job_no)
def submit_jobs(self, processing_stage=ProcessingStage.PUT_DATA): logger.debug("submit_jobs: processing_stage: " + str(processing_stage)) logger.info("----------") logger.info("Submitting " + self.PIPELINE_NAME + " jobs for") logger.info(" Project: " + self.project) logger.info(" Subject: " + self.subject) logger.info(" Session: " + self.session) logger.info(" Stage: " + str(processing_stage)) logger.info("----------") # make sure working directories do not have the same name based on # the same start time by sleeping a few seconds time.sleep(5) # build the working directory name self._working_directory_name = \ self.build_working_directory_name(self.project, self.PIPELINE_NAME, self.subject) logger.info("Making working directory: " + self._working_directory_name) os.makedirs(name=self._working_directory_name) # # get JSESSION ID # jsession_id = xnat_access.get_jsession_id( # server='db.humanconnectome.org', # username=self.username, # password=self.password) # logger.info("jsession_id: " + jsession_id) # # get XNAT Session ID (a.k.a. the experiment ID, e.g. ConnectomeDB_E1234) # xnat_session_id = xnat_access.get_session_id( # server='db.humanconnectome.org', # username=self.username, # password=self.password, # project=self.project, # subject=self.subject, # session=self.session) # logger.info("xnat_session_id: " + xnat_session_id) # # get XNAT Workflow ID # workflow_obj = xnat_access.Workflow(self.username, self.password, # 'https://db.humanconnectome.org', jsession_id) # self._workflow_id = workflow_obj.create_workflow(xnat_session_id, # self.project, # self.PIPELINE_NAME, # 'Queued') # logger.info("workflow_id: " + self._workflow_id) # determine output resource name self._output_resource_name = "Diffusion_bedpostx" # clean output resource if requested if self.clean_output_resource_first: logger.info("Deleting resource: " + self._output_resource_name + " for:") logger.info(" project: " + self.project) logger.info(" subject: " + self.subject) logger.info(" session: " + self.session) delete_resource.delete_resource( self.username, self.password, str_utils.get_server_name(self.server), self.project, self.subject, self.session, self._output_resource_name) # create scripts for various stages of processing if processing_stage >= ProcessingStage.PREPARE_SCRIPTS: # create script to get data self._create_get_data_script() self._create_process_script() self._create_clean_data_script() put_script_name = self._put_data_script_name() self.create_put_script(put_script_name, self.username, self.password, self.put_server, self.project, self.subject, self.session, self._working_directory_name, self._output_resource_name, self.PIPELINE_NAME) # run the script to get the data if processing_stage >= ProcessingStage.GET_DATA: stdout_file = open(self._get_data_script_name() + '.stdout', 'w') stderr_file = open(self._get_data_script_name() + '.stderr', 'w') logger.info("Running get data script") logger.info(" stdout: " + stdout_file.name) logger.info(" stderr: " + stderr_file.name) proc = subprocess.Popen(['bash', self._get_data_script_name()], stdout=stdout_file, stderr=stderr_file) proc.communicate() logger.info(" return code: " + str(proc.returncode)) stdout_file.close() stderr_file.close() if proc.returncode != 0: raise RuntimeError("get data script ended with non-zero return code") else: logger.info("Get data script not run") # run the script to submit processing jobs if processing_stage >= ProcessingStage.PROCESS_DATA: stdout_file = open(self._process_script_name() + '.stdout', 'w') stderr_file = open(self._process_script_name() + '.stderr', 'w') logger.info("Running script to submit processing jobs") logger.info(" stdout: " + stdout_file.name) logger.info(" stderr: " + stderr_file.name) proc = subprocess.Popen(['bash', self._process_script_name()], stdout=stdout_file, stderr=stderr_file) proc.communicate() stdout_file.close() stderr_file.close() logger.info(" return code: " + str(proc.returncode)) if proc.returncode != 0: raise RuntimeError("script to submit processing jobs ended with non-zero return code") else: logger.info("process data job not submitted") # submit the job to clean the data if processing_stage >= ProcessingStage.CLEAN_DATA: # figure out what the job id number is for the bedpostx postprocessing job postproc_file_name = self._working_directory_name + os.sep postproc_file_name += self.subject + os.sep postproc_file_name += 'T1w' + os.sep postproc_file_name += 'Diffusion.bedpostX' + os.sep postproc_file_name += 'logs' + os.sep postproc_file_name += 'postproc_ID' logger.info("Post-processing job ID file name: " + postproc_file_name) f = open(postproc_file_name, 'r') id_str = f.readline().rstrip() logger.info("Post-processing job ID: " + id_str) f.close() clean_submit_cmd = 'qsub -W depend=afterok:' + id_str + ' ' + self._clean_data_script_name() logger.info("clean_submit_cmd: " + clean_submit_cmd) completed_clean_submit_process = subprocess.run( clean_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) clean_job_no = str_utils.remove_ending_new_lines(completed_clean_submit_process.stdout) logger.info("clean_job_no: " + clean_job_no) else: logger.info("Clean data job not submitted") # submit the job to put the resulting data in the DB if processing_stage >= ProcessingStage.PUT_DATA: put_submit_cmd = 'qsub -W depend=afterok:' + clean_job_no + ' ' + put_script_name logger.info("put_submit_cmd: " + put_submit_cmd) completed_put_submit_process = subprocess.run( put_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) put_job_no = str_utils.remove_ending_new_lines(completed_put_submit_process.stdout) logger.info("put_job_no: " + put_job_no) else: logger.info("Put data job not submitted")
def submit_jobs(self, processing_stage=ProcessingStage.PUT_DATA): logger.debug(debug_utils.get_name() + ": processing_stage: " + str(processing_stage)) logger.info("-----") logger.info("Submitting " + self.PIPELINE_NAME + " jobs for") logger.info(" Project: " + self.project) logger.info(" Subject: " + self.subject) logger.info(" Session: " + self.session) logger.info(" Scan: " + self.scan) logger.info(" Stage: " + str(processing_stage)) # make sure working directories do not have the same name based on # the same start time by sleeping a few seconds time.sleep(5) # build the working directory name self._working_directory_name = \ self.build_working_directory_name(self.project, self.PIPELINE_NAME, self.subject, self.scan) logger.info("Making working directory: " + self._working_directory_name) os.makedirs(name=self._working_directory_name) # determine output resource name self._output_resource_name = self.scan + "_" + self.output_resource_suffix # clean output resource if requested if self.clean_output_resource_first: logger.info("Deleting resource: " + self._output_resource_name + " for:") logger.info(" project: " + self.project) logger.info(" subject: " + self.subject) logger.info(" session: " + self.session) delete_resource.delete_resource( self.username, self.password, str_utils.get_server_name(self.server), self.project, self.subject, self.session, self._output_resource_name) # create scripts for various stages of processing if processing_stage >= ProcessingStage.PREPARE_SCRIPTS: self._create_get_data_script() self._create_work_script() self._create_clean_data_script() put_script_name = self._put_data_script_name() self.create_put_script(put_script_name, self.username, self.password, self.put_server, self.project, self.subject, self.session, self._working_directory_name, self._output_resource_name, self.PIPELINE_NAME, leave_subject_id_level=True) # Submit the job to get the data if processing_stage >= ProcessingStage.GET_DATA: get_data_submit_cmd = 'qsub ' + self._get_data_script_name() logger.info("get_data_submit_cmd: " + get_data_submit_cmd) completed_get_data_submit_process = subprocess.run( get_data_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) get_data_job_no = str_utils.remove_ending_new_lines( completed_get_data_submit_process.stdout) logger.info("get_data_job_no: " + get_data_job_no) else: logger.info("Get data job not submitted") # Submit the job to process the data (do the work) if processing_stage >= ProcessingStage.PROCESS_DATA: work_submit_cmd = 'qsub -W depend=afterok:' + get_data_job_no + ' ' + self._work_script_name( ) logger.info("work_submit_cmd: " + work_submit_cmd) completed_work_submit_process = subprocess.run( work_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) work_job_no = str_utils.remove_ending_new_lines( completed_work_submit_process.stdout) logger.info("work_job_no: " + work_job_no) else: logger.info("Process data job not submitted") # Submit job to clean the data if processing_stage >= ProcessingStage.CLEAN_DATA: clean_submit_cmd = 'qsub -W depend=afterok:' + work_job_no + ' ' + self._clean_data_script_name( ) logger.info("clean_submit_cmd: " + clean_submit_cmd) completed_clean_submit_process = subprocess.run( clean_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) clean_job_no = str_utils.remove_ending_new_lines( completed_clean_submit_process.stdout) logger.info("clean_job_no: " + clean_job_no) else: logger.info("Clean data job not submitted") # Submit job to put the resulting data in the DB if processing_stage >= ProcessingStage.PUT_DATA: put_submit_cmd = 'qsub -W depend=afterok:' + clean_job_no + ' ' + put_script_name logger.info("put_submit_cmd: " + put_submit_cmd) completed_put_submit_process = subprocess.run( put_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) put_job_no = str_utils.remove_ending_new_lines( completed_put_submit_process.stdout) logger.info("put_job_no: " + put_job_no) else: logger.info("Put data job not submitted")
def submit_jobs(self, username, password, server, project, subject, session, structural_reference_project, structural_reference_session, put_server, setup_script, incomplete_only, scan, walltime_limit_hours, mem_limit_gbs, vmem_limit_gbs): subject_info = hcp7t_subject.Hcp7TSubjectInfo( project, structural_reference_project, subject) # determine names of the preprocessed resting state scans that are # available for the subject resting_state_scan_names = self.archive.available_resting_state_preproc_names( subject_info) inform("Preprocessed resting state scans available for subject: " + str(resting_state_scan_names)) # determine names of the preprocessed MOVIE task scans that are available for the subject movie_scan_names = self.archive.available_movie_preproc_names( subject_info) inform("Preprocessed movie scans available for subject " + str(movie_scan_names)) # build list of scans to process scan_list = [] if scan is None: scan_list = resting_state_scan_names + movie_scan_names else: scan_list.append(scan) # process the specified scans for scan_name in scan_list: if incomplete_only and self.archive.FIX_processing_repaired( subject_info, scan_name): inform("scan: " + scan_name + " FIX processing is already repaired") inform( "Only submitting jobs for incomplete scans - skipping " + scan_name) continue long_scan_name = self.archive.functional_scan_long_name(scan_name) output_resource_name = self.archive.FIX_processed_resource_name( scan_name) inform("") inform("-------------------------------------------------") inform("Submitting jobs for scan: " + long_scan_name) inform("Output resource name: " + output_resource_name) inform("") # make sure working directories don't have the same name based on the # same start time by sleeping a few seconds time.sleep(5) current_seconds_since_epoch = int(time.time()) working_directory_name = self.build_home working_directory_name += os.sep + project working_directory_name += os.sep + self.PIPELINE_NAME working_directory_name += '.' + subject working_directory_name += '.' + long_scan_name working_directory_name += '.' + str(current_seconds_since_epoch) inform("Making working directory: " + working_directory_name) os.makedirs(name=working_directory_name) script_file_start_name = working_directory_name script_file_start_name += os.sep + subject script_file_start_name += '.' + long_scan_name script_file_start_name += '.' + self.PIPELINE_NAME script_file_start_name += '.' + project script_file_start_name += '.' + session # Create script to submit to do the actual work work_script_name = script_file_start_name + '.XNAT_PBS_job.sh' with contextlib.suppress(FileNotFoundError): os.remove(work_script_name) work_script = open(work_script_name, 'w') work_script.write('#PBS -l nodes=1:ppn=1,walltime=' + str(walltime_limit_hours) + ':00:00,mem=' + str(mem_limit_gbs) + 'gb,vmem=' + str(vmem_limit_gbs) + 'gb' + os.linesep) work_script.write('#PBS -o ' + working_directory_name + os.linesep) work_script.write('#PBS -e ' + working_directory_name + os.linesep) work_script.write(os.linesep) work_script.write(self.xnat_pbs_jobs_home + os.sep + '7T' + os.sep + self.PIPELINE_NAME + os.sep + self.PIPELINE_NAME + '.XNAT.sh \\' + os.linesep) work_script.write(' --user="******" \\' + os.linesep) work_script.write(' --password="******" \\' + os.linesep) work_script.write(' --server="' + str_utils.get_server_name(server) + '" \\' + os.linesep) work_script.write(' --project="' + project + '" \\' + os.linesep) work_script.write(' --subject="' + subject + '" \\' + os.linesep) work_script.write(' --session="' + session + '" \\' + os.linesep) work_script.write(' --structural-reference-project="' + structural_reference_project + '" \\' + os.linesep) work_script.write(' --structural-reference-session="' + structural_reference_session + '" \\' + os.linesep) work_script.write(' --scan="' + long_scan_name + '" \\' + os.linesep) work_script.write(' --working-dir="' + working_directory_name + '" \\' + os.linesep) work_script.write(' --setup-script=' + setup_script + os.linesep) work_script.close() os.chmod(work_script_name, stat.S_IRWXU | stat.S_IRWXG) # Create script to put the results into the DB put_script_name = script_file_start_name + '.XNAT_PBS_PUT_job.sh' self.create_put_script(put_script_name, username, password, put_server, project, subject, session, working_directory_name, output_resource_name, scan_name + '_' + self.PIPELINE_NAME, leave_subject_id_level=True) # Submit the job to do the work work_submit_cmd = 'qsub ' + work_script_name inform("work_submit_cmd: " + work_submit_cmd) completed_work_submit_process = subprocess.run( work_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) work_job_no = str_utils.remove_ending_new_lines( completed_work_submit_process.stdout) inform("work_job_no: " + work_job_no) # Submit the job put the results in the DB put_submit_cmd = 'qsub -W depend=afterok:' + work_job_no + ' ' + put_script_name inform("put_submit_cmd: " + put_submit_cmd) completed_put_submit_process = subprocess.run( put_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) put_job_no = str_utils.remove_ending_new_lines( completed_put_submit_process.stdout) inform("put_job_no: " + put_job_no)
def submit_jobs( self, username, password, server, project, subject, session, structural_reference_project, structural_reference_session, put_server, clean_output_resource_first, setup_script, incomplete_only, scan, walltime_limit_hours, mem_limit_gbs, # UNUSED vmem_limit_gbs, ): subject_info = hcp7t_subject.Hcp7TSubjectInfo(project, structural_reference_project, subject) # determine names of preprocessed resting state scans that are # available for the subject preproc_resting_state_scan_names = self.archive.available_resting_state_preproc_names(subject_info) inform("Preprocessed resting state scans available for subject: " + str(preproc_resting_state_scan_names)) # determine names of the preprocessed MOVIE task scans that are available for the subject preproc_movie_scan_names = self.archive.available_movie_preproc_names(subject_info) inform("Preprocessed movie scans available for subject " + str(preproc_movie_scan_names)) # determine names of the FIX processed scans that are available for the subject fix_processed_scan_names = self.archive.available_FIX_processed_names(subject_info) inform("FIX processed scans available for subject " + str(fix_processed_scan_names)) # build list of scans to process scan_list = [] if scan is None: scan_list = fix_processed_scan_names else: scan_list.append(scan) # process specified scans for scan_name in scan_list: if incomplete_only: completion_checker = PostFixHCP7T_OneSubjectCompletionChecker.PostFixHCP7T_OneSubjectCompletionChecker() if completion_checker.is_processing_complete(self.archive, subject_info, scan_name): inform("scan: " + scan_name + " has already completed PostFixHCP7T processing") inform("Only submitting jobs for incomplete scans - skipping " + scan_name) continue long_scan_name = self.archive.functional_scan_long_name(scan_name) output_resource_name = self.archive.PostFix_processed_resource_name(scan_name) inform("") inform("-------------------------------------------------") inform("Submitting jobs for scan: " + long_scan_name) inform("Output resource name: " + output_resource_name) inform("-------------------------------------------------") inform("") # make sure working directories don't have the same name based on the # same start time by sleeping a few seconds time.sleep(5) current_seconds_since_epoch = int(time.time()) working_directory_name = self.build_home working_directory_name += os.sep + project working_directory_name += os.sep + self.PIPELINE_NAME working_directory_name += "." + subject working_directory_name += "." + long_scan_name working_directory_name += "." + str(current_seconds_since_epoch) # make the working directory inform("Making working directory: " + working_directory_name) os.makedirs(name=working_directory_name) # get JSESSION ID jsession_id = xnat_access.get_jsession_id( server="db.humanconnectome.org", username=username, password=password ) inform("jsession_id: " + jsession_id) # get XNAT Session ID (a.k.a. the experiment ID, e.g. ConnectomeDB_E1234) xnat_session_id = xnat_access.get_session_id( server="db.humanconnectome.org", username=username, password=password, project=project, subject=subject, session=session, ) inform("xnat_session_id: " + xnat_session_id) # get XNAT Workflow ID workflow_obj = xnat_access.Workflow(username, password, server, jsession_id) workflow_id = workflow_obj.create_workflow(xnat_session_id, project, self.PIPELINE_NAME, "Queued") inform("workflow_id: " + workflow_id) # Clean the output resource if requested if clean_output_resource_first: inform("Deleting resource: " + output_resource_name + " for:") inform(" project: " + project) inform(" subject: " + subject) inform(" session: " + session) delete_resource.delete_resource( username, password, str_utils.get_server_name(server), project, subject, session, output_resource_name, ) script_file_start_name = working_directory_name script_file_start_name += os.sep + subject script_file_start_name += "." + long_scan_name script_file_start_name += "." + self.PIPELINE_NAME script_file_start_name += "." + project script_file_start_name += "." + session # Create script to submit to do the actual work work_script_name = script_file_start_name + ".XNAT_PBS_job.sh" with contextlib.suppress(FileNotFoundError): os.remove(work_script_name) work_script = open(work_script_name, "w") nodes_spec = "nodes=1:ppn=1" walltime_spec = "walltime=" + str(walltime_limit_hours) + ":00:00" vmem_spec = "vmem=" + str(vmem_limit_gbs) + "gb" work_script.write("#PBS -l " + nodes_spec + "," + walltime_spec + "," + vmem_spec + os.linesep) work_script.write("#PBS -o " + working_directory_name + os.linesep) work_script.write("#PBS -e " + working_directory_name + os.linesep) work_script.write(os.linesep) work_script.write( self.xnat_pbs_jobs_home + os.sep + "7T" + os.sep + "PostFixHCP7T" + os.sep + "PostFixHCP7T.XNAT.sh \\" + os.linesep ) work_script.write(' --user="******" \\' + os.linesep) work_script.write(' --password="******" \\' + os.linesep) work_script.write(' --server="' + str_utils.get_server_name(server) + '" \\' + os.linesep) work_script.write(' --project="' + project + '" \\' + os.linesep) work_script.write(' --subject="' + subject + '" \\' + os.linesep) work_script.write(' --session="' + session + '" \\' + os.linesep) work_script.write(' --scan="' + long_scan_name + '" \\' + os.linesep) work_script.write(' --working-dir="' + working_directory_name + '" \\' + os.linesep) work_script.write(' --workflow-id="' + workflow_id + '" \\' + os.linesep) work_script.write(" --setup-script=" + setup_script + os.linesep) work_script.close() os.chmod(work_script_name, stat.S_IRWXU | stat.S_IRWXG) # Create script to put the results into the DB put_script_name = script_file_start_name + ".XNAT_PBS_PUT_job.sh" self.create_put_script( put_script_name, username, password, put_server, project, subject, session, working_directory_name, output_resource_name, scan_name + "_" + self.PIPELINE_NAME, ) # Submit the job to do the work work_submit_cmd = "qsub " + work_script_name inform("work_submit_cmd: " + work_submit_cmd) completed_work_submit_process = subprocess.run( work_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True ) work_job_no = str_utils.remove_ending_new_lines(completed_work_submit_process.stdout) inform("work_job_no: " + work_job_no) # Submit the job to put the results in the DB put_submit_cmd = "qsub -W depend=afterok:" + work_job_no + " " + put_script_name inform("put_submit_cmd: " + put_submit_cmd) completed_put_submit_process = subprocess.run( put_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True ) put_job_no = str_utils.remove_ending_new_lines(completed_put_submit_process.stdout) inform("put_job_no: " + put_job_no)
def submit_jobs(self): _debug("submit_jobs") if self.validate_parameters(): # make sure working directories don't have the same name based on the same # start time by sleeping a few seconds time.sleep(5) current_seconds_since_epoch = int(time.time()) working_directory_name = self.build_home working_directory_name += os.sep + self.project working_directory_name += os.sep + self.PIPELINE_NAME working_directory_name += '.' + self.subject working_directory_name += '.' + str(current_seconds_since_epoch) # make the working directory _inform("Making working directory: " + working_directory_name) os.makedirs(name=working_directory_name) # get JSESSION ID jsession_id = xnat_access.get_jsession_id( server=os_utils.getenv_required('XNAT_PBS_JOBS_XNAT_SERVER'), username=self.username, password=self.password) _inform("jsession_id: " + jsession_id) # get XNAT Session ID (a.k.a. the experiment ID, e.g. ConnectomeDB_E1234) xnat_session_id = xnat_access.get_session_id( server=os_utils.getenv_required('XNAT_PBS_JOBS_XNAT_SERVER'), username=self.username, password=self.password, project=self.project, subject=self.subject, session=self.session) _inform("xnat_session_id: " + xnat_session_id) # get XNAT Workflow ID workflow_obj = xnat_access.Workflow( self.username, self.password, self.server, jsession_id) workflow_id = workflow_obj.create_workflow( xnat_session_id, self.project, self.PIPELINE_NAME, 'Queued') _inform("workflow_id: " + workflow_id) # Determine the output resource name output_resource_name = self.archive.DEDRIFT_AND_RESAMPLE_RESOURCE_NAME _inform("output_resource_name: " + output_resource_name) # Clean the output resource if requested if self.clean_output_resource_first: _inform("Deleting resouce: " + output_resource_name + " for:") _inform(" project: " + self.project) _inform(" subject: " + self.subject) _inform(" session: " + self.session) delete_resource.delete_resource( self.username, self.password, str_utils.get_server_name(self.server), self.project, self.subject, self.session, output_resource_name, True) script_file_start_name = working_directory_name script_file_start_name += os.sep + self.subject script_file_start_name += '.' + self.PIPELINE_NAME script_file_start_name += '.' + self.project script_file_start_name += '.' + self.session # Create script to submit to do the actual work work_script_name = script_file_start_name + '.XNAT_PBS_job.sh' with contextlib.suppress(FileNotFoundError): os.remove(work_script_name) work_script = open(work_script_name, 'w') nodes_spec = 'nodes=1:ppn=1' walltime_spec = 'walltime=' + str(self.walltime_limit_hours) + ':00:00' vmem_spec = 'vmem=' + str(self.vmem_limit_gbs) + 'gb' mem_spec = 'mem=' + str(self.mem_limit_gbs) + 'gb' work_script.write('#PBS -l ' + nodes_spec + ',' + walltime_spec + ',' + vmem_spec + ',' + mem_spec + os.linesep) # work_script.write('#PBS -q HCPput' + os.linesep) work_script.write('#PBS -o ' + working_directory_name + os.linesep) work_script.write('#PBS -e ' + working_directory_name + os.linesep) work_script.write(os.linesep) work_script.write(self.xnat_pbs_jobs_home + os.sep + '7T' + os.sep + 'DeDriftAndResampleHCP7T' + os.sep + 'DeDriftAndResampleHCP7T.XNAT.sh \\' + os.linesep) work_script.write(' --user="******" \\' + os.linesep) work_script.write(' --password="******" \\' + os.linesep) work_script.write(' --server="' + str_utils.get_server_name(self.server) + '" \\' + os.linesep) work_script.write(' --project="' + self.project + '" \\' + os.linesep) work_script.write(' --subject="' + self.subject + '" \\' + os.linesep) work_script.write(' --session="' + self.session + '" \\' + os.linesep) work_script.write(' --structural-reference-project="' + self.structural_reference_project + '" \\' + os.linesep) work_script.write(' --structural-reference-session="' + self.structural_reference_session + '" \\' + os.linesep) work_script.write(' --working-dir="' + working_directory_name + '" \\' + os.linesep) work_script.write(' --workflow-id="' + workflow_id + '" \\' + os.linesep) # work_script.write(' --keep-all' + ' \\' + os.linesep) # work_script.write(' --prevent-push' + ' \\' + os.linesep) work_script.write(' --setup-script=' + self.setup_script + os.linesep) work_script.close() os.chmod(work_script_name, stat.S_IRWXU | stat.S_IRWXG) # Create script to put the results into the DB put_script_name = script_file_start_name + '.XNAT_PBS_PUT_job.sh' self.create_put_script(put_script_name, self.username, self.password, self.put_server, self.project, self.subject, self.session, working_directory_name, output_resource_name, self.PIPELINE_NAME) # Submit the job to do the work work_submit_cmd = 'qsub ' + work_script_name _inform("work_submit_cmd: " + work_submit_cmd) completed_work_submit_process = subprocess.run( work_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) work_job_no = str_utils.remove_ending_new_lines(completed_work_submit_process.stdout) _inform("work_job_no: " + work_job_no) # Submit the job to put the results in the DB put_submit_cmd = 'qsub -W depend=afterok:' + work_job_no + ' ' + put_script_name _inform("put_submit_cmd: " + put_submit_cmd) completed_put_submit_process = subprocess.run( put_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) put_job_no = str_utils.remove_ending_new_lines(completed_put_submit_process.stdout) _inform("put_job_no: " + put_job_no) else: _inform("Unable to submit jobs")
def submit_jobs(self): _debug("submit_jobs") if self.validate_parameters(): # subject_info = hcp7t_subject.Hcp7TSubjectInfo(self.project, # self.structural_reference_project, # self.subject) # make sure working directories don't have the same name based on the same # start time by sleeping a few seconds time.sleep(5) current_seconds_since_epoch = int(time.time()) working_directory_name = self.build_home working_directory_name += os.sep + self.project working_directory_name += os.sep + self.PIPELINE_NAME working_directory_name += '.' + self.subject working_directory_name += '.' + str(current_seconds_since_epoch) # make the working directory _inform("Making working directory: " + working_directory_name) os.makedirs(name=working_directory_name) # get JSESSION ID jsession_id = xnat_access.get_jsession_id( server='db.humanconnectome.org', username=self.username, password=self.password) _inform("jsession_id: " + jsession_id) # get XNAT Session ID (a.k.a. the experiment ID, e.g. ConnectomeDB_E1234) xnat_session_id = xnat_access.get_session_id( server='db.humanconnectome.org', username=self.username, password=self.password, project=self.project, subject=self.subject, session=self.session) _inform("xnat_session_id: " + xnat_session_id) # get XNAT Workflow ID workflow_obj = xnat_access.Workflow( self.username, self.password, self.server, jsession_id) workflow_id = workflow_obj.create_workflow( xnat_session_id, self.project, self.PIPELINE_NAME, 'Queued') _inform("workflow_id: " + workflow_id) # Determine the output resource name output_resource_name = self.archive.DEDRIFT_AND_RESAMPLE_RESOURCE_NAME _inform("output_resource_name: " + output_resource_name) # Clean the output resource if requested if self.clean_output_resource_first: _inform("Deleting resouce: " + output_resource_name + " for:") _inform(" project: " + self.project) _inform(" subject: " + self.subject) _inform(" session: " + self.session) delete_resource.delete_resource( self.username, self.password, str_utils.get_server_name(self.server), self.project, self.subject, self.session, output_resource_name, True) script_file_start_name = working_directory_name script_file_start_name += os.sep + self.subject script_file_start_name += '.' + self.PIPELINE_NAME script_file_start_name += '.' + self.project script_file_start_name += '.' + self.session # Create script to submit to do the actual work work_script_name = script_file_start_name + '.XNAT_PBS_job.sh' with contextlib.suppress(FileNotFoundError): os.remove(work_script_name) work_script = open(work_script_name, 'w') nodes_spec = 'nodes=1:ppn=1' walltime_spec = 'walltime=' + str(self.walltime_limit_hours) + ':00:00' vmem_spec = 'vmem=' + str(self.vmem_limit_gbs) + 'gb' work_script.write('#PBS -l ' + nodes_spec + ',' + walltime_spec + ',' + vmem_spec + os.linesep) # work_script.write('#PBS -q HCPput' + os.linesep) work_script.write('#PBS -o ' + working_directory_name + os.linesep) work_script.write('#PBS -e ' + working_directory_name + os.linesep) work_script.write(os.linesep) work_script.write(self.xnat_pbs_jobs_home + os.sep + '7T' + os.sep + 'DeDriftAndResampleHCP7T' + os.sep + 'DeDriftAndResampleHCP7T.XNAT.sh \\' + os.linesep) work_script.write(' --user="******" \\' + os.linesep) work_script.write(' --password="******" \\' + os.linesep) work_script.write(' --server="' + str_utils.get_server_name(self.server) + '" \\' + os.linesep) work_script.write(' --project="' + self.project + '" \\' + os.linesep) work_script.write(' --subject="' + self.subject + '" \\' + os.linesep) work_script.write(' --session="' + self.session + '" \\' + os.linesep) work_script.write(' --structural-reference-project="' + self.structural_reference_project + '" \\' + os.linesep) work_script.write(' --structural-reference-session="' + self.structural_reference_session + '" \\' + os.linesep) work_script.write(' --working-dir="' + working_directory_name + '" \\' + os.linesep) work_script.write(' --workflow-id="' + workflow_id + '" \\' + os.linesep) work_script.write(' --setup-script=' + self.setup_script + os.linesep) work_script.close() os.chmod(work_script_name, stat.S_IRWXU | stat.S_IRWXG) # Create script to put the results into the DB put_script_name = script_file_start_name + '.XNAT_PBS_PUT_job.sh' self.create_put_script(put_script_name, self.username, self.password, self.put_server, self.project, self.subject, self.session, working_directory_name, output_resource_name, self.PIPELINE_NAME) # Submit the job to do the work work_submit_cmd = 'qsub ' + work_script_name _inform("work_submit_cmd: " + work_submit_cmd) completed_work_submit_process = subprocess.run( work_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) work_job_no = str_utils.remove_ending_new_lines(completed_work_submit_process.stdout) _inform("work_job_no: " + work_job_no) # Submit the job to put the results in the DB put_submit_cmd = 'qsub -W depend=afterok:' + work_job_no + ' ' + put_script_name _inform("put_submit_cmd: " + put_submit_cmd) completed_put_submit_process = subprocess.run( put_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) put_job_no = str_utils.remove_ending_new_lines(completed_put_submit_process.stdout) _inform("put_job_no: " + put_job_no) else: _inform("Unable to submit jobs")
def submit_jobs(self, processing_stage=ProcessingStage.PUT_DATA): """ processing_stage is the last processing stage for which to submit the corresponding job. GET_DATA means just get the data. PROCESS_DATA means get the data and do the processing. PUT_DATA means get the data, processing it, and put the results back in the DB """ logger.debug("submit_jobs processing_stage: " + str(processing_stage)) if self.validate_parameters(): # determine what scans to run the RestingStateStats pipeline on for this subject # TBD: Does this get run on every scan for which the ICAFIX pipeline has been run, # or does it only get run on every resting state scan that has been fix processed. subject_info = hcp7t_subject.Hcp7TSubjectInfo( self.project, self.structural_reference_project, self.subject) fix_processed_scans = self.archive.available_FIX_processed_names( subject_info) fix_processed_scans_set = set(fix_processed_scans) logger.debug("fix_processed_scans_set = " + str(fix_processed_scans_set)) # resting_state_scans = self.archive.available_resting_state_preproc_names(subject_info) # resting_state_scans_set = set(resting_state_scans) # logger.debug("resting_state_scans_set = " + str(resting_state_scans_set)) # scans_to_process_set = resting_state_scans_set & fix_processed_scans_set scans_to_process_set = fix_processed_scans_set scans_to_process = list(scans_to_process_set) scans_to_process.sort() logger.debug("scans_to_process: " + str(scans_to_process)) incomplete_scans_to_process = list() for scan in scans_to_process: if (not is_complete(self.archive, subject_info, scan)): incomplete_scans_to_process.append(scan) logger.debug("incomplete_scans_to_process: " + str(incomplete_scans_to_process)) print("incomplete_scans_to_process:", incomplete_scans_to_process) # for scan in scans_to_process: for scan in incomplete_scans_to_process: logger.info("") logger.info( "--------------------------------------------------") logger.info("Submitting " + self.PIPELINE_NAME + " jobs for") logger.info(" Project: " + self.project) logger.info(" Subject: " + self.subject) logger.info(" Session: " + self.session) logger.info(" Structural Reference Project: " + self.structural_reference_project) logger.info(" Structural Reference Session: " + self.structural_reference_session) logger.info(" Scan: " + scan) logger.info(" Stage: " + str(processing_stage)) logger.info( "--------------------------------------------------") # make sure working directories do not have the same name based on # the same start time by sleeping a few seconds time.sleep(5) # build the working directory name self._working_directory_name = \ self.build_working_directory_name(self.project, self.PIPELINE_NAME, self.subject, scan) logger.info("Making working directory: " + self._working_directory_name) os.makedirs(name=self._working_directory_name) # get JSESSION ID jsession_id = xnat_access.get_jsession_id( server=os_utils.getenv_required( 'XNAT_PBS_JOBS_XNAT_SERVER'), username=self.username, password=self.password) logger.info("jsession_id: " + jsession_id) # get XNAT Session ID (a.k.a. the experiment ID, e.g. ConnectomeDB_E1234) xnat_session_id = xnat_access.get_session_id( server=os_utils.getenv_required( 'XNAT_PBS_JOBS_XNAT_SERVER'), username=self.username, password=self.password, project=self.project, subject=self.subject, session=self.session) logger.info("xnat_session_id: " + xnat_session_id) # get XNAT Workflow ID workflow_obj = xnat_access.Workflow( self.username, self.password, 'https://' + os_utils.getenv_required('XNAT_PBS_JOBS_XNAT_SERVER'), jsession_id) self._workflow_id = workflow_obj.create_workflow( xnat_session_id, self.project, self.PIPELINE_NAME + '_' + scan, 'Queued') logger.info("workflow_id: " + self._workflow_id) # determine output resource name self._output_resource_name = scan + "_RSS" # clean output resource if requested if self.clean_output_resource_first: logger.info("Deleting resource: " + self._output_resource_name + " for:") logger.info(" project: " + self.project) logger.info(" subject: " + self.subject) logger.info(" session: " + self.session) delete_resource.delete_resource( self.username, self.password, str_utils.get_server_name(self.server), self.project, self.subject, self.session, self._output_resource_name) # create scripts for various stages of processing if processing_stage >= ProcessingStage.PREPARE_SCRIPTS: # create script to get data self._create_get_data_script(scan) # create script to do work self._create_work_script(scan) # create script to clean data self._create_clean_data_script(scan) # create script to put the results into the DB put_script_name = self._put_data_script_name(scan) self.create_put_script(put_script_name, self.username, self.password, self.put_server, self.project, self.subject, self.session, self._working_directory_name, self._output_resource_name, self.PIPELINE_NAME + '_' + scan) # submit job to get the data if processing_stage >= ProcessingStage.GET_DATA: get_data_submit_cmd = 'qsub ' + self._get_data_script_name( scan) logger.info("get_data_submit_cmd: " + get_data_submit_cmd) completed_get_data_submit_process = subprocess.run( get_data_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) get_data_job_no = str_utils.remove_ending_new_lines( completed_get_data_submit_process.stdout) logger.info("get_data_job_no: " + get_data_job_no) else: logger.info("Get data job not submitted") # submit job to process the data if processing_stage >= ProcessingStage.PROCESS_DATA: work_submit_cmd = 'qsub -W depend=afterok:' + get_data_job_no + ' ' + self._work_script_name( scan) logger.info("work_submit_cmd: " + work_submit_cmd) completed_work_submit_process = subprocess.run( work_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) work_job_no = str_utils.remove_ending_new_lines( completed_work_submit_process.stdout) logger.info("work_job_no: " + work_job_no) else: logger.info("Process data job not submitted") # submit job to clean the data if processing_stage >= ProcessingStage.CLEAN_DATA: clean_submit_cmd = 'qsub -W depend=afterok:' + work_job_no + ' ' + self._clean_data_script_name( scan) logger.info("clean_submit_cmd: " + clean_submit_cmd) completed_clean_submit_process = subprocess.run( clean_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) clean_job_no = str_utils.remove_ending_new_lines( completed_clean_submit_process.stdout) logger.info("clean_job_no: " + clean_job_no) else: logger.info("Clean data job not submitted") # submit job to put the resulting data in the DB if processing_stage >= ProcessingStage.PUT_DATA: put_submit_cmd = 'qsub -W depend=afterok:' + clean_job_no + ' ' + put_script_name logger.info("put_submit_cmd: " + put_submit_cmd) completed_put_submit_process = subprocess.run( put_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) put_job_no = str_utils.remove_ending_new_lines( completed_put_submit_process.stdout) logger.info("put_job_no: " + put_job_no) else: logger.info("Put data job not submitted") else: logger.info("Unable to submit jobs")
def submit_jobs(self): _debug("submit_jobs") if self.validate_parameters(): _inform("") _inform("--------------------------------------------------") _inform("Submitting " + self.PIPELINE_NAME + " jobs for") _inform(" Project: " + self.project) _inform(" Subject: " + self.subject) _inform(" Session: " + self.session) _inform("--------------------------------------------------") # make sure working directories don't have the same name based on # the same start time by sleeping a few seconds time.sleep(5) current_seconds_since_epoch = int(time.time()) # build the working directory name self._working_directory_name = self.build_home self._working_directory_name += os.sep + self.project self._working_directory_name += os.sep + self.PIPELINE_NAME self._working_directory_name += '.' + self.subject self._working_directory_name += '.' + str( current_seconds_since_epoch) # make the working directory _inform("making working directory: " + self._working_directory_name) os.makedirs(name=self._working_directory_name) # get JSESSION ID jsession_id = xnat_access.get_jsession_id( server=os_utils.getenv_required('XNAT_PBS_JOBS_XNAT_SERVER'), username=self.username, password=self.password) _inform("jsession_id: " + jsession_id) # get XNAT Session ID (a.k.a. the experiment ID, e.g. ConnectomeDB_E1234) xnat_session_id = xnat_access.get_session_id( server=os_utils.getenv_required('XNAT_PBS_JOBS_XNAT_SERVER'), username=self.username, password=self.password, project=kself.project, subject=self.subject, session=self.session) _inform("xnat_session_id: " + xnat_session_id) # get XNAT Workflow ID workflow_obj = xnat_access.Workflow( self.username, self.password, os_utils.getenv_required('XNAT_PBS_JOBS_XNAT_SERVER'), jsession_id) self._workflow_id = workflow_obj.create_workflow( xnat_session_id, self.project, self.PIPELINE_NAME, 'Queued') _inform("workflow_id: " + self._workflow_id) # determine output resource name self._output_resource_name = 'Diffusion_preproc' # clean the output resource if requested if self.clean_output_resource_first: _inform("Deleting resource: " + self._output_resource_name + " for:") _inform(" project: " + self.project) _inform(" subject: " + self.subject) _inform(" session: " + self.session) delete_resource.delete_resource( self.username, self.password, str_utils.get_server_name(self.server), self.project, self.subject, self.session, self._output_resource_name) # create script to do the PreEddy work self._create_pre_eddy_script() # create script to do the Eddy work self._create_eddy_script() # create script to do the PostEddy work self._create_post_eddy_script() # create script to put the results into the DB put_script_name = self._get_scripts_start_name( ) + '.XNAT_PBS_PUT_job.sh' self.create_put_script(put_script_name, self.username, self.password, self.put_server, self.project, self.subject, self.session, self._working_directory_name, self._output_resource_name, self.PIPELINE_NAME) # Submit the job to do the Pre-Eddy work pre_eddy_submit_cmd = 'qsub ' + self._pre_eddy_script_name _inform("pre_eddy_submit_cmd: " + pre_eddy_submit_cmd) completed_pre_eddy_submit_process = subprocess.run( pre_eddy_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) pre_eddy_job_no = str_utils.remove_ending_new_lines( completed_pre_eddy_submit_process.stdout) _inform("pre_eddy_job_no: " + pre_eddy_job_no) # Submit the job to do the Eddy work eddy_submit_cmd = 'qsub -W depend=afterok:' + pre_eddy_job_no + ' ' + self._eddy_script_name _inform("eddy_submit_cmd: " + eddy_submit_cmd) completed_eddy_submit_process = subprocess.run( eddy_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) eddy_job_no = str_utils.remove_ending_new_lines( completed_eddy_submit_process.stdout) _inform("eddy_job_no: " + eddy_job_no) # Submit the job to do the Post-Eddy work post_eddy_submit_cmd = 'qsub -W depend=afterok:' + eddy_job_no + ' ' + self._post_eddy_script_name _inform("post_eddy_submit_cmd: " + post_eddy_submit_cmd) completed_post_eddy_submit_process = subprocess.run( post_eddy_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) post_eddy_job_no = str_utils.remove_ending_new_lines( completed_post_eddy_submit_process.stdout) _inform("post_eddy_job_no: " + post_eddy_job_no) # Submit the job to put the results in the DB put_submit_cmd = 'qsub -W depend=afterok:' + post_eddy_job_no + ' ' + put_script_name _inform("put_submit_cmd: " + put_submit_cmd) completed_put_submit_process = subprocess.run( put_submit_cmd, shell=True, check=True, stdout=subprocess.PIPE, universal_newlines=True) put_job_no = str_utils.remove_ending_new_lines( completed_put_submit_process.stdout) _inform("put_job_no: " + put_job_no) else: _inform("Unable to submit jobs")