def get_cp_pipeline_path(self):
     filename_regex_obj = re.compile(fntranslate(
         self.cp_pipeline_fnpattern))
     # Look both in PIPES folder and process folder of the pipe (named
     # after pipe).
     for location in (self.process_path, self.pipes_path):
         cp_pipeline_files = [
             os.path.join(location, filename)
             for filename in os.listdir(location)
             if filename_regex_obj.match(filename)
         ]
         # Check if found anything and complain otherwise.
         if len(cp_pipeline_files) > 1:
             raise BrainyProcessError(
                 'More than one CP pipeline settings '
                 'file found matching: %s in %s' %
                 (self.cp_pipeline_fnpattern, self.process_path))
         elif len(cp_pipeline_files) == 1:
             # First found pipeline. Rest is ignored.
             return cp_pipeline_files[0]
     # Failed to find anything.
     raise BrainyProcessError(
         'No CP pipeline settings file'
         ' found matching: %s in %s' %
         (self.cp_pipeline_fnpattern, self.process_path))
    def submit(self):
        results = list()
        for fused_filename in self.fused_files:
            matlab_code = self.get_matlab_code(fused_filename)
            fused_report = fused_filename.replace(
                '.mat', '.results_%s' % get_timestamp_str())
            submission_result = self.submit_matlab_job(
                matlab_code,
                report_file=fused_report,
            )
            results.append(submission_result)

        if not results:
            raise BrainyProcessError(warning='Failed to find complete '
                                     'measurements.. check or restart previous'
                                     ' step')

        print(
            '''
            <status action="%(step_name)s">submitting (%(results_count)d) fusion jobs..
            <output>%(submission_result)s</output>
            </status>
        ''' % {
                'step_name': self.step_name,
                'results_count': len(results),
                'submission_result': escape_xml(str(results)),
            })

        self.set_flag('submitted')
    def resubmit(self):
        results = list()
        for fused_filename in self.fused_files:
            matlab_code = self.get_matlab_code(fused_filename)
            fused_report = 'DataFusion_' + fused_filename.replace(
                '.mat', '.results_%s' % get_timestamp_str())
            resubmission_result = self.submit_matlab_job(
                matlab_code,
                report_file=fused_report,
                is_resubmitting=True,
            )
            results.append(resubmission_result)

        if not results:
            raise BrainyProcessError(warning='Failed to resubmit data fusion')

        print(
            '''
            <status action="%(step_name)s">resubmitting (%(results_count)d) fusion jobs..
            <output>%(resubmission_result)s</output>
            </status>
        ''' % {
                'step_name': self.step_name,
                'results_count': len(results),
                'resubmission_result': escape_xml(str(results)),
            })

        self.set_flag('resubmitted')
        super(CPDataFusion, self).resubmit()
    def submit(self):
        results = list()
        for batch_filename in self.batch_files:
            matlab_code = self.get_matlab_code(batch_filename)
            batch_report = batch_filename.replace(
                '.mat', '.results_%s' % get_timestamp_str())
            submission_result = self.submit_matlab_job(
                matlab_code,
                report_file=batch_report,
            )
            results.append(submission_result)

        if not results:
            # We want to explicitly check for errors and report them, if any
            # found.
            self.check_logs_for_errors()
            # If no errors were found, we still have to report a warning!
            raise BrainyProcessError(warning='Failed to find any batches.. '
                                     'check or restart previous step')

        print(
            '''
            <status action="%(step_name)s">submitting (%(batch_count)d) batches..
            <output>%(submission_result)s</output>
            </status>
        ''' % {
                'step_name': self.step_name,
                'batch_count': len(results),
                'submission_result': escape_xml(str(results)),
            })

        self.set_flag('submitted')
 def get_cp_pipeline_path(self):
     filename_regex_obj = re.compile(fntranslate(
         self.cp_pipeline_fnpattern))
     cp_pipeline_files = [
         filename for filename in os.listdir(self.process_path)
         if filename_regex_obj.match(filename)
     ]
     if len(cp_pipeline_files) > 1:
         raise BrainyProcessError(
             'More than one CP pipeline settings file'
             ' found matching: %s in %s' %
             (self.cp_pipeline_fnpattern, self.process_path))
     elif len(cp_pipeline_files) == 0:
         raise BrainyProcessError(
             'No CP pipeline settings file'
             ' found matching: %s in %s' %
             (self.cp_pipeline_fnpattern, self.process_path))
     return os.path.join(self.process_path, cp_pipeline_files[0])
    def resubmit(self):
        resubmission_results = list()
        output_batches = self.get_out_files()
        result_batches = self.get_result_files()
        for batch_filename in self.batch_files:
            # Resubmit only those files that have no data, i.e. failed with no
            # output.
            batch_prefix = self.parse_batch(batch_filename)
            if batch_prefix in output_batches \
                    and batch_prefix in result_batches:
                # This batch is complete. Do not resubmit it.
                print '<!-- CP batch %s is complete. Skipping.. -->' % \
                      batch_filename
                continue
            # Batch has not produced any output yet. We can resubmit it.
            matlab_code = self.get_matlab_code(batch_filename)
            batch_report = batch_filename.replace(
                '.mat', '.results_%s' % get_timestamp_str())
            resubmission_result = self.submit_matlab_job(
                matlab_code,
                report_file=batch_report,
                is_resubmitting=True,
            )
            resubmission_results.append(resubmission_result)

        if not resubmission_results:
            # We want to explicitly check for errors and report them, if any
            # found.
            self.check_logs_for_errors()
            # If no errors were found, we still have to report a warning!
            raise BrainyProcessError(warning='Failed to find any batches to '
                                     'resubmit. Check or restart previous '
                                     'step.')

        print(
            '''
            <status action="%(step_name)s">resubmitting (%(batch_count)d) batches..
            <output>%(resubmission_result)s</output>
            </status>
        ''' % {
                'step_name': self.step_name,
                'batch_count': len(resubmission_results),
                'resubmission_result': escape_xml(str(resubmission_results)),
            })

        self.set_flag('resubmitted')
        super(CPCluster, self).resubmit()
Exemple #7
0
    def has_data(self):
        '''
        Check if all files matching given patterns have been linked.
        '''
        print self.target_location
        if not os.path.exists(self.target_location):
            raise BrainyProcessError(warning='Expected target folder is not '
                                     'found: %s' % self.target_location)

        def get_name(root, name):
            return name

        linking_per_file_type = {
            'f': ['hardlink', 'symlink'],
            'd': ['symlink'],
        }

        for file_type in linking_per_file_type:
            linking = linking_per_file_type[file_type]
            for link_type in linking:
                if link_type in self.file_patterns:
                    patterns = self.file_patterns[link_type]
                    source_matches = list(
                        find_files(
                            path=self.source_location,
                            match=MatchAnyPatternsAndTypes(
                                filetypes=[file_type],
                                names=patterns,
                            ),
                            collect=get_name,
                            recursive=self.recursively,
                        ))
                    target_matches = list(
                        find_files(
                            path=self.target_location,
                            match=MatchAnyPatternsAndTypes(
                                filetypes=[file_type],
                                names=patterns,
                            ),
                            collect=get_name,
                            recursive=self.recursively,
                        ))
                    if not source_matches == target_matches:
                        return False
        return True
Exemple #8
0
 def build_linking_args(source_location, target_location,
                        nested_file_patterns, file_type, recursively):
     for link_type in ['hardlink', 'symlink']:
         if link_type in nested_file_patterns:
             if type(nested_file_patterns[link_type]) != list \
                     or len(nested_file_patterns[link_type]) == 0:
                 raise BrainyProcessError(
                     warning='LinkFiles process requires a non empty list '
                     'of file patterns which can be match to '
                     'files in source_location.')
             args = {
                 'source_location': source_location,
                 'target_location': target_location,
                 'file_patterns': nested_file_patterns[link_type],
                 'link_type': link_type,
                 'file_type': file_type,
                 'recursively': recursively,
             }
             yield args
Exemple #9
0
 def link(source_path,
          target_path,
          patterns,
          link_type='hard',
          file_type='f',
          recursively=False):
     '''
     Expect keys 'hardlink' and 'symlink' keys in
     description['file_patterns']. If pattern string starts and ends with
     '/' then it is a regexp, otherwise it is fnmatch.
     '''
     assert os.path.exists(source_path)
     assert os.path.exists(target_path)
     file_matches = find_files(
         path=source_path,
         match=MatchAnyPatternsAndTypes(
             filetypes=[file_type],
             names=patterns,
         ),
         recursive=recursively,
     )
     if link_type == 'hardlink' and file_type == 'f':
         make_link = os.link
     elif link_type == 'symlink':
         make_link = os.symlink
     else:
         raise IOError('Unsupported link type: %s' % link_type)
     for source_file in file_matches:
         link_path = os.path.join(target_path,
                                  os.path.basename(source_file))
         try:
             print 'Linking "%s" -> "%s"' % (source_file, link_path)
             make_link(source_file, link_path)
         except IOError as error:
             if 'File exists' in str(error):
                 message = 'It looks like linking was already done. Maybe '\
                     'you are trying to re-run project incorrectly. Make '\
                     'sure to clean previous results before retrying.'
             else:
                 message = 'Unknown input-output error.'
             raise BrainyProcessError(warning=message, output=str(error))