htsworkflow/submission/submission.py

   1 """Common submission elements
   2 """
   3 import logging
   4 import os
   5 import re
   6
   7 import RDF
   8
   9 from htsworkflow.util.rdfhelp import \
  10      blankOrUri, \
  11      dafTermOntology, \
  12      dump_model, \
  13      get_model, \
  14      libraryOntology, \
  15      owlNS, \
  16      rdfNS, \
  17      submissionLog, \
  18      submissionOntology, \
  19      toTypedNode, \
  20      fromTypedNode
  21 from htsworkflow.util.hashfile import make_md5sum
  22 from htsworkflow.submission.fastqname import FastqName
  23 from htsworkflow.submission.daf import \
  24      MetadataLookupException, \
  25      ModelException, \
  26      get_submission_uri
  27
  28 LOGGER = logging.getLogger(__name__)
  29
  30 class Submission(object):
  31     def __init__(self, name, model, host):
  32         self.name = name
  33         self.model = model
  34
  35         self.submissionSet = get_submission_uri(self.name)
  36         self.submissionSetNS = RDF.NS(str(self.submissionSet) + '#')
  37         self.libraryNS = RDF.NS('{0}/library/'.format(host))
  38
  39         self.__view_map = None
  40
  41     def scan_submission_dirs(self, result_map):
  42         """Examine files in our result directory
  43         """
  44         for lib_id, result_dir in result_map.items():
  45             LOGGER.info("Importing %s from %s" % (lib_id, result_dir))
  46             try:
  47                 self.import_analysis_dir(result_dir, lib_id)
  48             except MetadataLookupException, e:
  49                 LOGGER.error("Skipping %s: %s" % (lib_id, str(e)))
  50
  51     def import_analysis_dir(self, analysis_dir, library_id):
  52         """Import a submission directories and update our model as needed
  53         """
  54         #attributes = get_filename_attribute_map(paired)
  55         libNode = self.libraryNS[library_id + "/"]
  56
  57         self._add_library_details_to_model(libNode)
  58
  59         submission_files = os.listdir(analysis_dir)
  60         for filename in submission_files:
  61             pathname = os.path.abspath(os.path.join(analysis_dir, filename))
  62             self.construct_file_attributes(analysis_dir, libNode, pathname)
  63
  64     def construct_file_attributes(self, analysis_dir, libNode, pathname):
  65         """Looking for the best extension
  66         The 'best' is the longest match
  67
  68         :Args:
  69         filename (str): the filename whose extention we are about to examine
  70         """
  71         path, filename = os.path.split(pathname)
  72
  73         LOGGER.debug("Searching for view")
  74         file_type = self.find_best_match(filename)
  75         if file_type is None:
  76             LOGGER.warn("Unrecognized file: {0}".format(pathname))
  77             return None
  78         if str(file_type) == str(libraryOntology['ignore']):
  79             return None
  80
  81         an_analysis_name = self.make_submission_name(analysis_dir)
  82         an_analysis = self.get_submission_node(analysis_dir)
  83         an_analysis_uri = str(an_analysis.uri)
  84         file_classification = self.model.get_target(file_type,
  85                                                     rdfNS['type'])
  86         if file_classification is None:
  87             errmsg = 'Could not find class for {0}'
  88             LOGGER.warning(errmsg.format(str(file_type)))
  89             return
  90
  91         self.model.add_statement(
  92             RDF.Statement(self.submissionSetNS[''],
  93                           submissionOntology['has_submission'],
  94                           an_analysis))
  95         self.model.add_statement(RDF.Statement(an_analysis,
  96                                                submissionOntology['name'],
  97                                                toTypedNode(an_analysis_name)))
  98         self.model.add_statement(
  99             RDF.Statement(an_analysis,
 100                           rdfNS['type'],
 101                           submissionOntology['submission']))
 102         self.model.add_statement(RDF.Statement(an_analysis,
 103                                                submissionOntology['library'],
 104                                                libNode))
 105
 106         LOGGER.debug("Adding statements to {0}".format(str(an_analysis)))
 107         # add track specific information
 108         self.model.add_statement(
 109             RDF.Statement(an_analysis,
 110                           dafTermOntology['paired'],
 111                           toTypedNode(self._is_paired(libNode))))
 112         self.model.add_statement(
 113             RDF.Statement(an_analysis,
 114                           dafTermOntology['submission'],
 115                           an_analysis))
 116
 117         # add file specific information
 118         fileNode = self.make_file_node(pathname, an_analysis)
 119         self.add_md5s(filename, fileNode, analysis_dir)
 120         self.add_fastq_metadata(filename, fileNode)
 121         self.model.add_statement(
 122             RDF.Statement(fileNode,
 123                           rdfNS['type'],
 124                           file_type))
 125         LOGGER.debug("Done.")
 126
 127     def make_file_node(self, pathname, submissionNode):
 128         """Create file node and attach it to its submission.
 129         """
 130         # add file specific information
 131         path, filename = os.path.split(pathname)
 132         fileNode = RDF.Node(RDF.Uri('file://'+ os.path.abspath(pathname)))
 133         self.model.add_statement(
 134             RDF.Statement(submissionNode,
 135                           dafTermOntology['has_file'],
 136                           fileNode))
 137         self.model.add_statement(
 138             RDF.Statement(fileNode,
 139                           dafTermOntology['filename'],
 140                           filename))
 141         return fileNode
 142
 143     def add_md5s(self, filename, fileNode, analysis_dir):
 144         LOGGER.debug("Updating file md5sum")
 145         submission_pathname = os.path.join(analysis_dir, filename)
 146         md5 = make_md5sum(submission_pathname)
 147         if md5 is None:
 148             errmsg = "Unable to produce md5sum for {0}"
 149             LOGGER.warning(errmsg.format(submission_pathname))
 150         else:
 151             self.model.add_statement(
 152                 RDF.Statement(fileNode, dafTermOntology['md5sum'], md5))
 153
 154     def add_fastq_metadata(self, filename, fileNode):
 155         # How should I detect if this is actually a fastq file?
 156         try:
 157             fqname = FastqName(filename=filename)
 158         except ValueError:
 159             # currently its just ignore it if the fastq name parser fails
 160             return
 161
 162         terms = [('flowcell', libraryOntology['flowcell_id']),
 163                  ('lib_id', libraryOntology['library_id']),
 164                  ('lane', libraryOntology['lane_number']),
 165                  ('read', libraryOntology['read']),
 166                  ('cycle', libraryOntology['read_length'])]
 167         for file_term, model_term in terms:
 168             value = fqname.get(file_term)
 169             if value is not None:
 170                 s = RDF.Statement(fileNode, model_term, toTypedNode(value))
 171                 self.model.append(s)
 172
 173     def _add_library_details_to_model(self, libNode):
 174         # attributes that can have multiple values
 175         set_attributes = set((libraryOntology['has_lane'],
 176                               libraryOntology['has_mappings'],
 177                               dafTermOntology['has_file']))
 178         parser = RDF.Parser(name='rdfa')
 179         new_statements = parser.parse_as_stream(libNode.uri)
 180         toadd = []
 181         for s in new_statements:
 182             # always add "collections"
 183             if s.predicate in set_attributes:
 184                 toadd.append(s)
 185                 continue
 186             # don't override things we already have in the model
 187             targets = list(self.model.get_targets(s.subject, s.predicate))
 188             if len(targets) == 0:
 189                 toadd.append(s)
 190
 191         for s in toadd:
 192             self.model.append(s)
 193
 194         self._add_lane_details(libNode)
 195
 196     def _add_lane_details(self, libNode):
 197         """Import lane details
 198         """
 199         query = RDF.Statement(libNode, libraryOntology['has_lane'], None)
 200         lanes = []
 201         for lane_stmt in self.model.find_statements(query):
 202             lanes.append(lane_stmt.object)
 203
 204         parser = RDF.Parser(name='rdfa')
 205         for lane in lanes:
 206             LOGGER.debug("Importing %s" % (lane.uri,))
 207             try:
 208                 parser.parse_into_model(self.model, lane.uri)
 209             except RDF.RedlandError, e:
 210                 LOGGER.error("Error accessing %s" % (lane.uri,))
 211                 raise e
 212
 213
 214     def find_best_match(self, filename):
 215         """Search through potential filename matching patterns
 216         """
 217         if self.__view_map is None:
 218             self.__view_map = self._get_filename_view_map()
 219
 220         results = []
 221         for pattern, view in self.__view_map.items():
 222             if re.match(pattern, filename):
 223                 results.append(view)
 224
 225         if len(results) > 1:
 226             msg = "%s matched multiple views %s" % (
 227                 filename,
 228                 [str(x) for x in results])
 229             raise ModelException(msg)
 230         elif len(results) == 1:
 231             return results[0]
 232         else:
 233             return None
 234
 235     def _get_filename_view_map(self):
 236         """Query our model for filename patterns
 237
 238         return a dictionary of compiled regular expressions to view names
 239         """
 240         filename_query = RDF.Statement(
 241             None, dafTermOntology['filename_re'], None)
 242
 243         patterns = {}
 244         for s in self.model.find_statements(filename_query):
 245             view_name = s.subject
 246             literal_re = s.object.literal_value['string']
 247             LOGGER.debug("Found: %s" % (literal_re,))
 248             try:
 249                 filename_re = re.compile(literal_re)
 250             except re.error, e:
 251                 LOGGER.error("Unable to compile: %s" % (literal_re,))
 252             patterns[literal_re] = view_name
 253         return patterns
 254
 255     def make_submission_name(self, analysis_dir):
 256         analysis_dir = os.path.normpath(analysis_dir)
 257         analysis_dir_name = os.path.split(analysis_dir)[1]
 258         if len(analysis_dir_name) == 0:
 259             raise RuntimeError(
 260                 "Submission dir name too short: {0}".format(analysis_dir))
 261         return analysis_dir_name
 262
 263     def get_submission_node(self, analysis_dir):
 264         """Convert a submission directory name to a submission node
 265         """
 266         submission_name = self.make_submission_name(analysis_dir)
 267         return self.submissionSetNS[submission_name]
 268
 269     def _get_library_attribute(self, libNode, attribute):
 270         if not isinstance(attribute, RDF.Node):
 271             attribute = libraryOntology[attribute]
 272
 273         targets = list(self.model.get_targets(libNode, attribute))
 274         if len(targets) > 0:
 275             return self._format_library_attribute(targets)
 276         else:
 277             return None
 278
 279         #targets = self._search_same_as(libNode, attribute)
 280         #if targets is not None:
 281         #    return self._format_library_attribute(targets)
 282
 283         # we don't know anything about this attribute
 284         self._add_library_details_to_model(libNode)
 285
 286         targets = list(self.model.get_targets(libNode, attribute))
 287         if len(targets) > 0:
 288             return self._format_library_attribute(targets)
 289
 290         return None
 291
 292     def _format_library_attribute(self, targets):
 293         if len(targets) == 0:
 294             return None
 295         elif len(targets) == 1:
 296             return fromTypedNode(targets[0])
 297         elif len(targets) > 1:
 298             return [fromTypedNode(t) for t in targets]
 299
 300     def _is_paired(self, libNode):
 301         """Determine if a library is paired end"""
 302         library_type = self._get_library_attribute(libNode, 'library_type')
 303         if library_type is None:
 304             errmsg = "%s doesn't have a library type"
 305             raise ModelException(errmsg % (str(libNode),))
 306
 307         single = ['CSHL (lacking last nt)',
 308                   'Single End (non-multiplexed)',
 309                   'Small RNA (non-multiplexed)',]
 310         paired = ['Barcoded Illumina',
 311                   'Multiplexing',
 312                   'Nextera',
 313                   'Paired End (non-multiplexed)',]
 314         if library_type in single:
 315             return False
 316         elif library_type in paired:
 317             return True
 318         else:
 319             raise MetadataLookupException(
 320                 "Unrecognized library type %s for %s" % \
 321                 (library_type, str(libNode)))
 322
 323     def execute_query(self, template, context):
 324         """Execute the query, returning the results
 325         """
 326         formatted_query = template.render(context)
 327         LOGGER.debug(formatted_query)
 328         query = RDF.SPARQLQuery(str(formatted_query))
 329         rdfstream = query.execute(self.model)
 330         results = []
 331         for record in rdfstream:
 332             d = {}
 333             for key, value in record.items():
 334                 d[key] = fromTypedNode(value)
 335             results.append(d)
 336         return results