1 """Convert srf and qseq archive files to fastqs
5 from pprint import pformat,pprint
8 from urlparse import urljoin, urlparse
10 from htsworkflow.pipelines.sequences import scan_for_sequences, \
11 update_model_sequence_library
12 from htsworkflow.pipelines.samplekey import SampleKey
13 from htsworkflow.pipelines import qseq2fastq
14 from htsworkflow.pipelines import srf2fastq
15 from htsworkflow.pipelines import desplit_fastq
16 from htsworkflow.submission.fastqname import FastqName
17 from htsworkflow.util.rdfhelp import get_model, dump_model, load_into_model, \
20 from htsworkflow.util.rdfns import *
21 from htsworkflow.util.conversion import parse_flowcell_id
23 from django.conf import settings
24 from django.template import Context, loader
28 LOGGER = logging.getLogger(__name__)
31 class CondorFastqExtract(object):
32 def __init__(self, host, sequences_path,
36 """Extract fastqs from results archive
39 host (str): root of the htsworkflow api server
40 apidata (dict): id & key to post to the server
41 sequences_path (str): root of the directory tree to scan for files
42 log_path (str): where to put condor log files
43 force (bool): do we force overwriting current files?
46 self.model = get_model(model)
47 self.sequences_path = sequences_path
48 self.log_path = log_path
50 LOGGER.info("CondorFastq host={0}".format(self.host))
51 LOGGER.info("CondorFastq sequences_path={0}".format(self.sequences_path))
52 LOGGER.info("CondorFastq log_path={0}".format(self.log_path))
54 def create_scripts(self, result_map ):
56 Generate condor scripts to build any needed fastq files
59 result_map: htsworkflow.submission.results.ResultMap()
61 template_map = {'srf': 'srf.condor',
62 'qseq': 'qseq.condor',
63 'split_fastq': 'split_fastq.condor',
67 pythonpath = os.environ.get('PYTHONPATH', None)
68 if pythonpath is not None:
69 env = "PYTHONPATH=%s" % (pythonpath,)
70 condor_entries = self.build_condor_arguments(result_map)
71 for script_type in template_map.keys():
72 template = loader.get_template(template_map[script_type])
73 variables = {'python': sys.executable,
74 'logdir': self.log_path,
76 'args': condor_entries[script_type],
77 'root_url': self.host,
79 context = Context(variables)
81 with open(script_type + '.condor','w+') as outstream:
82 outstream.write(template.render(context))
84 def build_condor_arguments(self, result_map):
85 condor_entries = {'srf': [],
89 conversion_funcs = {'srf': self.condor_srf_to_fastq,
90 'qseq': self.condor_qseq_to_fastq,
91 'split_fastq': self.condor_desplit_fastq
93 sequences = self.find_archive_sequence_files(result_map)
94 needed_targets = self.update_fastq_targets(result_map, sequences)
96 for target_pathname, available_sources in needed_targets.items():
97 LOGGER.debug(' target : %s' % (target_pathname,))
98 LOGGER.debug(' candidate sources: %s' % (available_sources,))
99 for condor_type in available_sources.keys():
100 conversion = conversion_funcs.get(condor_type, None)
101 if conversion is None:
102 errmsg = "Unrecognized type: {0} for {1}"
103 LOGGER.error(errmsg.format(condor_type,
104 pformat(available_sources)))
106 sources = available_sources.get(condor_type, None)
108 if sources is not None:
109 condor_entries.setdefault(condor_type, []).append(
110 conversion(sources, target_pathname))
112 LOGGER.warn(" need file %s", target_pathname)
114 return condor_entries
116 def find_archive_sequence_files(self, result_map):
118 Find archived sequence files associated with our results.
120 self.import_libraries(result_map)
121 flowcell_ids = self.find_relevant_flowcell_ids()
122 self.import_sequences(flowcell_ids)
125 prefix libns: <http://jumpgate.caltech.edu/wiki/LibraryOntology#>
126 prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
127 prefix xsd: <http://www.w3.org/2001/XMLSchema#>
129 select ?filenode ?filetype ?cycle ?lane_number ?read
131 ?flowcell ?flowcell_id ?read_length
132 ?flowcell_type ?flowcell_status
134 ?filenode libns:cycle ?cycle ;
135 libns:lane_number ?lane_number ;
137 libns:flowcell ?flowcell ;
138 libns:flowcell_id ?flowcell_id ;
139 libns:library ?library ;
140 libns:library_id ?library_id ;
141 libns:file_type ?filetype ;
142 a libns:IlluminaResult .
143 ?flowcell libns:read_length ?read_length ;
144 libns:flowcell_type ?flowcell_type .
145 OPTIONAL { ?flowcell libns:flowcell_status ?flowcell_status }
146 FILTER(?filetype != libns:sequencer_result)
149 LOGGER.debug("find_archive_sequence_files query: %s",
151 query = RDF.SPARQLQuery(query_text)
153 for r in query.execute(self.model):
154 library_id = fromTypedNode(r['library_id'])
155 if library_id in result_map:
156 seq = SequenceResult(r)
157 LOGGER.debug("Creating sequence result for library %s: %s",
163 def import_libraries(self, result_map):
164 for lib_id in result_map.keys():
165 lib_id_encoded = lib_id.encode('utf-8')
166 liburl = urljoin(self.host, 'library/%s/' % (lib_id_encoded,))
167 library = RDF.Node(RDF.Uri(liburl))
168 self.import_library(library)
170 def import_library(self, library):
171 """Import library data into our model if we don't have it already
173 q = RDF.Statement(library, rdfNS['type'], libraryOntology['Library'])
175 if not self.model.contains_statement(q):
177 load_into_model(self.model, 'rdfa', library)
178 LOGGER.debug("Did we import %s: %s", library.uri, present)
180 def find_relevant_flowcell_ids(self):
181 """Generate set of flowcell ids that had samples of interest on them
183 flowcell_query = RDF.SPARQLQuery("""
184 prefix libns: <http://jumpgate.caltech.edu/wiki/LibraryOntology#>
186 select distinct ?flowcell ?flowcell_id
188 ?library a libns:Library ;
189 libns:has_lane ?lane .
190 ?lane libns:flowcell ?flowcell .
191 ?flowcell libns:flowcell_id ?flowcell_id .
194 for r in flowcell_query.execute(self.model):
195 flowcell_ids.add( fromTypedNode(r['flowcell_id']) )
197 a_lane = self.model.get_target(r['flowcell'],
198 libraryOntology['has_lane'])
201 # we lack information about which lanes were on this flowcell
202 load_into_model(self.model, 'rdfa', r['flowcell'])
203 LOGGER.debug("Did we imported %s: %s" % (r['flowcell'].uri,
208 def import_sequences(self, flowcell_ids):
210 for f in flowcell_ids:
211 seq_dirs.append(os.path.join(self.sequences_path, str(f)))
212 sequences = scan_for_sequences(seq_dirs)
213 for seq in sequences:
214 seq.save_to_model(self.model, self.host)
215 update_model_sequence_library(self.model, self.host)
217 def update_fastq_targets(self, result_map, raw_files):
218 """Return list of fastq files that need to be built.
220 Also update model with link between illumina result files
221 and our target fastq file.
223 fastq_paired_template = '%(lib_id)s_%(flowcell)s_c%(cycle)s_l%(lane)s_r%(read)s.fastq'
224 fastq_single_template = '%(lib_id)s_%(flowcell)s_c%(cycle)s_l%(lane)s.fastq'
225 # find what targets we're missing
227 for seq in raw_files:
230 filename_attributes = {
231 'flowcell': seq.flowcell_id,
232 'lib_id': seq.library_id,
233 'lane': seq.lane_number,
236 'is_paired': seq.ispaired
239 fqName = FastqName(**filename_attributes)
241 result_dir = result_map[seq.library_id]
242 target_pathname = os.path.join(result_dir, fqName.filename)
243 if self.force or not os.path.exists(target_pathname):
244 t = needed_targets.setdefault(target_pathname, {})
245 t.setdefault(seq.filetype, []).append(seq)
246 self.add_target_source_links(target_pathname, seq)
247 return needed_targets
249 def add_target_source_links(self, target, seq):
250 """Add link between target pathname and the 'lane' that produced it
251 (note lane objects are now post demultiplexing.)
253 target_uri = 'file://' + target.encode('utf-8')
254 target_node = RDF.Node(RDF.Uri(target_uri))
255 source_stmt = RDF.Statement(target_node, dcNS['source'], seq.filenode)
256 self.model.add_statement(source_stmt)
258 def condor_srf_to_fastq(self, sources, target_pathname):
260 raise ValueError("srf to fastq can only handle one file")
263 if sources[0].flowcell_id == '30DY0AAXX':
267 'sources': [sources[0].path],
268 'pyscript': srf2fastq.__file__,
269 'flowcell': sources[0].flowcell_id,
270 'ispaired': sources[0].ispaired,
271 'target': target_pathname,
272 'target_right': target_pathname.replace('_r1.fastq', '_r2.fastq'),
277 def condor_qseq_to_fastq(self, sources, target_pathname):
279 for source in sources:
280 paths.append(source.path)
283 'pyscript': qseq2fastq.__file__,
284 'flowcell': sources[0].flowcell_id,
285 'target': target_pathname,
287 'ispaired': sources[0].ispaired,
288 'istar': len(sources) == 1,
291 def condor_desplit_fastq(self, sources, target_pathname):
293 for source in sources:
294 paths.append(source.path)
297 'pyscript': desplit_fastq.__file__,
298 'target': target_pathname,
300 'ispaired': sources[0].ispaired,
304 def make_lane_dict(lib_db, lib_id):
306 Convert the lane_set in a lib_db to a dictionary
307 indexed by flowcell ID
310 for lane in lib_db[lib_id]['lane_set']:
311 flowcell_id, status = parse_flowcell_id(lane['flowcell'])
312 lane['flowcell'] = flowcell_id
313 result.append((lane['flowcell'], lane))
316 class SequenceResult(object):
317 """Convert the sparql query result from find_archive_sequence_files
319 def __init__(self, result):
320 self.filenode = result['filenode']
321 self._filetype = result['filetype']
322 self.cycle = fromTypedNode(result['cycle'])
323 self.lane_number = fromTypedNode(result['lane_number'])
324 self.read = fromTypedNode(result['read'])
325 if type(self.read) in types.StringTypes:
327 self.library = result['library']
328 self.library_id = fromTypedNode(result['library_id'])
329 self.flowcell = result['flowcell']
330 self.flowcell_id = fromTypedNode(result['flowcell_id'])
331 self.flowcell_type = fromTypedNode(result['flowcell_type'])
332 self.flowcell_status = fromTypedNode(result['flowcell_status'])
335 """is this sequence / flowcell 'good enough'"""
336 if self.flowcell_status is not None and \
337 self.flowcell_status.lower() == "failed":
340 isgood = property(_is_good)
342 def _get_ispaired(self):
343 if self.flowcell_type.lower() == "paired":
347 ispaired = property(_get_ispaired)
349 def _get_filetype(self):
350 return strip_namespace(libraryOntology, self._filetype)
351 filetype = property(_get_filetype)
354 url = urlparse(str(self.filenode.uri))
355 if url.scheme == 'file':
358 errmsg = u"Unsupported scheme {0} for {1}"
359 raise ValueError(errmsg.format(url.scheme, unicode(url)))
360 path = property(_get_path)
363 return "SequenceResult({0},{1},{2})".format(
365 str(self.library_id),
366 str(self.flowcell_id))