2 Provide access to information stored in the GERALD directory.
4 from datetime import datetime, date
12 from htsworkflow.pipeline.runfolder import \
17 from htsworkflow.util.ethelp import indent, flatten
18 from htsworkflow.util.opener import autoopen
22 Capture meaning out of the GERALD directory
26 RUN_PARAMETERS='RunParameters'
29 class LaneParameters(object):
31 Make it easy to access elements of LaneSpecificRunParameters from python
33 def __init__(self, gerald, key):
37 def __get_attribute(self, xml_tag):
38 subtree = self._gerald.tree.find('LaneSpecificRunParameters')
39 container = subtree.find(xml_tag)
42 if len(container.getchildren()) > LANES_PER_FLOWCELL:
43 raise RuntimeError('GERALD config.xml file changed')
44 lanes = [x.tag.split('_')[1] for x in container.getchildren()]
45 index = lanes.index(self._key)
46 element = container[index]
48 def _get_analysis(self):
49 return self.__get_attribute('ANALYSIS')
50 analysis = property(_get_analysis)
52 def _get_eland_genome(self):
53 genome = self.__get_attribute('ELAND_GENOME')
54 # default to the chipwide parameters if there isn't an
55 # entry in the lane specific paramaters
57 subtree = self._gerald.tree.find('ChipWideRunParameters')
58 container = subtree.find('ELAND_GENOME')
59 genome = container.text
61 eland_genome = property(_get_eland_genome)
63 def _get_read_length(self):
64 return self.__get_attribute('READ_LENGTH')
65 read_length = property(_get_read_length)
67 def _get_use_bases(self):
68 return self.__get_attribute('USE_BASES')
69 use_bases = property(_get_use_bases)
71 class LaneSpecificRunParameters(object):
73 Provide access to LaneSpecificRunParameters
75 def __init__(self, gerald):
78 def __getitem__(self, key):
79 return Gerald.LaneParameters(self._gerald, key)
81 if self._keys is None:
82 tree = self._gerald.tree
83 analysis = tree.find('LaneSpecificRunParameters/ANALYSIS')
84 # according to the pipeline specs I think their fields
85 # are sampleName_laneID, with sampleName defaulting to s
86 # since laneIDs are constant lets just try using
88 self._keys = [ x.tag.split('_')[1] for x in analysis]
91 return [ self[x] for x in self.keys() ]
93 return zip(self.keys(), self.values())
95 return len(self.keys())
97 def __init__(self, xml=None):
101 # parse lane parameters out of the config.xml file
102 self.lanes = Gerald.LaneSpecificRunParameters(self)
105 self.eland_results = None
108 self.set_elements(xml)
111 if self.tree is None:
112 return datetime.today()
113 timestamp = self.tree.findtext('ChipWideRunParameters/TIME_STAMP')
114 epochstamp = time.mktime(time.strptime(timestamp, '%c'))
115 return datetime.fromtimestamp(epochstamp)
116 date = property(_get_date)
119 return time.mktime(self.date.timetuple())
120 time = property(_get_time, doc='return run time as seconds since epoch')
122 def _get_version(self):
123 if self.tree is None:
125 return self.tree.findtext('ChipWideRunParameters/SOFTWARE_VERSION')
126 version = property(_get_version)
130 Debugging function, report current object
132 print 'Gerald version:', self.version
133 print 'Gerald run date:', self.date
134 print 'Gerald config.xml:', self.tree
137 def get_elements(self):
138 if self.tree is None or self.summary is None:
141 gerald = ElementTree.Element(Gerald.GERALD,
142 {'version': unicode(Gerald.XML_VERSION)})
143 gerald.append(self.tree)
144 gerald.append(self.summary.get_elements())
145 if self.eland_results:
146 gerald.append(self.eland_results.get_elements())
149 def set_elements(self, tree):
150 if tree.tag != Gerald.GERALD:
151 raise ValueError('exptected GERALD')
152 xml_version = int(tree.attrib.get('version', 0))
153 if xml_version > Gerald.XML_VERSION:
154 logging.warn('XML tree is a higher version than this class')
155 for element in list(tree):
156 tag = element.tag.lower()
157 if tag == Gerald.RUN_PARAMETERS.lower():
159 elif tag == Gerald.SUMMARY.lower():
160 self.summary = Summary(xml=element)
161 elif tag == ELAND.ELAND.lower():
162 self.eland_results = ELAND(xml=element)
164 logging.warn("Unrecognized tag %s" % (element.tag,))
167 def gerald(pathname):
169 g.pathname = pathname
170 path, name = os.path.split(pathname)
171 config_pathname = os.path.join(pathname, 'config.xml')
172 g.tree = ElementTree.parse(config_pathname).getroot()
174 # parse Summary.htm file
175 summary_pathname = os.path.join(pathname, 'Summary.htm')
176 g.summary = Summary(summary_pathname)
178 g.eland_results = eland(g.pathname, g)
183 Convert a value to int if its an int otherwise a float.
187 except ValueError, e:
191 def parse_mean_range(value):
193 Parse values like 123 +/- 4.5
195 if value.strip() == 'unknown':
198 average, pm, deviation = value.split()
200 raise RuntimeError("Summary.htm file format changed")
201 return tonumber(average), tonumber(deviation)
203 def make_mean_range_element(parent, name, mean, deviation):
205 Make an ElementTree subelement <Name mean='mean', deviation='deviation'/>
207 element = ElementTree.SubElement(parent, name,
208 { 'mean': unicode(mean),
209 'deviation': unicode(deviation)})
212 def parse_mean_range_element(element):
214 Grab mean/deviation out of element
216 return (tonumber(element.attrib['mean']),
217 tonumber(element.attrib['deviation']))
219 def parse_summary_element(element):
221 Determine if we have a simple element or a mean/deviation element
223 if len(element.attrib) > 0:
224 return parse_mean_range_element(element)
228 class Summary(object):
230 Extract some useful information from the Summary.htm file
235 class LaneResultSummary(object):
237 Parse the LaneResultSummary table out of Summary.htm
238 Mostly for the cluster number
240 LANE_RESULT_SUMMARY = 'LaneResultSummary'
242 'LaneYield': 'lane_yield',
243 'Cluster': 'cluster', # Raw
244 'ClusterPF': 'cluster_pass_filter',
245 'AverageFirstCycleIntensity': 'average_first_cycle_intensity',
246 'PercentIntensityAfter20Cycles': 'percent_intensity_after_20_cycles',
247 'PercentPassFilterClusters': 'percent_pass_filter_clusters',
248 'PercentPassFilterAlign': 'percent_pass_filter_align',
249 'AverageAlignmentScore': 'average_alignment_score',
250 'PercentErrorRate': 'percent_error_rate'
253 def __init__(self, html=None, xml=None):
255 self.lane_yield = None
257 self.cluster_pass_filter = None
258 self.average_first_cycle_intensity = None
259 self.percent_intensity_after_20_cycles = None
260 self.percent_pass_filter_clusters = None
261 self.percent_pass_filter_align = None
262 self.average_alignment_score = None
263 self.percent_error_rate = None
266 self.set_elements_from_html(html)
268 self.set_elements(xml)
270 def set_elements_from_html(self, data):
271 if not len(data) in (8,10):
272 raise RuntimeError("Summary.htm file format changed")
274 # same in pre-0.3.0 Summary file and 0.3 summary file
278 parsed_data = [ parse_mean_range(x) for x in data[1:] ]
279 # this is the < 0.3 Pipeline version
280 self.cluster = parsed_data[0]
281 self.average_first_cycle_intensity = parsed_data[1]
282 self.percent_intensity_after_20_cycles = parsed_data[2]
283 self.percent_pass_filter_clusters = parsed_data[3]
284 self.percent_pass_filter_align = parsed_data[4]
285 self.average_alignment_score = parsed_data[5]
286 self.percent_error_rate = parsed_data[6]
287 elif len(data) == 10:
288 parsed_data = [ parse_mean_range(x) for x in data[2:] ]
289 # this is the >= 0.3 summary file
290 self.lane_yield = data[1]
291 self.cluster = parsed_data[0]
292 self.cluster_pass_filter = parsed_data[1]
293 self.average_first_cycle_intensity = parsed_data[2]
294 self.percent_intensity_after_20_cycles = parsed_data[3]
295 self.percent_pass_filter_clusters = parsed_data[4]
296 self.percent_pass_filter_align = parsed_data[5]
297 self.average_alignment_score = parsed_data[6]
298 self.percent_error_rate = parsed_data[7]
300 def get_elements(self):
301 lane_result = ElementTree.Element(
302 Summary.LaneResultSummary.LANE_RESULT_SUMMARY,
304 for tag, variable_name in Summary.LaneResultSummary.TAGS.items():
305 value = getattr(self, variable_name)
308 # it looks like a sequence
309 elif type(value) in (types.TupleType, types.ListType):
310 element = make_mean_range_element(
316 element = ElementTree.SubElement(lane_result, tag)
320 def set_elements(self, tree):
321 if tree.tag != Summary.LaneResultSummary.LANE_RESULT_SUMMARY:
322 raise ValueError('Expected %s' % (
323 Summary.LaneResultSummary.LANE_RESULT_SUMMARY))
324 self.lane = tree.attrib['lane']
325 tags = Summary.LaneResultSummary.TAGS
326 for element in list(tree):
328 variable_name = tags[element.tag]
329 setattr(self, variable_name,
330 parse_summary_element(element))
332 logging.warn('Unrecognized tag %s' % (element.tag,))
334 def __init__(self, filename=None, xml=None):
335 self.lane_results = {}
337 if filename is not None:
338 self._extract_lane_results(filename)
340 self.set_elements(xml)
342 def __getitem__(self, key):
343 return self.lane_results[key]
346 return len(self.lane_results)
349 return self.lane_results.keys()
352 return self.lane_results.values()
355 return self.lane_results.items()
357 def _flattened_row(self, row):
359 flatten the children of a <tr>...</tr>
361 return [flatten(x) for x in row.getchildren() ]
363 def _parse_table(self, table):
365 assumes the first line is the header of a table,
366 and that the remaining rows are data
368 rows = table.getchildren()
371 data.append(self._flattened_row(r))
374 def _extract_named_tables(self, pathname):
376 extract all the 'named' tables from a Summary.htm file
377 and return as a dictionary
379 Named tables are <h2>...</h2><table>...</table> pairs
380 The contents of the h2 tag is considered to the name
383 tree = ElementTree.parse(pathname).getroot()
384 body = tree.find('body')
386 for i in range(len(body)):
387 if body[i].tag == 'h2' and body[i+1].tag == 'table':
388 # we have an interesting table
389 name = flatten(body[i])
391 data = self._parse_table(table)
395 def _extract_lane_results(self, pathname):
397 extract the Lane Results Summary table
400 tables = self._extract_named_tables(pathname)
402 # parse lane result summary
403 lane_summary = tables['Lane Results Summary']
404 # this is version 1 of the summary file
405 if len(lane_summary[-1]) == 8:
407 headers = lane_summary[0]
408 # grab the lane by lane data
409 lane_summary = lane_summary[1:]
411 # this is version 2 of the summary file
412 if len(lane_summary[-1]) == 10:
413 # lane_summary[0] is a different less specific header row
414 headers = lane_summary[1]
415 lane_summary = lane_summary[2:10]
416 # after the last lane, there's a set of chip wide averages
418 for r in lane_summary:
419 lrs = Summary.LaneResultSummary(html=r)
420 self.lane_results[lrs.lane] = lrs
422 def get_elements(self):
423 summary = ElementTree.Element(Summary.SUMMARY,
424 {'version': unicode(Summary.XML_VERSION)})
425 for lane in self.lane_results.values():
426 summary.append(lane.get_elements())
429 def set_elements(self, tree):
430 if tree.tag != Summary.SUMMARY:
431 return ValueError("Expected %s" % (Summary.SUMMARY,))
432 xml_version = int(tree.attrib.get('version', 0))
433 if xml_version > Summary.XML_VERSION:
434 logging.warn('Summary XML tree is a higher version than this class')
435 for element in list(tree):
436 lrs = Summary.LaneResultSummary()
437 lrs.set_elements(element)
438 self.lane_results[lrs.lane] = lrs
442 Debugging function, report current object
447 def build_genome_fasta_map(genome_dir):
448 # build fasta to fasta file map
449 genome = genome_dir.split(os.path.sep)[-1]
451 for vld_file in glob(os.path.join(genome_dir, '*.vld')):
453 if os.path.islink(vld_file):
455 vld_file = os.path.realpath(vld_file)
456 path, vld_name = os.path.split(vld_file)
457 name, ext = os.path.splitext(vld_name)
459 fasta_map[name] = name
461 fasta_map[name] = os.path.join(genome, name)
464 class ElandLane(object):
466 Process an eland result file
470 SAMPLE_NAME = 'SampleName'
472 GENOME_MAP = 'GenomeMap'
473 GENOME_ITEM = 'GenomeItem'
474 MAPPED_READS = 'MappedReads'
475 MAPPED_ITEM = 'MappedItem'
476 MATCH_CODES = 'MatchCodes'
480 def __init__(self, pathname=None, genome_map=None, xml=None):
481 self.pathname = pathname
482 self._sample_name = None
485 self._mapped_reads = None
486 self._match_codes = None
487 if genome_map is None:
489 self.genome_map = genome_map
492 self.set_elements(xml)
496 Actually read the file and actually count the reads
498 # can't do anything if we don't have a file to process
499 if self.pathname is None:
502 if os.stat(self.pathname)[stat.ST_SIZE] == 0:
503 raise RuntimeError("Eland isn't done, try again later.")
508 match_codes = {'NM':0, 'QC':0, 'RM':0,
509 'U0':0, 'U1':0, 'U2':0,
510 'R0':0, 'R1':0, 'R2':0,
512 for line in autoopen(self.pathname,'r'):
514 fields = line.split()
516 # match_codes[code] = match_codes.setdefault(code, 0) + 1
517 # the QC/NM etc codes are in the 3rd field and always present
518 match_codes[fields[2]] += 1
519 # ignore lines that don't have a fasta filename
522 fasta = self.genome_map.get(fields[6], fields[6])
523 mapped_reads[fasta] = mapped_reads.setdefault(fasta, 0) + 1
524 self._match_codes = match_codes
525 self._mapped_reads = mapped_reads
528 def _update_name(self):
529 # extract the sample name
530 if self.pathname is None:
533 path, name = os.path.split(self.pathname)
534 split_name = name.split('_')
535 self._sample_name = split_name[0]
536 self._lane_id = split_name[1]
538 def _get_sample_name(self):
539 if self._sample_name is None:
541 return self._sample_name
542 sample_name = property(_get_sample_name)
544 def _get_lane_id(self):
545 if self._lane_id is None:
548 lane_id = property(_get_lane_id)
550 def _get_reads(self):
551 if self._reads is None:
554 reads = property(_get_reads)
556 def _get_mapped_reads(self):
557 if self._mapped_reads is None:
559 return self._mapped_reads
560 mapped_reads = property(_get_mapped_reads)
562 def _get_match_codes(self):
563 if self._match_codes is None:
565 return self._match_codes
566 match_codes = property(_get_match_codes)
568 def get_elements(self):
569 lane = ElementTree.Element(ElandLane.LANE,
571 unicode(ElandLane.XML_VERSION)})
572 sample_tag = ElementTree.SubElement(lane, ElandLane.SAMPLE_NAME)
573 sample_tag.text = self.sample_name
574 lane_tag = ElementTree.SubElement(lane, ElandLane.LANE_ID)
575 lane_tag.text = self.lane_id
576 genome_map = ElementTree.SubElement(lane, ElandLane.GENOME_MAP)
577 for k, v in self.genome_map.items():
578 item = ElementTree.SubElement(
579 genome_map, ElandLane.GENOME_ITEM,
580 {'name':k, 'value':unicode(v)})
581 mapped_reads = ElementTree.SubElement(lane, ElandLane.MAPPED_READS)
582 for k, v in self.mapped_reads.items():
583 item = ElementTree.SubElement(
584 mapped_reads, ElandLane.MAPPED_ITEM,
585 {'name':k, 'value':unicode(v)})
586 match_codes = ElementTree.SubElement(lane, ElandLane.MATCH_CODES)
587 for k, v in self.match_codes.items():
588 item = ElementTree.SubElement(
589 match_codes, ElandLane.MATCH_ITEM,
590 {'name':k, 'value':unicode(v)})
591 reads = ElementTree.SubElement(lane, ElandLane.READS)
592 reads.text = unicode(self.reads)
596 def set_elements(self, tree):
597 if tree.tag != ElandLane.LANE:
598 raise ValueError('Exptecting %s' % (ElandLane.LANE,))
601 self._mapped_reads = {}
602 self._match_codes = {}
605 tag = element.tag.lower()
606 if tag == ElandLane.SAMPLE_NAME.lower():
607 self._sample_name = element.text
608 elif tag == ElandLane.LANE_ID.lower():
609 self._lane_id = element.text
610 elif tag == ElandLane.GENOME_MAP.lower():
611 for child in element:
612 name = child.attrib['name']
613 value = child.attrib['value']
614 self.genome_map[name] = value
615 elif tag == ElandLane.MAPPED_READS.lower():
616 for child in element:
617 name = child.attrib['name']
618 value = child.attrib['value']
619 self._mapped_reads[name] = int(value)
620 elif tag == ElandLane.MATCH_CODES.lower():
621 for child in element:
622 name = child.attrib['name']
623 value = int(child.attrib['value'])
624 self._match_codes[name] = value
625 elif tag == ElandLane.READS.lower():
626 self._reads = int(element.text)
628 logging.warn("ElandLane unrecognized tag %s" % (element.tag,))
630 def extract_eland_sequence(instream, outstream, start, end):
632 Extract a chunk of sequence out of an eland file
634 for line in instream:
635 record = line.split()
637 result = [record[0], record[1][start:end]]
639 result = [record[0][start:end]]
640 outstream.write("\t".join(result))
641 outstream.write(os.linesep)
645 Summarize information from eland files
649 ELAND = 'ElandCollection'
653 def __init__(self, xml=None):
654 # we need information from the gerald config.xml
658 self.set_elements(xml)
661 return len(self.results)
664 return self.results.keys()
667 return self.results.values()
670 return self.results.items()
672 def __getitem__(self, key):
673 return self.results[key]
675 def get_elements(self):
676 root = ElementTree.Element(ELAND.ELAND,
677 {'version': unicode(ELAND.XML_VERSION)})
678 for lane_id, lane in self.results.items():
679 eland_lane = lane.get_elements()
680 eland_lane.attrib[ELAND.LANE_ID] = unicode(lane_id)
681 root.append(eland_lane)
684 def set_elements(self, tree):
685 if tree.tag.lower() != ELAND.ELAND.lower():
686 raise ValueError('Expecting %s', ELAND.ELAND)
687 for element in list(tree):
688 lane_id = element.attrib[ELAND.LANE_ID]
689 lane = ElandLane(xml=element)
690 self.results[lane_id] = lane
692 def eland(basedir, gerald=None, genome_maps=None):
695 file_list = glob(os.path.join(basedir, "*_eland_result.txt"))
696 if len(file_list) == 0:
697 # lets handle compressed eland files too
698 file_list = glob(os.path.join(basedir, "*_eland_result.txt.bz2"))
700 for pathname in file_list:
701 # yes the lane_id is also being computed in ElandLane._update
702 # I didn't want to clutter up my constructor
703 # but I needed to persist the sample_name/lane_id for
704 # runfolder summary_report
705 path, name = os.path.split(pathname)
706 split_name = name.split('_')
707 lane_id = split_name[1]
709 if genome_maps is not None:
710 genome_map = genome_maps[lane_id]
711 elif gerald is not None:
712 genome_dir = gerald.lanes[lane_id].eland_genome
713 genome_map = build_genome_fasta_map(genome_dir)
717 eland_result = ElandLane(pathname, genome_map)
718 e.results[lane_id] = eland_result