htsworkflow/pipelines/summary.py

   1 """
   2 Analyze the Summary.htm file produced by GERALD
   3 """
   4 import types
   5
   6 from htsworkflow.pipelines.runfolder import ElementTree
   7 from htsworkflow.util.ethelp import indent, flatten
   8
   9 class Summary(object):
  10     """
  11     Extract some useful information from the Summary.htm file
  12     """
  13     XML_VERSION = 2
  14     SUMMARY = 'Summary'
  15
  16     class LaneResultSummary(object):
  17         """
  18         Parse the LaneResultSummary table out of Summary.htm
  19         Mostly for the cluster number
  20         """
  21         LANE_RESULT_SUMMARY = 'LaneResultSummary'
  22         TAGS = {
  23           'LaneYield': 'lane_yield',
  24           'Cluster': 'cluster', # Raw
  25           'ClusterPF': 'cluster_pass_filter',
  26           'AverageFirstCycleIntensity': 'average_first_cycle_intensity',
  27           'PercentIntensityAfter20Cycles': 'percent_intensity_after_20_cycles',
  28           'PercentPassFilterClusters': 'percent_pass_filter_clusters',
  29           'PercentPassFilterAlign': 'percent_pass_filter_align',
  30           'AverageAlignmentScore': 'average_alignment_score',
  31           'PercentErrorRate': 'percent_error_rate'
  32         }
  33
  34         def __init__(self, html=None, xml=None):
  35             self.lane = None
  36             self.lane_yield = None
  37             self.cluster = None
  38             self.cluster_pass_filter = None
  39             self.average_first_cycle_intensity = None
  40             self.percent_intensity_after_20_cycles = None
  41             self.percent_pass_filter_clusters = None
  42             self.percent_pass_filter_align = None
  43             self.average_alignment_score = None
  44             self.percent_error_rate = None
  45
  46             if html is not None:
  47                 self.set_elements_from_html(html)
  48             if xml is not None:
  49                 self.set_elements(xml)
  50
  51         def set_elements_from_html(self, data):
  52             if not len(data) in (8,10):
  53                 raise RuntimeError("Summary.htm file format changed")
  54
  55             # same in pre-0.3.0 Summary file and 0.3 summary file
  56             self.lane = data[0]
  57
  58             if len(data) == 8:
  59                 parsed_data = [ parse_mean_range(x) for x in data[1:] ]
  60                 # this is the < 0.3 Pipeline version
  61                 self.cluster = parsed_data[0]
  62                 self.average_first_cycle_intensity = parsed_data[1]
  63                 self.percent_intensity_after_20_cycles = parsed_data[2]
  64                 self.percent_pass_filter_clusters = parsed_data[3]
  65                 self.percent_pass_filter_align = parsed_data[4]
  66                 self.average_alignment_score = parsed_data[5]
  67                 self.percent_error_rate = parsed_data[6]
  68             elif len(data) == 10:
  69                 parsed_data = [ parse_mean_range(x) for x in data[2:] ]
  70                 # this is the >= 0.3 summary file
  71                 self.lane_yield = data[1]
  72                 self.cluster = parsed_data[0]
  73                 self.cluster_pass_filter = parsed_data[1]
  74                 self.average_first_cycle_intensity = parsed_data[2]
  75                 self.percent_intensity_after_20_cycles = parsed_data[3]
  76                 self.percent_pass_filter_clusters = parsed_data[4]
  77                 self.percent_pass_filter_align = parsed_data[5]
  78                 self.average_alignment_score = parsed_data[6]
  79                 self.percent_error_rate = parsed_data[7]
  80
  81         def get_elements(self):
  82             lane_result = ElementTree.Element(
  83                             Summary.LaneResultSummary.LANE_RESULT_SUMMARY,
  84                             {'lane': self.lane})
  85             for tag, variable_name in Summary.LaneResultSummary.TAGS.items():
  86                 value = getattr(self, variable_name)
  87                 if value is None:
  88                     continue
  89                 # it looks like a sequence
  90                 elif type(value) in (types.TupleType, types.ListType):
  91                     element = make_mean_range_element(
  92                       lane_result,
  93                       tag,
  94                       *value
  95                     )
  96                 else:
  97                     element = ElementTree.SubElement(lane_result, tag)
  98                     element.text = value
  99             return lane_result
 100
 101         def set_elements(self, tree):
 102             if tree.tag != Summary.LaneResultSummary.LANE_RESULT_SUMMARY:
 103                 raise ValueError('Expected %s' % (
 104                         Summary.LaneResultSummary.LANE_RESULT_SUMMARY))
 105             self.lane = tree.attrib['lane']
 106             tags = Summary.LaneResultSummary.TAGS
 107             for element in list(tree):
 108                 try:
 109                     variable_name = tags[element.tag]
 110                     setattr(self, variable_name,
 111                             parse_summary_element(element))
 112                 except KeyError, e:
 113                     logging.warn('Unrecognized tag %s' % (element.tag,))
 114
 115     def __init__(self, filename=None, xml=None):
 116         self.lane_results = {}
 117
 118         if filename is not None:
 119             self._extract_lane_results(filename)
 120         if xml is not None:
 121             self.set_elements(xml)
 122
 123     def __getitem__(self, key):
 124         return self.lane_results[key]
 125
 126     def __len__(self):
 127         return len(self.lane_results)
 128
 129     def keys(self):
 130         return self.lane_results.keys()
 131
 132     def values(self):
 133         return self.lane_results.values()
 134
 135     def items(self):
 136         return self.lane_results.items()
 137
 138     def _flattened_row(self, row):
 139         """
 140         flatten the children of a <tr>...</tr>
 141         """
 142         return [flatten(x) for x in row.getchildren() ]
 143
 144     def _parse_table(self, table):
 145         """
 146         assumes the first line is the header of a table,
 147         and that the remaining rows are data
 148         """
 149         rows = table.getchildren()
 150         data = []
 151         for r in rows:
 152             data.append(self._flattened_row(r))
 153         return data
 154
 155     def _extract_named_tables(self, pathname):
 156         """
 157         extract all the 'named' tables from a Summary.htm file
 158         and return as a dictionary
 159
 160         Named tables are <h2>...</h2><table>...</table> pairs
 161         The contents of the h2 tag is considered to the name
 162         of the table.
 163         """
 164         # tree = ElementTree.parse(pathname).getroot()
 165         # hack for 1.1rc1, this should be removed when possible.
 166         file_body = open(pathname).read()
 167         file_body = file_body.replace('CHASTITY<=', 'CHASTITY&lt;=')
 168         tree = ElementTree.fromstring(file_body)
 169         body = tree.find('body')
 170         tables = {}
 171         for i in range(len(body)):
 172             if body[i].tag == 'h2' and body[i+1].tag == 'table':
 173                 # we have an interesting table
 174                 name = flatten(body[i])
 175                 table = body[i+1]
 176                 data = self._parse_table(table)
 177                 tables[name] = data
 178         return tables
 179
 180     def _extract_lane_results(self, pathname):
 181         """
 182         extract the Lane Results Summary table
 183         """
 184
 185         tables = self._extract_named_tables(pathname)
 186
 187         # parse lane result summary
 188         lane_summary = tables['Lane Results Summary']
 189         # this is version 1 of the summary file
 190         if len(lane_summary[-1]) == 8:
 191             # strip header
 192             headers = lane_summary[0]
 193             # grab the lane by lane data
 194             lane_summary = lane_summary[1:]
 195
 196         # this is version 2 of the summary file
 197         if len(lane_summary[-1]) == 10:
 198             # lane_summary[0] is a different less specific header row
 199             headers = lane_summary[1]
 200             lane_summary = lane_summary[2:10]
 201             # after the last lane, there's a set of chip wide averages
 202
 203         for r in lane_summary:
 204             lrs = Summary.LaneResultSummary(html=r)
 205             self.lane_results[lrs.lane] = lrs
 206
 207     def get_elements(self):
 208         summary = ElementTree.Element(Summary.SUMMARY,
 209                                       {'version': unicode(Summary.XML_VERSION)})
 210         for lane in self.lane_results.values():
 211             summary.append(lane.get_elements())
 212         return summary
 213
 214     def set_elements(self, tree):
 215         if tree.tag != Summary.SUMMARY:
 216             return ValueError("Expected %s" % (Summary.SUMMARY,))
 217         xml_version = int(tree.attrib.get('version', 0))
 218         if xml_version > Summary.XML_VERSION:
 219             logging.warn('Summary XML tree is a higher version than this class')
 220         for element in list(tree):
 221             lrs = Summary.LaneResultSummary()
 222             lrs.set_elements(element)
 223             self.lane_results[lrs.lane] = lrs
 224
 225     def dump(self):
 226         """
 227         Debugging function, report current object
 228         """
 229         pass
 230
 231 def tonumber(v):
 232     """
 233     Convert a value to int if its an int otherwise a float.
 234     """
 235     try:
 236         v = int(v)
 237     except ValueError, e:
 238         v = float(v)
 239     return v
 240
 241 def parse_mean_range(value):
 242     """
 243     Parse values like 123 +/- 4.5
 244     """
 245     if value.strip() == 'unknown':
 246         return 0, 0
 247
 248     average, pm, deviation = value.split()
 249     if pm != '+/-':
 250         raise RuntimeError("Summary.htm file format changed")
 251     return tonumber(average), tonumber(deviation)
 252
 253 def make_mean_range_element(parent, name, mean, deviation):
 254     """
 255     Make an ElementTree subelement <Name mean='mean', deviation='deviation'/>
 256     """
 257     element = ElementTree.SubElement(parent, name,
 258                                      { 'mean': unicode(mean),
 259                                        'deviation': unicode(deviation)})
 260     return element
 261
 262 def parse_mean_range_element(element):
 263     """
 264     Grab mean/deviation out of element
 265     """
 266     return (tonumber(element.attrib['mean']),
 267             tonumber(element.attrib['deviation']))
 268
 269 def parse_summary_element(element):
 270     """
 271     Determine if we have a simple element or a mean/deviation element
 272     """
 273     if len(element.attrib) > 0:
 274         return parse_mean_range_element(element)
 275     else:
 276         return element.text