Updating endure parser to read new csv output format

scripts/slave/process_log_utils.py

 # Copyright (c) 2012 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 """Parser and evaluator for performance tests.
 
 Several performance tests have complicated log output, this module is intended
 to help buildsteps parse these logs and identify if tests had anomalies.
 
 
@@ skipping 820 matching lines @@
     graphs = {}
     for name, graph in self._graphs.iteritems():
       graphs[name] = {'name': name,
                       'important': graph.IsImportant(),
                       'units': graph.units}
     self._output[GRAPH_LIST] = json.dumps(graphs).split('\n')
 
 
 class GraphingEndureLogProcessor(GraphingLogProcessor):
   """Handles additional processing for Chrome Endure data."""
-  ENDURE_HEADER_LINE_REGEX = re.compile(r'^url,')
-  ENDURE_RESULT_LINE_REGEX = re.compile(r'^http')
+  ENDURE_HEADER_LINE_REGEX = re.compile(r'^url,|^page_name,')
+  ENDURE_RESULT_LINE_REGEX = re.compile(r'^http|^endure_')
   ENDURE_FIELD_NAME_REGEX = re.compile(
       r'(?P<TRACE>.*)_(?P<COORDINATE>[XY]) \((?P<UNITS>.*)\)')
 
   def __init__(self, *args, **kwargs):
     GraphingLogProcessor.__init__(self, *args, **kwargs)
     # A dictionary of trace names to EndureGraph objects, which is generated
     # when parsing the header line.  This template will be used as the default
     # EndureGraph when parsing results lines.
     self._graph_template = {}
 
@@ skipping 40 matching lines @@
   def _ProcessEndureResultLine(self, line):
     """Parse each regular results line in the csv input.
 
     A typical results line will look like this:
     https://www.google.com/calendar/,4,446,4,2847
     """
     assert self._graph_template
 
     values = csv.reader([line]).next()
 
-    # Assume url is the first column.
-    test_name = self.url_as_file_safe_name(values[0])
+    # Assume test name is the first column.
+    test_name = self.str_as_file_safe_name(values[0])
 
     # Iterate over all trace names discovered from the header.
     for trace_name in self._graph_template:
       graph_name = test_name + '-' + trace_name
 
       # The default EndureGraph is copied from the template, which already
       # contains all units and index information.
       graph = self._graphs.get(
           graph_name,
           copy.deepcopy(self._graph_template[trace_name]))
 
       trace = graph.traces.get(trace_name, EndureTrace())
       trace.values.append([values[graph.csv_index_x],
                            values[graph.csv_index_y]])
 
       graph.traces[trace_name] = trace
       self._graphs[graph_name] = graph
 
   @staticmethod
-  def url_as_file_safe_name(url):
-    # Just replace all special characters in the url with underscore.
-    return re.sub('[^a-zA-Z0-9]', '_', url)
+  def str_as_file_safe_name(string):
+    # Just replace all special characters in the string with underscores.
+    return re.sub('[^a-zA-Z0-9]', '_', string)
 
   def _FinalizeProcessing(self):
     self.__CreateSummaryOutput()
     self.__GenerateGraphInfo()
 
   def __BuildSummaryJSON(self, graph):
     """Sorts the traces and returns a summary JSON encoding of the graph.
 
     Although JS objects are not ordered, according to the spec, in practice
     everyone iterates in order, since not doing so is a compatibility problem.
@@ skipping 156 matching lines @@
     for page in self._page_list:
       times = page_times[page]
       mean, stddev = chromium_utils.FilteredMeanAndStandardDeviation(times)
       file_data.append('%s (%s+/-%s): %s' % (page,
                                              FormatFloat(mean),
                                              FormatFloat(stddev),
                                              JoinWithSpacesAndNewLine(times)))
 
     filename = '%s_%s.dat' % (self._revision, trace_name)
     return {filename: file_data}