[tools/perf] Implement csv_output result formatter

Bug: 981349
Change-Id: I8faa2aa790327a838978d3f002ad706db70e9d18
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1829337
Commit-Queue: Mikhail Khokhlov <khokhlov@google.com>
Reviewed-by: Juan Antonio Navarro Pérez <perezju@chromium.org>
Cr-Commit-Position: refs/heads/master@{#701167}

tools/perf/core/results_processor/formatters/__init__.py

@@ -2,6 +2,7 @@
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
+from core.results_processor.formatters import csv_output
 from core.results_processor.formatters import histograms_output
 from core.results_processor.formatters import html_output
 from core.results_processor.formatters import json3_output
@@ -11,4 +12,5 @@ FORMATTERS = {
     'json-test-results': json3_output,
     'histograms': histograms_output,
     'html': html_output,
+    'csv': csv_output,
 }

tools/perf/core/results_processor/formatters/csv_output.py

+# Copyright 2019 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.
+
+"""Output formatter for CSV format."""
+
+import collections
+import csv
+import json
+import os
+
+from core.results_processor.formatters import histograms_output
+from py_utils import tempfile_ext
+from tracing.value import histograms_to_csv
+
+
+OUTPUT_FILENAME = 'results.csv'
+
+
+def _ReadCsv(input_stream):
+  dicts = []
+  header = None
+  for row in csv.reader(input_stream):
+    if header is None:
+      header = row
+    elif row:
+      dicts.append(collections.OrderedDict(zip(header, row)))
+  return dicts
+
+
+def _WriteCsv(dicts, output_stream):
+  header = []
+  for d in dicts:
+    for k in d.iterkeys():
+      if k not in header:
+        header.append(k)
+  rows = [header]
+  for d in dicts:
+    rows.append([d.get(k, '') for k in header])
+  csv.writer(output_stream).writerows(rows)
+
+
+def Process(intermediate_results, options):
+  """Process intermediate results and write output in output_dir."""
+  histogram_dicts = histograms_output.Convert(intermediate_results,
+                                              options.results_label)
+
+  with tempfile_ext.NamedTemporaryFile() as hist_file:
+    json.dump(histogram_dicts, hist_file)
+    hist_file.close()
+    vinn_result = histograms_to_csv.HistogramsToCsv(hist_file.name)
+    csv_dicts = _ReadCsv(vinn_result.stdout.splitlines())
+
+  output_file = os.path.join(options.output_dir, OUTPUT_FILENAME)
+  if not options.reset_results and os.path.isfile(output_file):
+    with open(output_file) as input_stream:
+      csv_dicts += _ReadCsv(input_stream)
+
+  with open(output_file, 'w') as output_stream:
+    _WriteCsv(csv_dicts, output_stream)

tools/perf/core/results_processor/processor_test.py

@@ -9,12 +9,14 @@ standalone results processor on them to check that output files are produced
 with their expected contents.
 """
 
+import csv
 import json
 import os
 import shutil
 import tempfile
 import unittest
 
+from core.results_processor.formatters import csv_output
 from core.results_processor.formatters import json3_output
 from core.results_processor.formatters import histograms_output
 from core.results_processor.formatters import html_output
@@ -328,3 +330,121 @@ class ResultsProcessorIntegrationTests(unittest.TestCase):
     diag_values = [list(v) for v in  out_histograms.shared_diagnostics]
     self.assertIn(['label1'], diag_values)
     self.assertIn(['label2'], diag_values)
+
+  def testCsvOutput(self):
+    hist_file = os.path.join(self.output_dir,
+                             histograms_output.HISTOGRAM_DICTS_NAME)
+    test_hist = histogram.Histogram('a', 'ms')
+    test_hist.AddSample(3000)
+    with open(hist_file, 'w') as f:
+      json.dump([test_hist.AsDict()], f)
+
+    self.SerializeIntermediateResults(
+        test_results=[
+            testing.TestResult(
+                'benchmark/story',
+                artifacts={'histogram_dicts.json': testing.Artifact(hist_file)},
+            ),
+        ],
+        diagnostics={
+            'benchmarks': ['benchmark'],
+            'osNames': ['linux'],
+            'documentationUrls': [['documentation', 'url']],
+        },
+    )
+
+    processor.main([
+        '--output-format', 'csv',
+        '--output-dir', self.output_dir,
+        '--intermediate-dir', self.intermediate_dir,
+        '--results-label', 'label',
+    ])
+
+    with open(os.path.join(self.output_dir, csv_output.OUTPUT_FILENAME)) as f:
+      lines = [line for line in f]
+
+    actual = list(zip(*csv.reader(lines)))
+    expected = [
+        ('name', 'a'), ('unit', 'ms'), ('avg', '3000'), ('count', '1'),
+        ('max', '3000'), ('min', '3000'), ('std', '0'), ('sum', '3000'),
+        ('architectures', ''), ('benchmarks', 'benchmark'),
+        ('benchmarkStart', ''), ('bots', ''),
+        ('builds', ''), ('deviceIds', ''), ('displayLabel', 'label'),
+        ('masters', ''), ('memoryAmounts', ''), ('osNames', 'linux'),
+        ('osVersions', ''), ('productVersions', ''),
+        ('stories', ''), ('storysetRepeats', ''),
+        ('traceStart', ''), ('traceUrls', '')
+    ]
+    self.assertEqual(actual, expected)
+
+  def testCsvOutputResetResults(self):
+    hist_file = os.path.join(self.output_dir,
+                             histograms_output.HISTOGRAM_DICTS_NAME)
+    with open(hist_file, 'w') as f:
+      json.dump([histogram.Histogram('a', 'unitless').AsDict()], f)
+
+    self.SerializeIntermediateResults(
+        test_results=[
+            testing.TestResult(
+                'benchmark/story',
+                artifacts={'histogram_dicts.json': testing.Artifact(hist_file)},
+            ),
+        ],
+    )
+
+    processor.main([
+        '--output-format', 'csv',
+        '--output-dir', self.output_dir,
+        '--intermediate-dir', self.intermediate_dir,
+        '--results-label', 'label1',
+    ])
+
+    processor.main([
+        '--output-format', 'csv',
+        '--output-dir', self.output_dir,
+        '--intermediate-dir', self.intermediate_dir,
+        '--results-label', 'label2',
+        '--reset-results',
+    ])
+
+    with open(os.path.join(self.output_dir, csv_output.OUTPUT_FILENAME)) as f:
+      lines = [line for line in f]
+
+    self.assertEqual(len(lines), 2)
+    self.assertIn('label2', lines[1])
+
+  def testCsvOutputAppendResults(self):
+    hist_file = os.path.join(self.output_dir,
+                             histograms_output.HISTOGRAM_DICTS_NAME)
+    with open(hist_file, 'w') as f:
+      json.dump([histogram.Histogram('a', 'unitless').AsDict()], f)
+
+    self.SerializeIntermediateResults(
+        test_results=[
+            testing.TestResult(
+                'benchmark/story',
+                artifacts={'histogram_dicts.json': testing.Artifact(hist_file)},
+            ),
+        ],
+    )
+
+    processor.main([
+        '--output-format', 'csv',
+        '--output-dir', self.output_dir,
+        '--intermediate-dir', self.intermediate_dir,
+        '--results-label', 'label1',
+    ])
+
+    processor.main([
+        '--output-format', 'csv',
+        '--output-dir', self.output_dir,
+        '--intermediate-dir', self.intermediate_dir,
+        '--results-label', 'label2',
+    ])
+
+    with open(os.path.join(self.output_dir, csv_output.OUTPUT_FILENAME)) as f:
+      lines = [line for line in f]
+
+    self.assertEqual(len(lines), 3)
+    self.assertIn('label2', lines[1])
+    self.assertIn('label1', lines[2])