AFDO-Bisect: Write a (very) basic AFDO prof parser

This CL introduces a script which does very basic parsing of AFDO
text-based profiles which associates each top-level function with its
associated profile data.

BUG=None
TEST=The added test is successful.

Change-Id: I6c8d3b3b1049b81a5890cc9699e2798dda641eb9
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/third_party/toolchain-utils/+/1679371
Reviewed-by: Caroline Tice <cmtice@chromium.org>
Reviewed-by: George Burgess <gbiv@chromium.org>
Tested-by: Emma Vukelj <emmavukelj@google.com>

2 files changed, 116 insertions, 0 deletions

diff --git a/afdo_tools/bisection/afdo_parse.py b/afdo_tools/bisection/afdo_parse.py
new file mode 100755
index 00000000..b9da2174
--- /dev/null
+++ b/afdo_tools/bisection/afdo_parse.py
@@ -0,0 +1,61 @@
+#!/usr/bin/env python2
+# -*- coding: utf-8 -*-
+# Copyright 2019 The Chromium OS Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+"""Performs basic parsing of an AFDO text-based profile.
+
+This short script performs very basic parsing of an AFDO text-based profile
+into a dictionary which associates each top-level function with its profile
+data (as plain text), and these results are dumped to a pickle file.
+"""
+
+from __future__ import print_function
+from absl import app
+from absl import flags
+
+import json
+import pprint
+
+flags.DEFINE_string('afdo_text', None, 'AFDO text-based profile to be parsed')
+flags.DEFINE_string('output_file', None, 'File to write JSON results to')
+FLAGS = flags.FLAGS
+
+
+def parse_afdo(f):
+  """Performs basic parsing of an AFDO text-based profile.
+
+  This parsing expects an input file of the form generated by bin/llvm-profdata
+  (within an LLVM build).
+  """
+  results = {}
+  curr_func = None
+  curr_data = []
+  for line in f:
+    if not line.startswith(' '):
+      if curr_func:
+        results[curr_func] = ''.join(curr_data)
+        curr_data = []
+      curr_func = line.split(':')[0].strip()
+    else:
+      curr_data.append(line)
+
+  if curr_func:
+    results[curr_func] = ''.join(curr_data)
+  return results
+
+
+def main(_):
+  with open(FLAGS.afdo_text) as f:
+    results = parse_afdo(f)
+    if FLAGS.output_file:
+      with open(FLAGS.output_file, 'wb') as f_out:
+        json.dump(results, f_out, indent=2)
+    else:
+      pprint.pprint(results)
+
+
+if __name__ == '__main__':
+  flags.mark_flag_as_required('afdo_text')
+  app.run(main)
diff --git a/afdo_tools/bisection/afdo_parse_test.py b/afdo_tools/bisection/afdo_parse_test.py
new file mode 100755
index 00000000..3e17b8f9
--- /dev/null
+++ b/afdo_tools/bisection/afdo_parse_test.py
@@ -0,0 +1,55 @@
+#!/usr/bin/env python2
+# -*- coding: utf-8 -*-
+# Copyright 2019 The Chromium OS Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+"""Tests for afdo_parse."""
+
+from __future__ import print_function
+
+import StringIO
+import unittest
+
+import afdo_parse
+
+
+class SimpleAfdoParseTest(unittest.TestCase):
+  """Test class for AFDO parsing."""
+
+  def test_parse_afdo(self):
+    test_data = StringIO.StringIO('deflate_slow:87460059:3\n'
+                                  ' 3: 24\n'
+                                  ' 14: 54767\n'
+                                  ' 15: 664 fill_window:22\n'
+                                  ' 16: 661\n'
+                                  ' 19: 637\n'
+                                  ' 41: 36692 longest_match:36863\n'
+                                  ' 44: 36692\n'
+                                  ' 44.2: 5861\n'
+                                  ' 46: 13942\n'
+                                  ' 46.1: 14003\n')
+    expected = {
+        'deflate_slow': ' 3: 24\n'
+                        ' 14: 54767\n'
+                        ' 15: 664 fill_window:22\n'
+                        ' 16: 661\n'
+                        ' 19: 637\n'
+                        ' 41: 36692 longest_match:36863\n'
+                        ' 44: 36692\n'
+                        ' 44.2: 5861\n'
+                        ' 46: 13942\n'
+                        ' 46.1: 14003\n'
+    }
+    actual = afdo_parse.parse_afdo(test_data)
+    self.assertEqual(actual, expected)
+    test_data.close()
+
+  def test_parse_empty_afdo(self):
+    expected = {}
+    actual = afdo_parse.parse_afdo('')
+    self.assertEqual(actual, expected)
+
+
+if __name__ == '__main__':
+  unittest.main()