Reland "Refractor manage_render_test_goldens.py and add manage_wpr_archives.py"

This is a reland of 41dddd7d4e

changes: fix logging.info and logging.error syntax issue.

Original change's description:
> Refractor manage_render_test_goldens.py and add manage_wpr_archives.py
>
> R=bsheedy@chromium.org,harringtond@chromium.org
>
> Bug: 1060399
> Change-Id: Ifb06c6a94a51a6b1a2815bf94fa8f5a53dec1db8
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2219525
> Reviewed-by: Haiyang Pan <hypan@google.com>
> Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
> Reviewed-by: Dan H <harringtond@chromium.org>
> Reviewed-by: Brian Sheedy <bsheedy@chromium.org>
> Commit-Queue: Arthur Wang <wuwang@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#774253}

Bug: 1060399
Change-Id: Id5e989cfca591e5f1aacd4f5024d29e894dae919
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2225243
Reviewed-by: Robbie Iannucci <iannucci@chromium.org>
Reviewed-by: Dan H <harringtond@chromium.org>
Reviewed-by: Brian Sheedy <bsheedy@chromium.org>
Commit-Queue: Arthur Wang <wuwang@chromium.org>
Cr-Commit-Position: refs/heads/master@{#774501}

.gitignore

@@ -315,6 +315,9 @@ vs-chromium-project.txt
 # Ignore any Android RenderTest goldens
 **/render_tests/*.png
 
+# Ignore any WPR archive files
+**/wpr_tests/*.wprgo
+
 # Ignore IntelliJ files.
 .idea/
 

.vpython

@@ -112,6 +112,13 @@ wheel: <
   >
 >
 
+# Used by:
+#   chrome/test/data/android/upload_download_utils_test.py
+wheel: <
+  name: "infra/python/wheels/pyfakefs-py2_py3"
+  version: "version:3.7.2"
+>
+
 # Used by:
 #   tools/perf/fetch_benchmark_deps.py
 wheel: <

DEPS

@@ -117,6 +117,10 @@ vars = {
   'checkout_traffic_annotation_tools': 'checkout_configuration != "small"',
   'checkout_instrumented_libraries': 'checkout_linux and checkout_configuration != "small"',
 
+  # By default bot checkouts the WPR archive files only when this
+  # flag is set True.
+  'checkout_wpr_archives': False,
+
   # By default, do not check out WebKit for iOS, as it is not needed unless
   # running against ToT WebKit rather than system WebKit. This can be overridden
   # e.g. with custom_vars.
@@ -4573,6 +4577,16 @@ hooks = [
                 'download',
     ],
   },
+  # Pull down WPR Archive files
+  {
+    'name': 'Fetch WPR archive files',
+    'pattern': '.',
+    'condition': 'checkout_android and (checkout_wpr_archives or checkout_src_internal)',
+    'action': [ 'python',
+                'src/chrome/test/data/android/manage_wpr_archives.py',
+                'download',
+    ],
+  },
   {
     'name': 'Fetch Android AFDO profile',
     'pattern': '.',

chrome/android/feed/core/javatests/src/org/chromium/chrome/browser/feed/wpr_tests/FeedNewTabPageCardInstrumentationTest#launchNTP_withMultipleFeedCardsRendered.wprgo.sha1

@@ -0,0 +1 @@
+0846bac870a018d9ed4820943700b2c978c5ccca

chrome/test/data/android/manage_render_test_goldens.py

@@ -8,11 +8,9 @@
 # render tests or upload any newly generated ones.
 
 import argparse
-import hashlib
-import multiprocessing
 import os
-import subprocess
-
+from upload_download_utils import download
+from upload_download_utils import upload
 
 STORAGE_BUCKET = 'chromium-android-render-test-goldens'
 THIS_DIR = os.path.abspath(os.path.dirname(__file__))
@@ -46,19 +44,8 @@ ALLOWED_DEVICE_SDK_COMBINATIONS = [
 ]
 
 
-# Assume a quad core if we can't get the actual core count.
-try:
-  THREAD_COUNT = multiprocessing.cpu_count()
-  # cpu_count only gets the physical core count. There doesn't appear to be a
-  # simple way of determining whether a CPU supports simultaneous multithreading
-  # in Python, so assume that anything with 6 or more cores supports it.
-  if THREAD_COUNT >= 6:
-    THREAD_COUNT *= 2
-except NotImplementedError:
-  THREAD_COUNT = 4
-
-
-def is_file_of_interest(f):
+def _is_file_of_interest(f):
+  """Filter through png files with right device sdk combo in the names."""
   if not f.endswith('.png'):
     return False
   for combo in ALLOWED_DEVICE_SDK_COMBINATIONS:
@@ -67,81 +54,22 @@ def is_file_of_interest(f):
   return False
 
 
-def download(directory):
-  # If someone removes a SHA1 file, we want to remove the associated PNG file
-  # the next time images are updated.
-  images_to_delete = []
-  for f in os.listdir(directory):
-    if not is_file_of_interest(f):
-      continue
-    sha1_path = os.path.join(directory, f + '.sha1')
-    if not os.path.exists(sha1_path):
-      images_to_delete.append(os.path.join(directory, f))
-  for image_path in images_to_delete:
-    os.remove(image_path)
-
-  # Downloading the files can be very spammy, so only show the output if
-  # something actually goes wrong.
-  try:
-    subprocess.check_output([
-        'download_from_google_storage',
-        '--bucket', STORAGE_BUCKET,
-        '-d', directory,
-        '-t', str(THREAD_COUNT),
-    ], stderr=subprocess.STDOUT)
-  except subprocess.CalledProcessError as e:
-    print ('Downloading RenderTest goldens in directory %s failed with error '
-           '%d: %s') % (directory, e.returncode, e.output)
-
-
-def upload(directory, dry_run):
-  files_to_upload = []
-  for f in os.listdir(directory):
-    # Skip any files that we don't care about.
-    if not is_file_of_interest(f):
-      continue
-
-    png_path = os.path.join(directory, f)
-    # upload_to_google_storage will upload a file even if it already exists
-    # in the bucket. As an optimization, hash locally and only pass files to
-    # the upload script if they don't have a matching .sha1 file already.
-    sha_path = png_path + '.sha1'
-    if os.path.isfile(sha_path):
-      with open(sha_path) as sha_file:
-        with open(png_path, 'rb') as png_file:
-          h = hashlib.sha1()
-          h.update(png_file.read())
-          if sha_file.read() == h.hexdigest():
-            continue
-    files_to_upload.append(png_path)
-
-  if len(files_to_upload):
-    if dry_run:
-      print ('Will upload the following files:')
-      for f in files_to_upload:
-        print ('  ' + f)
-      return
-    subprocess.check_call([
-        'upload_to_google_storage.py',
-        '--bucket', STORAGE_BUCKET,
-        '-t', str(THREAD_COUNT),
-    ] + files_to_upload)
-
-
 def main():
   parser = argparse.ArgumentParser()
   parser.add_argument('action', choices=['download', 'upload'],
                       help='Which action to perform')
   parser.add_argument('--dry_run', action='store_true',
-                      help='Dry run for uploading')
+                      default=False, help='Dry run for uploading')
   args = parser.parse_args()
 
   if args.action == 'download':
     for d in GOLDEN_DIRECTORIES:
-      download(d)
+      download(d, _is_file_of_interest,
+               'RenderTest Goldens', STORAGE_BUCKET)
   else:
     for d in GOLDEN_DIRECTORIES:
-      upload(d, args.dry_run)
+      upload(d, _is_file_of_interest,
+             'RenderTest Goldens', STORAGE_BUCKET, args.dry_run)
 
 
 if __name__ == '__main__':

chrome/test/data/android/manage_wpr_archives.py

@@ -0,0 +1,50 @@
+#!/usr/bin/env python
+# Copyright 2020 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.
+
+# Simple script to automatically download all current wpr archive files
+# for Android WPR record replay tests or upload any newly generated ones.
+
+import argparse
+import os
+
+from upload_download_utils import download
+from upload_download_utils import upload
+
+STORAGE_BUCKET = 'chrome-wpr-archives'
+THIS_DIR = os.path.abspath(os.path.dirname(__file__))
+CHROMIUM_SRC = os.path.abspath(os.path.join(THIS_DIR, '..', '..', '..', '..'))
+WPR_RECORD_REPLAY_TEST_DIRECTORIES = [
+  os.path.join(
+      CHROMIUM_SRC, 'chrome', 'android', 'feed', 'core', 'javatests',
+      'src', 'org', 'chromium', 'chrome', 'browser', 'feed', 'wpr_tests'),
+]
+
+
+def _is_file_of_interest(f):
+  """Filter all wprgo archive file through."""
+  return f.endswith('.wprgo')
+
+
+def main():
+  parser = argparse.ArgumentParser()
+  parser.add_argument('action', choices=['download', 'upload'],
+                      help='Which action to perform')
+  parser.add_argument('--dry_run', action='store_true',
+                      default=False, help='Dry run for uploading')
+  args = parser.parse_args()
+
+  if args.action == 'download':
+    for d in WPR_RECORD_REPLAY_TEST_DIRECTORIES:
+      download(d, _is_file_of_interest,
+               'WPR archives', STORAGE_BUCKET)
+  else:
+    for d in WPR_RECORD_REPLAY_TEST_DIRECTORIES:
+      upload(d, _is_file_of_interest,
+             'WPR archives', STORAGE_BUCKET, args.dry_run)
+
+
+if __name__ == '__main__':
+  main()
+

chrome/test/data/android/upload_download_utils.py

@@ -0,0 +1,130 @@
+# Copyright 2020 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.
+
+# Script lib to download all current files from GCS or upload any newly
+# generated ones.
+
+import hashlib
+import logging
+import multiprocessing
+import os
+import subprocess
+
+
+def _get_thread_count():
+  """Gets a thread_count based on the multiprocessing.cpu_count()."""
+  try:
+    thread_count = multiprocessing.cpu_count()
+    # cpu_count only gets the physical core count. There doesn't appear to be a
+    # simple way of determining whether a CPU supports simultaneous
+    # multithreading in Python, so assume that anything with 6 or more cores
+    # supports it.
+    if thread_count >= 6:
+      thread_count *= 2
+  except NotImplementedError:
+    # Assume a quad core if we can't get the actual core count.
+    thread_count = 4
+  return thread_count
+
+
+def download(directory, filter, scenario, gcs_bucket):
+  """Downloads the files from cloud as according to the given directory.
+
+  Args:
+    directory: A local disk directory.
+    filter: A method to filter target files inside this directory.
+    scenario: A scenario names that utilize this download method.
+    gcs_bucket: A GCS bucket name to be used in the download.
+  """
+  # If someone removes a SHA1 file, we want to remove the associated cloud file
+  # the next time file are updated.
+  files_to_delete = _get_files_to_delete(directory, filter)
+  for file_path in files_to_delete:
+    os.remove(file_path)
+
+  # Downloading the files can be very spammy, so only show the output if
+  # something actually goes wrong.
+  try:
+    subprocess.check_output([
+        'download_from_google_storage',
+        '--bucket', gcs_bucket,
+        '-d', directory,
+        '-t', str(_get_thread_count()),
+    ], stderr=subprocess.STDOUT)
+  except subprocess.CalledProcessError as e:
+    logging.error('Downloading %s in directory %s failed with error '
+                  '%d: %s', scenario, directory, e.returncode, e.output)
+
+
+def _get_files_to_delete(directory, filter):
+  """Returns a list of local files to delete.
+
+  Args:
+    directory: A local disk directory.
+    filter: A method to filter target files inside the directory.
+  """
+  files_to_delete = []
+  for f in os.listdir(directory):
+    if not filter(f):
+      continue
+    sha1_path = os.path.join(directory, f + '.sha1')
+    if not os.path.exists(sha1_path):
+      files_to_delete.append(os.path.join(directory, f))
+  return files_to_delete
+
+
+def upload(directory, filter, scenario, gcs_bucket, dry_run=False):
+  """Uploads the files to cloud storage.
+
+  Args:
+    directory: A local disk directory.
+    filter: A method to filter target files inside the directory.
+    scenarios: A scenario that calls this upload.
+    gcs_bucket: A GCS bucket name to be used in upload.
+    dry_run: A flag to indicate whether it is just a dry_run or a real upload.
+  """
+  files_to_upload = _get_files_to_upload(directory, filter)
+
+  if len(files_to_upload):
+    if dry_run:
+      logging.info('Will upload the following %s: %s',
+                   scenario, ', '.join(files_to_upload))
+      logging.info('Destination gcs bucket: %s', gcs_bucket)
+      return
+    subprocess.check_call([
+        'upload_to_google_storage.py',
+        '--bucket', gcs_bucket,
+        '-t', str(_get_thread_count()),
+    ] + files_to_upload)
+
+
+def _get_files_to_upload(directory, filter):
+  """Returns a list of files to upload to cloud storage.
+
+  Args:
+    directory: A local disk directory.
+    filter: A method to filter target files inside the directory.
+  """
+  files_to_upload = []
+  for f in os.listdir(directory):
+    # Skip any files that we don't care about.
+    if not filter(f):
+      continue
+
+    file_path = os.path.join(directory, f)
+    # upload_to_google_storage will upload a file even if it already exists
+    # in the bucket. As an optimization, hash locally and only pass files to
+    # the upload script if they don't have a matching .sha1 file already.
+    sha_path = file_path + '.sha1'
+    if os.path.isfile(sha_path):
+      with open(sha_path) as sha_file:
+        with open(file_path, 'rb') as source_file:
+          h = hashlib.sha1()
+          h.update(source_file.read())
+          if sha_file.read() == h.hexdigest():
+            continue
+    files_to_upload.append(file_path)
+  return files_to_upload
+
+

chrome/test/data/android/upload_download_utils_test.py

@@ -0,0 +1,129 @@
+#!/usr/bin/env vpython
+# Copyright 2020 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.
+
+# Lint as: python3
+"""Tests for upload_download_utils.
+
+To Execute:
+  PYTHONPATH=chrome/test chrome/test/data/android/upload_download_utils_test.py
+"""
+import multiprocessing
+import os
+import sys
+import unittest
+import upload_download_utils
+
+THIS_DIR = os.path.abspath(os.path.dirname(__file__))
+CHROMIUM_SRC = os.path.abspath(os.path.join(THIS_DIR, '..', '..', '..', '..'))
+PYMOCK_PATH = os.path.join(
+    CHROMIUM_SRC, 'third_party', 'pymock')
+sys.path.append(PYMOCK_PATH)
+import mock
+
+from pyfakefs import fake_filesystem_unittest
+
+
+def filter_all(f):
+  """Filter everything through."""
+  return True
+
+def filter_image_only(f):
+  """Filter only png file through."""
+  if f.endswith('png'):
+    return True
+  return False
+
+class UploadDownloadUtilsTest(fake_filesystem_unittest.TestCase):
+  def setUp(self):
+    """Set up the fake file system."""
+    self.setUpPyfakefs()
+
+  def test_get_files_to_delete(self):
+    pathname = os.path.join(THIS_DIR, 'tmp')
+    self.fs.create_dir(pathname)
+    self.fs.create_file(os.path.join(pathname, 'image1.png'))
+    self.fs.create_file(os.path.join(pathname, 'image1.sha1'))
+    self.fs.create_file(os.path.join(pathname, 'image2.png'))
+    self.fs.create_file(os.path.join(pathname, 'image3.png'))
+    self.fs.create_file(os.path.join(pathname, 'image3.png.sha1'))
+    self.fs.create_file(os.path.join(pathname, 'scenario1.wprgo'))
+    self.fs.create_file(os.path.join(pathname, 'scenario1.wprgo.sha1'))
+
+    expected_file_list = [
+        os.path.join(pathname, 'image1.png'),
+        os.path.join(pathname, 'image1.sha1'),
+        os.path.join(pathname, 'image2.png'),
+        os.path.join(pathname, 'image3.png.sha1'),
+        os.path.join(pathname, 'scenario1.wprgo.sha1'),
+    ].sort()
+    self.assertEquals(
+        upload_download_utils._get_files_to_delete(pathname, filter_all).sort(),
+        expected_file_list)
+
+    expected_file_list = [
+        os.path.join(pathname, 'image1.png'),
+        os.path.join(pathname, 'image2.png'),
+    ].sort()
+    self.assertEquals(
+        upload_download_utils._get_files_to_delete(pathname, filter_image_only)
+            .sort(),
+        expected_file_list)
+
+  def test_get_files_to_upload_with_entries(self):
+    pathname = os.path.join(THIS_DIR, 'tmp')
+    self.fs.create_dir(pathname)
+    self.fs.create_file(os.path.join(pathname, 'image1.png'))
+    self.fs.create_file(os.path.join(pathname, 'image1.png.sha1'))
+    self.fs.create_file(os.path.join(pathname, 'image2.png'))
+    self.fs.create_file(os.path.join(pathname, 'scenario1.wprgo'))
+    expected_file_list = [
+        os.path.join(pathname, 'image1.png'),
+        os.path.join(pathname, 'image1.png.sha1'),
+        os.path.join(pathname, 'image2.png'),
+        os.path.join(pathname, 'scenario1.wprgo'),
+    ].sort()
+    self.assertEquals(
+        upload_download_utils._get_files_to_upload(pathname, filter_all).sort(),
+        expected_file_list)
+
+    expected_file_list = [
+        os.path.join(pathname, 'image1.png'),
+        os.path.join(pathname, 'image2.png'),
+    ].sort()
+    self.assertEquals(
+        upload_download_utils._get_files_to_upload(pathname, filter_image_only)
+            .sort(),
+        expected_file_list)
+
+  def test_get_files_to_upload_no_entries(self):
+    pathname = os.path.join(THIS_DIR, 'tmp')
+    self.fs.create_dir(pathname)
+    self.fs.create_file(os.path.join(pathname, 'scenario1.wprgo'))
+    self.fs.create_file(os.path.join(pathname, 'scenario1.wprgo.sha1'))
+    self.fs.create_file(os.path.join(pathname, 'scenario2.wprgo'))
+    self.fs.create_file(os.path.join(pathname, 'scenario3.wprgo'))
+    self.assertEquals(
+        upload_download_utils._get_files_to_upload(pathname, filter_image_only),
+        [])
+
+  @mock.patch.object(multiprocessing, 'cpu_count')
+  def test_get_thread_count_not_implemented(self, multiplecore_mock):
+    multiplecore_mock.side_effect = NotImplementedError('abc')
+    self.assertEqual(upload_download_utils._get_thread_count(), 4)
+
+  @mock.patch.object(multiprocessing, 'cpu_count')
+  def test_get_thread_count_with_quad_core(self, multiplecore_mock):
+    multiplecore_mock.return_value = 4
+    self.assertEqual(upload_download_utils._get_thread_count(), 4)
+
+  @mock.patch.object(multiprocessing, 'cpu_count')
+  def test_get_thread_count_with_hexa_core(self, multiplecore_mock):
+    multiplecore_mock.return_value = 6
+    self.assertEqual(upload_download_utils._get_thread_count(), 12)
+
+
+if __name__ == '__main__':
+  unittest.main(verbosity=2)
+