new tool: download all GM images for a given builder, ready for skdiff

gm/rebaseline_server/download.py

Index: gm/rebaseline_server/download.py
diff --git a/gm/rebaseline_server/download.py b/gm/rebaseline_server/download.py
new file mode 100755
index 0000000000000000000000000000000000000000..daa6503e69fe653aea5e7d7d95e54b978782fe44
--- /dev/null
+++ b/gm/rebaseline_server/download.py
@@ -0,0 +1,136 @@
+#!/usr/bin/python
+
+"""
+Copyright 2014 Google Inc.
+
+Use of this source code is governed by a BSD-style license that can be
+found in the LICENSE file.
+
+Download actual GM results for a particular builder.
+"""
+
+# System-level imports
+import optparse
+import os
+import re
+import sys
+
+# Imports from within Skia
+#
+# We need to add the 'gm' directory, so that we can import gm_json.py within
+# that directory.  That script allows us to parse the actual-results.json file
+# written out by the GM tool.
+# Make sure that the 'gm' dir is in the PYTHONPATH, but add it at the *end*
+# so any dirs that are already in the PYTHONPATH will be preferred.
+PARENT_DIRECTORY = os.path.dirname(os.path.realpath(__file__))
+GM_DIRECTORY = os.path.dirname(PARENT_DIRECTORY)
+if GM_DIRECTORY not in sys.path:
+  sys.path.append(GM_DIRECTORY)
+import gm_json
+import url_or_path
+
+DEFAULT_ACTUALS_BASE_URL = 'http://skia-autogen.googlecode.com/svn/gm-actual'
+DEFAULT_JSON_FILENAME = 'actual-results.json'
+
+
+class Download(object):
+
+  def __init__(self, actuals_base_url=DEFAULT_ACTUALS_BASE_URL,
+               json_filename=DEFAULT_JSON_FILENAME,
+               gm_actuals_root_url=gm_json.GM_ACTUALS_ROOT_HTTP_URL):
+    """
+    Args:
+      actuals_base_url: URL or local filepath pointing at the root directory
+          containing all actual-results.json files, e.g.,
+          http://domain.name/path/to/dir  OR
+          /absolute/path/to/localdir  (on Linux) OR
+          relative\path\to\localdir   (on Windows)
+      json_filename: The JSON filename to read from within each directory.
+      gm_actuals_root_url: Base URL under which the actually-generated-by-bots
+          GM images are stored.
+    """
+    self._actuals_base_url = actuals_base_url
+    self._json_filename = json_filename
+    self._gm_actuals_root_url = gm_actuals_root_url
+    self._image_filename_re = re.compile(gm_json.IMAGE_FILENAME_PATTERN)
+
+  def fetch(self, builder_name, dest_dir):
+    """ Downloads actual GM results for a particular builder.
+
+    Args:
+      builder_name: which builder to download results of
+      dest_dir: path to directory where the image files will be written;
+                if the directory does not exist yet, it will be created
+
+    TODO(epoger): Display progress info.  Right now, it can take a long time
+    to download all of the results, and there is no indication of progress.
+
+    TODO(epoger): Download multiple images in parallel to speed things up.
+    """
+    json_url = url_or_path.join(self._actuals_base_url, builder_name,
+                                self._json_filename)
+    json_contents = url_or_path.read_as_string(json_url)
+    results_dict = gm_json.LoadFromString(json_contents)
+
+    actual_results_dict = results_dict[gm_json.JSONKEY_ACTUALRESULTS]
+    for result_type in sorted(actual_results_dict.keys()):
+      results_of_this_type = actual_results_dict[result_type]
+      if not results_of_this_type:
+        continue
+      for image_name in sorted(results_of_this_type.keys()):
+        (test, config) = self._image_filename_re.match(image_name).groups()
+        (hash_type, hash_digest) = results_of_this_type[image_name]
+        source_url = gm_json.CreateGmActualUrl(
+            test_name=test, hash_type=hash_type, hash_digest=hash_digest,
+            gm_actuals_root_url=self._gm_actuals_root_url)
+        dest_path = os.path.join(dest_dir, config, test + '.png')
+        url_or_path.copy_contents(source_path=source_url, dest_path=dest_path,
+                                  create_subdirs_if_needed=True)
+
+
+def main():
+  parser = optparse.OptionParser()
+  parser.add_option('--actuals-base-url',
+                    action='store', type='string',
+                    default=DEFAULT_ACTUALS_BASE_URL,
+                    help=('Base URL from which to read files containing JSON '
+                          'summaries of actual GM results; defaults to '
+                          '"%default". To get a specific revision (useful for '
+                          'trybots) replace "svn" with "svn-history/r123".'))
+  parser.add_option('--builder',
+                    action='store', type='string',
+                    default=None,
+                    help=('REQUIRED: Which builder to download results for. '
+                          'To see a list of builders, run "svn ls %s".' %

[epoger, 2014/01/23 04:15:32]

At a cost, yes.

If you look at
https://code.google.com/p/skia/source/browse/trunk/tools/rebaseline.py?r=1271...
, you will see that rebaseline.py doesn't provide an up-to-date list of
builders, and an indication of why not.

For now, I added information that the caller can use to easily generate the list
herself... does that seem sufficient?

If you'd like me to add a TODO to provide more, I'm all for it--please tell me
what would be your preferred methodology (display the builder list within --help
output, add a --list-builders flag that lists builders instead of doing anything
else, or ???)

[rmistry, 2014/01/23 12:48:23]

IMO having a --list-builders step would make it the easiest for users to list
everything out and then pick and choose as needed. This would be consistent with
the submit_try tool.

[bsalomon, 2014/01/23 14:20:33]

This seems totally sufficient. Thanks!

[epoger, 2014/01/23 15:34:19]

Discussed live: I will add this as a TODO, but "svn ls" is the best way to go
for now anyway because the builders are spread across multiple masters anyway.

[epoger, 2014/01/24 02:00:46]

Done.

+                          DEFAULT_ACTUALS_BASE_URL))
+  parser.add_option('--dest-dir',
+                    action='store', type='string',
+                    default=None,
+                    help=('REQUIRED: Directory where all images should be '
+                          'written. If this directory does not exist yet, it '
+                          'will be created.'))
+  parser.add_option('--json-filename',
+                    action='store', type='string',
+                    default=DEFAULT_JSON_FILENAME,
+                    help=('JSON summary filename to read for each builder; '
+                          'defaults to "%default".'))
+  (params, remaining_args) = parser.parse_args()
+
+  # Make sure all required options were set (no params should have value None),
+  # and that there were no items left over in the command line.
+  param_dict = vars(params)
+  for param_name, param_value in param_dict.items():
+    if param_value is None:
+      raise Exception('required option \'%s\' was not set' % param_name)
+  if len(remaining_args) is not 0:
+    raise Exception('extra items specified in the command line: %s' %
+                    remaining_args)
+
+  downloader = Download(actuals_base_url=params.actuals_base_url)
+  downloader.fetch(builder_name=params.builder,
+                   dest_dir=params.dest_dir)
+
+
+
+if __name__ == '__main__':
+  main()