teach rebaseline_server to generate diffs of rendered SKPs

gm/rebaseline_server/compare_rendered_pictures.py

 #!/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.
 
 Compare results of two render_pictures runs.
+
+EPOGER: Rename this module, because it is more generally useful now... it can download and compare any types of images.
 """
 
 # System-level imports
 import logging
 import os
+import shutil
+import tempfile
 import time
 
 # Must fix up PYTHONPATH before importing from within Skia
 import fix_pythonpath  # pylint: disable=W0611
 
 # Imports from within Skia
 from py.utils import url_utils
+import buildbot_globals
 import gm_json
 import imagediffdb
 import imagepair
 import imagepairset
 import results
 
 # URL under which all render_pictures images can be found in Google Storage.
 #
-# pylint: disable=C0301
-# TODO(epoger): Move this default value into
-# https://skia.googlesource.com/buildbot/+/master/site_config/global_variables.json
-# pylint: enable=C0301
-DEFAULT_IMAGE_BASE_URL = (
-    'http://chromium-skia-gm.commondatastorage.googleapis.com/'
-    'render_pictures/images')
+# EPOGER: In order to allow live-view of GMs and other images, read this from the input summary files, or allow the caller to set it within the GET_live_results call?
+DEFAULT_IMAGE_BASE_GS_URL = 'gs://' + buildbot_globals.Get('skp_images_bucket')
 
 
 class RenderedPicturesComparisons(results.BaseComparisons):
-  """Loads results from two different render_pictures runs into an ImagePairSet.
+  """Loads results from multiple render_pictures runs into an ImagePairSet.
   """
 
-  def __init__(self, subdirs, actuals_root,
-               generated_images_root=results.DEFAULT_GENERATED_IMAGES_ROOT,
-               image_base_url=DEFAULT_IMAGE_BASE_URL,
-               diff_base_url=None):
+  def __init__(self, actuals_dirs, expectations_dirs, image_diff_db,
+               image_base_gs_url=DEFAULT_IMAGE_BASE_GS_URL,
+               diff_base_url=None, actuals_label='actuals',
+               expectations_label='expectations'):
     """
     Args:
-      actuals_root: root directory containing all render_pictures-generated
-          JSON files
-      subdirs: (string, string) tuple; pair of subdirectories within
-          actuals_root to compare
-      generated_images_root: directory within which to create all pixel diffs;
-          if this directory does not yet exist, it will be created
-      image_base_url: URL under which all render_pictures result images can
+      actuals_dirs: list of root directories to copy all JSON summaries from,
+          and to use as actual results
+      expectations_dirs: list of root directories to copy all JSON summaries
+          from, and to use as expected results
+      image_diff_db: ImageDiffDB instance
+      image_base_gs_url: "gs://" URL pointing at the Google Storage bucket/dir
+          under which all render_pictures result images can
           be found; this will be used to read images for comparison within
-          this code, and included in the ImagePairSet so its consumers know
-          where to download the images from
+          this code, and included in the ImagePairSet (as an HTTP URL) so its
+          consumers know where to download the images from
       diff_base_url: base URL within which the client should look for diff
           images; if not specified, defaults to a "file:///" URL representation
-          of generated_images_root
+          of image_diff_db's storage_root
+      actuals_label: description to use for actual results
+      expectations_label: description to use for expected results
     """
-    time_start = int(time.time())
-    self._image_diff_db = imagediffdb.ImageDiffDB(generated_images_root)
-    self._image_base_url = image_base_url
+    self._image_diff_db = image_diff_db
+    self._image_base_gs_url = image_base_gs_url
     self._diff_base_url = (
         diff_base_url or
-        url_utils.create_filepath_url(generated_images_root))
-    self._load_result_pairs(actuals_root, subdirs)
-    self._timestamp = int(time.time())
-    logging.info('Results complete; took %d seconds.' %
-                 (self._timestamp - time_start))
+        url_utils.create_filepath_url(image_diff_db.storage_root))
+    self._actuals_label = actuals_label
+    self._expectations_label = expectations_label
 
-  def _load_result_pairs(self, actuals_root, subdirs):
-    """Loads all JSON files found within two subdirs in actuals_root,
-    compares across those two subdirs, and stores the summary in self._results.
+    tempdir = tempfile.mkdtemp()
+    try:
+      actuals_root = os.path.join(tempdir, 'actuals')
+      expectations_root = os.path.join(tempdir, 'expectations')
+      # Copy all summary files into actuals_root and expectations_root
+      # EPOGER: this won't work when *_dirs have more than one entry
+      for dir in actuals_dirs:
+        shutil.copytree(dir, actuals_root)
+      for dir in expectations_dirs:
+        shutil.copytree(dir, expectations_root)
+
+      time_start = int(time.time())
+      self._load_result_pairs(actuals_root, expectations_root)
+      self._timestamp = int(time.time())
+      logging.info('Results complete; took %d seconds.' %
+                   (self._timestamp - time_start))
+    finally:
+      shutil.rmtree(tempdir)
+
+  def _load_result_pairs(self, actuals_root, expectations_root):
+    """Loads all JSON image summaries from 2 directory trees and compares them.
+
+    The summary of all image diff results is stored in in self._results.
+    EPOGER: we don't want to store the results in self._results anymore... now, we just want to return them
 
     Args:
-      actuals_root: root directory containing all render_pictures-generated
-          JSON files
-      subdirs: (string, string) tuple; pair of subdirectories within
-          actuals_root to compare
+      actuals_root: root directory containing JSON summaries of actual results
+      expectations_root: root dir containing JSON summaries of expected results
     """
-    logging.info(
-        'Reading actual-results JSON files from %s subdirs within %s...' % (
-            subdirs, actuals_root))
-    subdirA, subdirB = subdirs
-    subdirA_dicts = self._read_dicts_from_root(
-        os.path.join(actuals_root, subdirA))
-    subdirB_dicts = self._read_dicts_from_root(
-        os.path.join(actuals_root, subdirB))
-    logging.info('Comparing subdirs %s and %s...' % (subdirA, subdirB))
+    logging.info('Reading JSON image summaries from dirs %s and %s...' % (
+        actuals_root, expectations_root))
+    actuals_dicts = self._read_dicts_from_root(actuals_root)
+    expectations_dicts = self._read_dicts_from_root(expectations_root)
+    logging.info('Comparing summary dicts...')
 
     all_image_pairs = imagepairset.ImagePairSet(
-        descriptions=subdirs,
+        descriptions=(self._actuals_label, self._expectations_label),
         diff_base_url=self._diff_base_url)
     failing_image_pairs = imagepairset.ImagePairSet(
-        descriptions=subdirs,
+        descriptions=(self._actuals_label, self._expectations_label),
         diff_base_url=self._diff_base_url)
 
     all_image_pairs.ensure_extra_column_values_in_summary(
         column_id=results.KEY__EXTRACOLUMNS__RESULT_TYPE, values=[
             results.KEY__RESULT_TYPE__FAILED,
             results.KEY__RESULT_TYPE__NOCOMPARISON,
             results.KEY__RESULT_TYPE__SUCCEEDED,
         ])
     failing_image_pairs.ensure_extra_column_values_in_summary(
         column_id=results.KEY__EXTRACOLUMNS__RESULT_TYPE, values=[
             results.KEY__RESULT_TYPE__FAILED,
             results.KEY__RESULT_TYPE__NOCOMPARISON,
         ])
 
-    common_dict_paths = sorted(set(subdirA_dicts.keys() + subdirB_dicts.keys()))
+    # Every actuals_dict should be paired up with a corresponding
+    # expectations_dict.
+    actuals_dict_paths = sorted(actuals_dicts.keys())
+    expectations_dict_paths = sorted(expectations_dicts.keys())
+    if actuals_dict_paths != expectations_dict_paths:
+      raise Exception('actuals_dict_paths %s != expectations_dict_paths %s' % (
+          actuals_dict_paths, expectations_dict_paths))
+    common_dict_paths = actuals_dict_paths
+
     num_common_dict_paths = len(common_dict_paths)
     dict_num = 0
     for dict_path in common_dict_paths:
       dict_num += 1
       logging.info('Generating pixel diffs for dict #%d of %d, "%s"...' %
                    (dict_num, num_common_dict_paths, dict_path))
-      dictA = subdirA_dicts[dict_path]
-      dictB = subdirB_dicts[dict_path]
+      dictA = actuals_dicts[dict_path]
+      dictB = expectations_dicts[dict_path]
       self._validate_dict_version(dictA)
       self._validate_dict_version(dictB)
       dictA_results = dictA[gm_json.JSONKEY_ACTUALRESULTS]
       dictB_results = dictB[gm_json.JSONKEY_ACTUALRESULTS]
       skp_names = sorted(set(dictA_results.keys() + dictB_results.keys()))
       for skp_name in skp_names:
         imagepairs_for_this_skp = []
 
         whole_image_A = RenderedPicturesComparisons.get_multilevel(
             dictA_results, skp_name, gm_json.JSONKEY_SOURCE_WHOLEIMAGE)
@@ skipping 96 matching lines @@
         results.KEY__EXTRACOLUMNS__TEST: test,
         # TODO(epoger): Right now, the client UI crashes if it receives
         # results that do not include this column.
         # Until we fix that, keep the client happy.
         results.KEY__EXTRACOLUMNS__BUILDER: 'TODO',
     }
 
     try:
       return imagepair.ImagePair(
           image_diff_db=self._image_diff_db,
-          base_url=self._image_base_url,
+          base_url=self._image_base_gs_url,
           imageA_relative_url=imageA_relative_url,
           imageB_relative_url=imageB_relative_url,
           extra_columns=extra_columns_dict)
     except (KeyError, TypeError):
       logging.exception(
           'got exception while creating ImagePair for'
           ' test="%s", config="%s", urlPair=("%s","%s")' % (
               test, config, imageA_relative_url, imageB_relative_url))
       return None
-
-
-# TODO(epoger): Add main() so this can be called by vm_run_skia_try.sh