1#!/usr/bin/python
2
3'''
4Copyright 2013 Google Inc.
5
6Use of this source code is governed by a BSD-style license that can be
7found in the LICENSE file.
8'''
9
10'''
11Gathers diffs between 2 JSON expectations files, or between actual and
12expected results within a single JSON actual-results file,
13and generates an old-vs-new diff dictionary.
14
15TODO(epoger): Fix indentation in this file (2-space indents, not 4-space).
16'''
17
18# System-level imports
19import argparse
20import json
21import os
22import sys
23import urllib2
24
25# Imports from within Skia
26#
27# We need to add the 'gm' directory, so that we can import gm_json.py within
28# that directory.  That script allows us to parse the actual-results.json file
29# written out by the GM tool.
30# Make sure that the 'gm' dir is in the PYTHONPATH, but add it at the *end*
31# so any dirs that are already in the PYTHONPATH will be preferred.
32#
33# This assumes that the 'gm' directory has been checked out as a sibling of
34# the 'tools' directory containing this script, which will be the case if
35# 'trunk' was checked out as a single unit.
36GM_DIRECTORY = os.path.realpath(
37    os.path.join(os.path.dirname(os.path.dirname(__file__)), 'gm'))
38if GM_DIRECTORY not in sys.path:
39    sys.path.append(GM_DIRECTORY)
40import gm_json
41
42
43# Object that generates diffs between two JSON gm result files.
44class GMDiffer(object):
45
46    def __init__(self):
47        pass
48
49    def _GetFileContentsAsString(self, filepath):
50        """Returns the full contents of a file, as a single string.
51        If the filename looks like a URL, download its contents.
52        If the filename is None, return None."""
53        if filepath is None:
54            return None
55        elif filepath.startswith('http:') or filepath.startswith('https:'):
56            return urllib2.urlopen(filepath).read()
57        else:
58            return open(filepath, 'r').read()
59
60    def _GetExpectedResults(self, contents):
61        """Returns the dictionary of expected results from a JSON string,
62        in this form:
63
64        {
65          'test1' : 14760033689012826769,
66          'test2' : 9151974350149210736,
67          ...
68        }
69
70        We make these simplifying assumptions:
71        1. Each test has either 0 or 1 allowed results.
72        2. All expectations are of type JSONKEY_HASHTYPE_BITMAP_64BITMD5.
73
74        Any tests which violate those assumptions will cause an exception to
75        be raised.
76
77        Any tests for which we have no expectations will be left out of the
78        returned dictionary.
79        """
80        result_dict = {}
81        json_dict = gm_json.LoadFromString(contents)
82        all_expectations = json_dict[gm_json.JSONKEY_EXPECTEDRESULTS]
83        for test_name in all_expectations.keys():
84            test_expectations = all_expectations[test_name]
85            allowed_digests = test_expectations[
86                gm_json.JSONKEY_EXPECTEDRESULTS_ALLOWEDDIGESTS]
87            if allowed_digests:
88                num_allowed_digests = len(allowed_digests)
89                if num_allowed_digests > 1:
90                    raise ValueError(
91                        'test %s has %d allowed digests' % (
92                            test_name, num_allowed_digests))
93                digest_pair = allowed_digests[0]
94                if digest_pair[0] != gm_json.JSONKEY_HASHTYPE_BITMAP_64BITMD5:
95                    raise ValueError(
96                        'test %s has unsupported hashtype %s' % (
97                            test_name, digest_pair[0]))
98                result_dict[test_name] = digest_pair[1]
99        return result_dict
100
101    def _GetActualResults(self, contents):
102        """Returns the dictionary of actual results from a JSON string,
103        in this form:
104
105        {
106          'test1' : 14760033689012826769,
107          'test2' : 9151974350149210736,
108          ...
109        }
110
111        We make these simplifying assumptions:
112        1. All results are of type JSONKEY_HASHTYPE_BITMAP_64BITMD5.
113
114        Any tests which violate those assumptions will cause an exception to
115        be raised.
116
117        Any tests for which we have no actual results will be left out of the
118        returned dictionary.
119        """
120        result_dict = {}
121        json_dict = gm_json.LoadFromString(contents)
122        all_result_types = json_dict[gm_json.JSONKEY_ACTUALRESULTS]
123        for result_type in all_result_types.keys():
124            results_of_this_type = all_result_types[result_type]
125            if results_of_this_type:
126                for test_name in results_of_this_type.keys():
127                    digest_pair = results_of_this_type[test_name]
128                    if digest_pair[0] != gm_json.JSONKEY_HASHTYPE_BITMAP_64BITMD5:
129                        raise ValueError(
130                            'test %s has unsupported hashtype %s' % (
131                                test_name, digest_pair[0]))
132                    result_dict[test_name] = digest_pair[1]
133        return result_dict
134
135    def _DictionaryDiff(self, old_dict, new_dict):
136        """Generate a dictionary showing the diffs between old_dict and new_dict.
137        Any entries which are identical across them will be left out."""
138        diff_dict = {}
139        all_keys = set(old_dict.keys() + new_dict.keys())
140        for key in all_keys:
141            if old_dict.get(key) != new_dict.get(key):
142                new_entry = {}
143                new_entry['old'] = old_dict.get(key)
144                new_entry['new'] = new_dict.get(key)
145                diff_dict[key] = new_entry
146        return diff_dict
147
148    def GenerateDiffDict(self, oldfile, newfile=None):
149        """Generate a dictionary showing the diffs:
150        old = expectations within oldfile
151        new = expectations within newfile
152
153        If newfile is not specified, then 'new' is the actual results within
154        oldfile.
155        """
156        return self.GenerateDiffDictFromStrings(self._GetFileContentsAsString(oldfile),
157                                                self._GetFileContentsAsString(newfile))
158
159    def GenerateDiffDictFromStrings(self, oldjson, newjson=None):
160        """Generate a dictionary showing the diffs:
161        old = expectations within oldjson
162        new = expectations within newjson
163
164        If newfile is not specified, then 'new' is the actual results within
165        oldfile.
166        """
167        old_results = self._GetExpectedResults(oldjson)
168        if newjson:
169            new_results = self._GetExpectedResults(newjson)
170        else:
171            new_results = self._GetActualResults(oldjson)
172        return self._DictionaryDiff(old_results, new_results)
173
174
175def _Main():
176    parser = argparse.ArgumentParser()
177    parser.add_argument(
178        'old',
179        help='Path to JSON file whose expectations to display on ' +
180        'the "old" side of the diff. This can be a filepath on ' +
181        'local storage, or a URL.')
182    parser.add_argument(
183        'new', nargs='?',
184        help='Path to JSON file whose expectations to display on ' +
185        'the "new" side of the diff; if not specified, uses the ' +
186        'ACTUAL results from the "old" JSON file. This can be a ' +
187        'filepath on local storage, or a URL.')
188    args = parser.parse_args()
189    differ = GMDiffer()
190    diffs = differ.GenerateDiffDict(oldfile=args.old, newfile=args.new)
191    json.dump(diffs, sys.stdout, sort_keys=True, indent=2)
192
193
194if __name__ == '__main__':
195    _Main()
196