tools/trace: Tool to compare json state dumps.
authorJosé Fonseca <jfonseca@vmware.com>
Thu, 20 Jun 2013 18:40:47 +0000 (19:40 +0100)
committerJosé Fonseca <jfonseca@vmware.com>
Fri, 21 Jun 2013 13:30:20 +0000 (14:30 +0100)
Copied verbatim from apitrace's scripts/jsondiff.py
Reviewed-by: Brian Paul <brianp@vmware.com>
src/gallium/tools/trace/diff_state.py [new file with mode: 0755]

diff --git a/src/gallium/tools/trace/diff_state.py b/src/gallium/tools/trace/diff_state.py
new file mode 100755 (executable)
index 0000000..470aeda
--- /dev/null
@@ -0,0 +1,324 @@
+#!/usr/bin/env python
+##########################################################################
+#
+# Copyright 2011 Jose Fonseca
+# All Rights Reserved.
+#
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in
+# all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+# THE SOFTWARE.
+#
+##########################################################################/
+
+
+import json
+import optparse
+import re
+import sys
+
+
+def strip_object_hook(obj):
+    if '__class__' in obj:
+        return None
+    for name in obj.keys():
+        if name.startswith('__') and name.endswith('__'):
+            del obj[name]
+    return obj
+
+
+class Visitor:
+
+    def visit(self, node, *args, **kwargs):
+        if isinstance(node, dict):
+            return self.visitObject(node, *args, **kwargs)
+        elif isinstance(node, list):
+            return self.visitArray(node, *args, **kwargs)
+        else:
+            return self.visitValue(node, *args, **kwargs)
+
+    def visitObject(self, node, *args, **kwargs):
+        pass
+
+    def visitArray(self, node, *args, **kwargs):
+        pass
+
+    def visitValue(self, node, *args, **kwargs):
+        pass
+
+
+class Dumper(Visitor):
+
+    def __init__(self, stream = sys.stdout):
+        self.stream = stream
+        self.level = 0
+
+    def _write(self, s):
+        self.stream.write(s)
+
+    def _indent(self):
+        self._write('  '*self.level)
+
+    def _newline(self):
+        self._write('\n')
+
+    def visitObject(self, node):
+        self.enter_object()
+
+        members = node.keys()
+        members.sort()
+        for i in range(len(members)):
+            name = members[i]
+            value = node[name]
+            self.enter_member(name)
+            self.visit(value)
+            self.leave_member(i == len(members) - 1)
+        self.leave_object()
+
+    def enter_object(self):
+        self._write('{')
+        self._newline()
+        self.level += 1
+
+    def enter_member(self, name):
+        self._indent()
+        self._write('%s: ' % name)
+
+    def leave_member(self, last):
+        if not last:
+            self._write(',')
+        self._newline()
+
+    def leave_object(self):
+        self.level -= 1
+        self._indent()
+        self._write('}')
+        if self.level <= 0:
+            self._newline()
+
+    def visitArray(self, node):
+        self.enter_array()
+        for i in range(len(node)):
+            value = node[i]
+            self._indent()
+            self.visit(value)
+            if i != len(node) - 1:
+                self._write(',')
+            self._newline()
+        self.leave_array()
+
+    def enter_array(self):
+        self._write('[')
+        self._newline()
+        self.level += 1
+
+    def leave_array(self):
+        self.level -= 1
+        self._indent()
+        self._write(']')
+
+    def visitValue(self, node):
+        self._write(json.dumps(node))
+
+
+
+class Comparer(Visitor):
+
+    def __init__(self, ignore_added = False, tolerance = 2.0 ** -24):
+        self.ignore_added = ignore_added
+        self.tolerance = tolerance
+
+    def visitObject(self, a, b):
+        if not isinstance(b, dict):
+            return False
+        if len(a) != len(b) and not self.ignore_added:
+            return False
+        ak = a.keys()
+        bk = b.keys()
+        ak.sort()
+        bk.sort()
+        if ak != bk and not self.ignore_added:
+            return False
+        for k in ak:
+            ae = a[k]
+            try:
+                be = b[k]
+            except KeyError:
+                return False
+            if not self.visit(ae, be):
+                return False
+        return True
+
+    def visitArray(self, a, b):
+        if not isinstance(b, list):
+            return False
+        if len(a) != len(b):
+            return False
+        for ae, be in zip(a, b):
+            if not self.visit(ae, be):
+                return False
+        return True
+
+    def visitValue(self, a, b):
+        if isinstance(a, float) or isinstance(b, float):
+            if a == 0:
+                return abs(b) < self.tolerance
+            else:
+                return abs((b - a)/a) < self.tolerance
+        else:
+            return a == b
+
+
+class Differ(Visitor):
+
+    def __init__(self, stream = sys.stdout, ignore_added = False):
+        self.dumper = Dumper(stream)
+        self.comparer = Comparer(ignore_added = ignore_added)
+
+    def visit(self, a, b):
+        if self.comparer.visit(a, b):
+            return
+        Visitor.visit(self, a, b)
+
+    def visitObject(self, a, b):
+        if not isinstance(b, dict):
+            self.replace(a, b)
+        else:
+            self.dumper.enter_object()
+            names = set(a.keys())
+            if not self.comparer.ignore_added:
+                names.update(b.keys())
+            names = list(names)
+            names.sort()
+
+            for i in range(len(names)):
+                name = names[i]
+                ae = a.get(name, None)
+                be = b.get(name, None)
+                if not self.comparer.visit(ae, be):
+                    self.dumper.enter_member(name)
+                    self.visit(ae, be)
+                    self.dumper.leave_member(i == len(names) - 1)
+
+            self.dumper.leave_object()
+
+    def visitArray(self, a, b):
+        if not isinstance(b, list):
+            self.replace(a, b)
+        else:
+            self.dumper.enter_array()
+            max_len = max(len(a), len(b))
+            for i in range(max_len):
+                try:
+                    ae = a[i]
+                except IndexError:
+                    ae = None
+                try:
+                    be = b[i]
+                except IndexError:
+                    be = None
+                self.dumper._indent()
+                if self.comparer.visit(ae, be):
+                    self.dumper.visit(ae)
+                else:
+                    self.visit(ae, be)
+                if i != max_len - 1:
+                    self.dumper._write(',')
+                self.dumper._newline()
+
+            self.dumper.leave_array()
+
+    def visitValue(self, a, b):
+        if a != b:
+            self.replace(a, b)
+
+    def replace(self, a, b):
+        self.dumper.visit(a)
+        self.dumper._write(' -> ')
+        self.dumper.visit(b)
+
+
+#
+# Unfortunately JSON standard does not include comments, but this is a quite
+# useful feature to have on regressions tests
+#
+
+_token_res = [
+    r'//[^\r\n]*', # comment
+    r'"[^"\\]*(\\.[^"\\]*)*"', # string
+]
+
+_tokens_re = re.compile(r'|'.join(['(' + token_re + ')' for token_re in _token_res]), re.DOTALL)
+
+
+def _strip_comment(mo):
+    if mo.group(1):
+        return ''
+    else:
+        return mo.group(0)
+
+
+def _strip_comments(data):
+    '''Strip (non-standard) JSON comments.'''
+    return _tokens_re.sub(_strip_comment, data)
+
+
+assert _strip_comments('''// a comment
+"// a comment in a string
+"''') == '''
+"// a comment in a string
+"'''
+
+
+def load(stream, strip_images = True, strip_comments = True):
+    if strip_images:
+        object_hook = strip_object_hook
+    else:
+        object_hook = None
+    if strip_comments:
+        data = stream.read()
+        data = _strip_comments(data)
+        return json.loads(data, strict=False, object_hook = object_hook)
+    else:
+        return json.load(stream, strict=False, object_hook = object_hook)
+
+
+def main():
+    optparser = optparse.OptionParser(
+        usage="\n\t%prog [options] <ref_json> <src_json>")
+    optparser.add_option(
+        '--keep-images',
+        action="store_false", dest="strip_images", default=True,
+        help="compare images")
+
+    (options, args) = optparser.parse_args(sys.argv[1:])
+
+    if len(args) != 2:
+        optparser.error('incorrect number of arguments')
+
+    a = load(open(sys.argv[1], 'rt'), options.strip_images)
+    b = load(open(sys.argv[2], 'rt'), options.strip_images)
+
+    if False:
+        dumper = Dumper()
+        dumper.visit(a)
+
+    differ = Differ()
+    differ.visit(a, b)
+
+
+if __name__ == '__main__':
+    main()