Add pretty printing for rappor.xml

tools/metrics/rappor/pretty_print.py

Index: tools/metrics/rappor/pretty_print.py
diff --git a/tools/metrics/rappor/pretty_print.py b/tools/metrics/rappor/pretty_print.py
new file mode 100755
index 0000000000000000000000000000000000000000..8141d416f70d784afc62c1ddb730fecba5822abd
--- /dev/null
+++ b/tools/metrics/rappor/pretty_print.py
@@ -0,0 +1,162 @@
+#!/usr/bin/env python
+# Copyright 2015 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.
+
+import logging
+import sys
+import os
+
+# Import the metrics/common module for pretty print xml.
+sys.path.append(os.path.join(os.path.dirname(__file__), '..', 'common'))
+import models
+import presubmit_util
+
+
+# Model definitions for rappor.xml content
+summaryType = models.TextNodeType('summary')
+
+parametersType = models.ObjectNodeType('parameters',

[Alexei Svitkine (slow), 2015/02/13 14:36:37]

Nit: camelCase isn't a thing in Python style...

I think these should be CONSTANT_CASED, probably with leading _ to indicate
they're internal to this file.

[Steven Holte, 2015/02/14 01:11:18]

Done.

+    int_attributes=[
+      'num-cohorts',
+      'bytes',
+      'hash-functions',
+    ],
+    float_attributes=[
+      'fake-prob',
+      'fake-one-prob',
+      'one-coin-prob',
+      'zero-coin-prob',
+    ],
+    string_attributes=[
+      'reporting-level'
+    ])
+
+rapporParametersType = models.ObjectNodeType('rappor-parameters',
+    extra_newlines=(1, 1, 1),
+    string_attributes=['name'],
+    children=[
+      models.ChildType('summary', summaryType, False),
+      models.ChildType('parameters', parametersType, False),
+    ])
+
+rapporParameterTypesType = models.ObjectNodeType('rappor-parameter-types',
+    extra_newlines=(1, 1, 1),
+    dont_indent=True,
+    children=[
+      models.ChildType('types', rapporParametersType, True),
+    ])
+
+ownerType = models.TextNodeType('owner', single_line=True)
+
+rapporMetricType = models.ObjectNodeType('rappor-metric',
+    extra_newlines=(1, 1, 1),
+    string_attributes=['name', 'type'],
+    children=[
+      models.ChildType('owners', ownerType, True),
+      models.ChildType('summary', summaryType, False),
+    ])
+
+rapporMetricsType = models.ObjectNodeType('rappor-metrics',
+    extra_newlines=(1, 1, 1),
+    dont_indent=True,
+    children=[
+      models.ChildType('metrics', rapporMetricType, True),
+    ])
+
+rapporConfigurationType = models.ObjectNodeType('rappor-configuration',
+    dont_indent=True,
+    children=[
+      models.ChildType('parameterTypes', rapporParameterTypesType, False),
+      models.ChildType('metrics', rapporMetricsType, False),
+    ])
+
+rapporXmlType = models.DocumentType(rapporConfigurationType)
+
+
+def HasMissingOwners(metrics):

[Alexei Svitkine (slow), 2015/02/13 14:36:37]

Can you add some tests?

[Steven Holte, 2015/02/14 01:11:18]

Done.

+  """Check that all of the metrics have owners.
+
+  Args:
+    metrics: A list of rappor metric description objects.
+
+  Returns:
+    True iff some metrics are missing owners.
+  """
+  missing_owners = [m for m in metrics if not m['owners']]
+  for metric in missing_owners:
+    logging.error('Rappor metric "%s" is missing an owner.', metric['name'])
+    print metric
+  return bool(missing_owners)
+
+
+def HasInvalidTypes(type_names, metrics):
+  """Check that all of the metrics have valid types.
+
+  Args:
+    type_names: The set of valid type names.
+    metrics: A list of rappor metric description objects.
+
+  Returns:
+    True iff some metrics have invalid types.
+  """
+  invalid_types = [m for m in metrics if m['type'] not in type_names]
+  for metric in invalid_types:
+    logging.error('Rappor metric "%s" has invalid type "%s"',
+        metric['name'], metric['type'])
+  return bool(invalid_types)
+
+
+def HasErrors(obj):
+  """Check that rappor.xml passes some basic validation checks.
+
+  Args:
+    obj: The parsed rappor.xml contents.
+
+  Returns:
+    True iff there are validation errors.
+  """
+  metrics = obj['metrics']['metrics']
+  type_names = set(p['name'] for p in obj['parameterTypes']['types'])
+  return (HasMissingOwners(metrics) or
+          HasInvalidTypes(type_names, metrics))
+
+
+def Cleanup(obj):
+  """Preform cleanup on description contents, such as sorting metrics.
+
+  Args:
+    obj: The parsed rappor.xml contents.
+  """
+  types = obj['parameterTypes']['types']
+  types.sort(key=lambda x: x['name'])
+  metrics = obj['metrics']['metrics']
+  metrics.sort(key=lambda x: x['name'])
+
+
+def UpdateXml(original_xml):
+  """Parse the original xml and return a pretty printed version.
+
+  Args:
+    original_xml: A string containing the original xml file contents.
+
+  Returns:
+    A Pretty printed xml string.
+  """
+  comments, obj = rapporXmlType.Parse(original_xml)
+
+  if HasErrors(obj):
+    return None
+
+  Cleanup(obj)
+
+  return rapporXmlType.PrettyPrint(comments, obj)
+
+
+def main(argv):
+  presubmit_util.DoPresubmitMain(argv, 'rappor.xml', 'rappor.old.xml',
+                                 'pretty_print.py', UpdateXml)
+
+
+if '__main__' == __name__:
+  sys.exit(main(sys.argv))