From 96b4914fef3b530f2f85cdd8aee8adfaec8ced25 Mon Sep 17 00:00:00 2001
From: cahrens <christina@edx.org>
Date: Thu, 15 Aug 2013 11:03:31 -0400
Subject: [PATCH] Make our deserializing consistent with serializing.

Also fixes problems with double-quoted strings (historical artifact).

STUD-640

Conflicts:

	common/lib/xmodule/xmodule/video_module.py
---
 .../lib/xmodule/xmodule/tests/test_video.py   | 56 +++++++++++++++++++
 common/lib/xmodule/xmodule/video_module.py    | 32 +++++++----
 common/lib/xmodule/xmodule/xml_module.py      |  5 ++
 3 files changed, 82 insertions(+), 11 deletions(-)

diff --git a/common/lib/xmodule/xmodule/tests/test_video.py b/common/lib/xmodule/xmodule/tests/test_video.py
index ac1ab1d247..5076f39c79 100644
--- a/common/lib/xmodule/xmodule/tests/test_video.py
+++ b/common/lib/xmodule/xmodule/tests/test_video.py
@@ -264,6 +264,62 @@ class VideoDescriptorImportTestCase(unittest.TestCase):
             'data': ''
         })
 
+    def test_from_xml_double_quotes(self):
+        """
+        Make sure we can handle the double-quoted string format (which was used for exporting for
+        a few weeks).
+        """
+        module_system = DummySystem(load_error_modules=True)
+        xml_data ='''
+            <video display_name="&quot;display_name&quot;"
+                html5_sources="[&quot;source_1&quot;, &quot;source_2&quot;]"
+                show_captions="false"
+                source="&quot;http://download_video&quot;"
+                sub="&quot;html5_subtitles&quot;"
+                track="&quot;http://download_track&quot;"
+                youtube_id_0_75="&quot;OEoXaMPEzf65&quot;"
+                youtube_id_1_25="&quot;OEoXaMPEzf125&quot;"
+                youtube_id_1_5="&quot;OEoXaMPEzf15&quot;"
+                youtube_id_1_0="&quot;OEoXaMPEzf10&quot;"
+                />
+        '''
+        output = VideoDescriptor.from_xml(xml_data, module_system)
+        self.assert_attributes_equal(output, {
+            'youtube_id_0_75': 'OEoXaMPEzf65',
+            'youtube_id_1_0': 'OEoXaMPEzf10',
+            'youtube_id_1_25': 'OEoXaMPEzf125',
+            'youtube_id_1_5': 'OEoXaMPEzf15',
+            'show_captions': False,
+            'start_time': 0.0,
+            'end_time': 0.0,
+            'track': 'http://download_track',
+            'source': 'http://download_video',
+            'html5_sources': ["source_1", "source_2"],
+            'data': ''
+        })
+
+    def test_from_xml_double_quote_concatenated_youtube(self):
+        module_system = DummySystem(load_error_modules=True)
+        xml_data = '''
+            <video display_name="Test Video"
+                   youtube="1.0:&quot;p2Q6BrNhdh8&quot;,1.25:&quot;1EeWXzPdhSA&quot;">
+            </video>
+        '''
+        output = VideoDescriptor.from_xml(xml_data, module_system)
+        self.assert_attributes_equal(output, {
+            'youtube_id_0_75': '',
+            'youtube_id_1_0': 'p2Q6BrNhdh8',
+            'youtube_id_1_25': '1EeWXzPdhSA',
+            'youtube_id_1_5': '',
+            'show_captions': True,
+            'start_time': 0.0,
+            'end_time': 0.0,
+            'track': '',
+            'source': '',
+            'html5_sources': [],
+            'data': ''
+        })
+
     def test_old_video_format(self):
         """
         Test backwards compatibility with VideoModule's XML format.
diff --git a/common/lib/xmodule/xmodule/video_module.py b/common/lib/xmodule/xmodule/video_module.py
index e0fae8cdbe..805ca48930 100644
--- a/common/lib/xmodule/xmodule/video_module.py
+++ b/common/lib/xmodule/xmodule/video_module.py
@@ -291,19 +291,19 @@ class VideoDescriptor(VideoFields, TabsEditingDescriptor, EmptyDataRawDescriptor
         XML-based courses.
         """
         ret = {'0.75': '', '1.00': '', '1.25': '', '1.50': ''}
-        if data == '':
-            return ret
+
         videos = data.split(',')
         for video in videos:
             pieces = video.split(':')
-            # HACK
-            # To elaborate somewhat: in many LMS tests, the keys for
-            # Youtube IDs are inconsistent. Sometimes a particular
-            # speed isn't present, and formatting is also inconsistent
-            # ('1.0' versus '1.00'). So it's necessary to either do
-            # something like this or update all the tests to work
-            # properly.
-            ret['%.2f' % float(pieces[0])] = pieces[1]
+            try:
+                speed = '%.2f' % float(pieces[0])  # normalize speed
+                # Handle the fact that youtube IDs got double-quoted for a period of time.
+                # Note: we pass in "VideoFields.youtube_id_1_0" so we deserialize as a String--
+                # it doesn't matter what the actual speed is for the purposes of deserializing.
+                youtube_id = VideoDescriptor._deserialize(VideoFields.youtube_id_1_0.name, pieces[1])
+                ret[speed] = youtube_id
+            except (ValueError, IndexError):
+                log.warning('Invalid YouTube ID: %s' % video)
         return ret
 
     @staticmethod
@@ -316,7 +316,6 @@ class VideoDescriptor(VideoFields, TabsEditingDescriptor, EmptyDataRawDescriptor
         model_data = {}
 
         conversions = {
-            'show_captions': json.loads,
             'start_time': VideoDescriptor._parse_time,
             'end_time': VideoDescriptor._parse_time
         }
@@ -355,10 +354,21 @@ class VideoDescriptor(VideoFields, TabsEditingDescriptor, EmptyDataRawDescriptor
                 #  Convert XML attrs into Python values.
                 if attr in conversions:
                     value = conversions[attr](value)
+                else:
+                # We export values with json.dumps (well, except for Strings, but
+                # for about a month we did it for Strings also).
+                    value = VideoDescriptor._deserialize(attr, value)
                 model_data[attr] = value
 
         return model_data
 
+    @classmethod
+    def _deserialize(cls, attr, value):
+        """
+        Handles deserializing values that may have been encoded with json.dumps.
+        """
+        return cls.get_map_for_field(attr).from_xml(value)
+
     @staticmethod
     def _parse_time(str_time):
         """Converts s in '12:34:45' format to seconds. If s is
diff --git a/common/lib/xmodule/xmodule/xml_module.py b/common/lib/xmodule/xmodule/xml_module.py
index 5b8d2c8aee..b0b7f300ed 100644
--- a/common/lib/xmodule/xmodule/xml_module.py
+++ b/common/lib/xmodule/xmodule/xml_module.py
@@ -167,6 +167,11 @@ class XmlDescriptor(XModuleDescriptor):
 
     @classmethod
     def get_map_for_field(cls, attr):
+        """
+        Returns a serialize/deserialize AttrMap for the given field of a class.
+
+        Searches through fields defined by cls to find one named attr.
+        """
         for field in set(cls.fields + cls.lms.fields):
             if field.name == attr:
                 from_xml = lambda val: deserialize_field(field, val)