Make initial import-export tests pass.

TODO:
* need unique slugs for errors so they don't overwrite each other on export.
  - try to preserve origin slug.  If not possible, generate random one.

* figure out what metadata to strip.  e.g.
 ({'data': '<problem><p><h1>Finger Exercise 1...'},
    {'data': '<problem points="1" type="lecture"><p><h1>Finger Exercise 1...'},
    False)
  - where did points and type come from?  Do we want them there?


* separate broken and non-broken test courses

common/lib/xmodule/xmodule/html_module.py

@@ -36,7 +36,6 @@ class HtmlDescriptor(XmlDescriptor, EditingDescriptor):
     # are being edited in the cms
     @classmethod
     def backcompat_paths(cls, path):
-        origpath = path
         if path.endswith('.html.xml'):
             path = path[:-9] + '.html'  #backcompat--look for html instead of xml
         candidates = []
@@ -45,9 +44,11 @@ class HtmlDescriptor(XmlDescriptor, EditingDescriptor):
             _, _, path = path.partition(os.sep)
 
         # also look for .html versions instead of .xml
-        if origpath.endswith('.xml'):
-            candidates.append(origpath[:-4] + '.html')
-        return candidates
+        nc = []
+        for candidate in candidates:
+            if candidate.endswith('.xml'):
+                nc.append(candidate[:-4] + '.html')
+        return candidates + nc
 
     # NOTE: html descriptors are special.  We do not want to parse and
     # export them ourselves, because that can break things (e.g. lxml
@@ -80,7 +81,7 @@ class HtmlDescriptor(XmlDescriptor, EditingDescriptor):
             # online and has imported all current (fall 2012) courses from xml
             if not system.resources_fs.exists(filepath):
                 candidates = cls.backcompat_paths(filepath)
-                #log.debug("candidates = {0}".format(candidates))
+                log.debug("candidates = {0}".format(candidates))
                 for candidate in candidates:
                     if system.resources_fs.exists(candidate):
                         filepath = candidate

common/lib/xmodule/xmodule/tests/test_export.py

+import unittest
+
 from fs.osfs import OSFS
 from nose.tools import assert_equals, assert_true
-from nose import SkipTest
 from path import path
 from tempfile import mkdtemp
 
@@ -18,10 +19,7 @@ DATA_DIR = TEST_DIR / 'data'
 
 def strip_metadata(descriptor, key):
     """
-    HACK: data_dir metadata tags break equality because they aren't real metadata
-    remove them.
-
-    Recursively strips same tag from all children.
+    Recursively strips tag from all children.
     """
     print "strip {key} from {desc}".format(key=key, desc=descriptor.location.url())
     descriptor.metadata.pop(key, None)
@@ -35,50 +33,66 @@ def check_gone(descriptor, key):
     for d in descriptor.get_children():
         check_gone(d, key)
 
-def check_export_roundtrip(data_dir, course_dir):
-    print "Starting import"
-    initial_import = XMLModuleStore(data_dir, eager=True, course_dirs=[course_dir])
 
-    courses = initial_import.get_courses()
-    assert_equals(len(courses), 1)
-    initial_course = courses[0]
 
-    print "Starting export"
-    export_dir = mkdtemp()
-    print "export_dir: {0}".format(export_dir)
-    fs = OSFS(export_dir)
-    export_course_dir = 'export'
-    export_fs = fs.makeopendir(export_course_dir)
+class RoundTripTestCase(unittest.TestCase):
+    '''Check that our test courses roundtrip properly'''
+    def check_export_roundtrip(self, data_dir, course_dir):
+        print "Starting import"
+        initial_import = XMLModuleStore(data_dir, eager=True, course_dirs=[course_dir])
+
+        courses = initial_import.get_courses()
+        self.assertEquals(len(courses), 1)
+        initial_course = courses[0]
+
+        print "Starting export"
+        export_dir = mkdtemp()
+        print "export_dir: {0}".format(export_dir)
+        fs = OSFS(export_dir)
+        export_course_dir = 'export'
+        export_fs = fs.makeopendir(export_course_dir)
+
+        xml = initial_course.export_to_xml(export_fs)
+        with export_fs.open('course.xml', 'w') as course_xml:
+            course_xml.write(xml)
+
+        print "Starting second import"
+        second_import = XMLModuleStore(export_dir, eager=True,
+                                       course_dirs=[export_course_dir])
+
+        courses2 = second_import.get_courses()
+        self.assertEquals(len(courses2), 1)
+        exported_course = courses2[0]
 
-    xml = initial_course.export_to_xml(export_fs)
-    with export_fs.open('course.xml', 'w') as course_xml:
-        course_xml.write(xml)
+        print "Checking course equality"
+        # HACK: data_dir metadata tags break equality because they
+        # aren't real metadata, and depend on paths. Remove them.
+        strip_metadata(initial_course, 'data_dir')
+        strip_metadata(exported_course, 'data_dir')
 
-    print "Starting second import"
-    second_import = XMLModuleStore(export_dir, eager=True,
-                                   course_dirs=[export_course_dir])
+        self.assertEquals(initial_course, exported_course)
 
-    courses2 = second_import.get_courses()
-    assert_equals(len(courses2), 1)
-    exported_course = courses2[0]
+        print "Checking key equality"
+        self.assertEquals(sorted(initial_import.modules.keys()),
+                          sorted(second_import.modules.keys()))
 
-    print "Checking course equality"
-    strip_metadata(initial_course, 'data_dir')
-    strip_metadata(exported_course, 'data_dir')
+        print "Checking module equality"
+        for location in initial_import.modules.keys():
+            print "Checking", location
+            if location.category == 'html':
+                print ("Skipping html modules--they can't import in"
+                       " final form without writing files...")
+                continue
+            self.assertEquals(initial_import.modules[location],
+                              second_import.modules[location])
 
-    assert_equals(initial_course, exported_course)
 
-    print "Checking key equality"
-    assert_equals(initial_import.modules.keys(), second_import.modules.keys())
+    def setUp(self):
+        self.maxDiff = None
 
-    print "Checking module equality"
-    for location in initial_import.modules.keys():
-        print "Checking", location
-        assert_equals(initial_import.modules[location], second_import.modules[location])
+    def test_toy_roundtrip(self):
+        self.check_export_roundtrip(DATA_DIR, "toy")
 
 
-def test_toy_roundtrip():
-    dir = ""
-    # TODO: add paths and make this run.
-    raise SkipTest()
-    check_export_roundtrip(dir)
+    def test_full_roundtrip(self):
+        self.check_export_roundtrip(DATA_DIR, "full")

common/lib/xmodule/xmodule/tests/test_import.py

@@ -51,8 +51,6 @@ class DummySystem(XMLParsingSystem):
 
 class ImportTestCase(unittest.TestCase):
     '''Make sure module imports work properly, including for malformed inputs'''
-
-
     @staticmethod
     def get_system():
         '''Get a dummy system'''

common/lib/xmodule/xmodule/xml_module.py

@@ -76,7 +76,7 @@ class XmlDescriptor(XModuleDescriptor):
 
     metadata_to_strip = ('data_dir',
            # VS[compat] -- remove the below attrs once everything is in the CMS
-           'course', 'org', 'url_name')
+           'course', 'org', 'url_name', 'filename')
 
     # A dictionary mapping xml attribute names AttrMaps that describe how
     # to import and export them

common/test/data/simple/course.xml

@@ -2,7 +2,7 @@
   <chapter name="Overview">
     <video name="Welcome" youtube="0.75:izygArpw-Qo,1.0:p2Q6BrNhdh8,1.25:1EeWXzPdhSA,1.50:rABDYkeK0x8"/>
     <videosequence format="Lecture Sequence" name="A simple sequence">
-      <html id="toylab" filename="toylab"/>
+      <html name="toylab" filename="toylab"/>
       <video name="S0V1: Video Resources" youtube="0.75:EuzkdzfR0i8,1.0:1bK-WdDi6Qw,1.25:0v1VzoDVUTM,1.50:Bxk_-ZJb240"/>
     </videosequence>
     <section name="Lecture 2">