Skip to content
Snippets Groups Projects
Commit 9254cfa9 authored by Don Mitchell's avatar Don Mitchell
Browse files

Import depth first to help w/ publishing

parent 2726b437
No related merge requests found
Hide whitespace changes
Inline Side-by-side
common/lib/xmodule/xmodule/modulestore/xml_importer.py
+ 94
− 77
View file @ 9254cfa9
......@@ -11,8 +11,8 @@ can't represent that virtual state (2nd row in table)
In the table body, the tuples represent virtual modulestore result. The row headers represent the pre-import
modulestore state.
Modulestore virtual \ XML physical (draft, published)
(draft, published) \ (-, -) | (x, -) | (x, x) | (x, y) | (-, x)
Modulestore virtual | XML physical (draft, published)
(draft, published) | (-, -) | (x, -) | (x, x) | (x, y) | (-, x)
----------------------+--------------------------------------------
(-, -) | (-, -) | (x, -) | (x, x) | (x, y) | (-, x)
(-, a) | (-, a) | (x, a) | (x, x) | (x, y) | (-, x) : deleted from draft before import
......@@ -207,11 +207,12 @@ def import_from_xml(
continue
with store.bulk_write_operations(dest_course_id):
source_course = xml_module_store.get_course(course_key)
# STEP 1: find and import course module
course, course_data_path = _import_course_module(
xml_module_store, store, runtime, user_id,
data_dir, course_key, dest_course_id, do_import_static,
verbose
store, runtime, user_id,
data_dir, course_key, dest_course_id, source_course,
do_import_static, verbose
)
new_courses.append(course)
......@@ -221,18 +222,39 @@ def import_from_xml(
)
# STEP 3: import PUBLISHED items
# now loop through all the modules
# now loop through all the modules depth first and then orphans
with store.branch_setting(ModuleStoreEnum.Branch.published_only, dest_course_id):
for module in xml_module_store.modules[course_key].itervalues():
if module.scope_ids.block_type == 'course':
# we've already saved the course module up above
continue
all_locs = set(xml_module_store.modules[course_key].keys())
all_locs.remove(source_course.location)
def depth_first(subtree):
"""
Import top down just so import code can make assumptions about parents always being available
"""
if subtree.has_children:
for child in subtree.get_children():
all_locs.remove(child.location)
if verbose:
log.debug('importing module location {loc}'.format(loc=child.location))
_import_module_and_update_references(
child, store,
user_id,
course_key,
dest_course_id,
do_import_static=do_import_static,
runtime=course.runtime
)
depth_first(child)
depth_first(source_course)
for leftover in all_locs:
if verbose:
log.debug('importing module location {loc}'.format(loc=module.location))
log.debug('importing module location {loc}'.format(loc=leftover))
_import_module_and_update_references(
module, store,
xml_module_store.get_item(leftover), store,
user_id,
course_key,
dest_course_id,
......@@ -256,7 +278,7 @@ def import_from_xml(
def _import_course_module(
xml_module_store, store, runtime, user_id, data_dir, course_key, dest_course_id, do_import_static,
store, runtime, user_id, data_dir, course_key, dest_course_id, source_course, do_import_static,
verbose,
):
if verbose:
......@@ -265,70 +287,65 @@ def _import_course_module(
# Quick scan to get course module as we need some info from there.
# Also we need to make sure that the course module is committed
# first into the store
for module in xml_module_store.modules[course_key].itervalues():
if module.scope_ids.block_type == 'course':
course_data_path = path(data_dir) / module.data_dir
log.debug(u'======> IMPORTING course {course_key}'.format(
course_key=course_key,
))
if not do_import_static:
# for old-style xblock where this was actually linked to kvs
module.static_asset_path = module.data_dir
module.save()
log.debug('course static_asset_path={path}'.format(
path=module.static_asset_path
))
log.debug('course data_dir={0}'.format(module.data_dir))
course_data_path = path(data_dir) / source_course.data_dir
log.debug(u'======> IMPORTING course {course_key}'.format(
course_key=course_key,
))
if not do_import_static:
# for old-style xblock where this was actually linked to kvs
source_course.static_asset_path = source_course.data_dir
source_course.save()
log.debug('course static_asset_path={path}'.format(
path=source_course.static_asset_path
))
log.debug('course data_dir={0}'.format(source_course.data_dir))
course = _import_module_and_update_references(
source_course, store, user_id,
course_key,
dest_course_id,
do_import_static=do_import_static,
runtime=runtime,
)
course = _import_module_and_update_references(
module, store, user_id,
course_key,
dest_course_id,
do_import_static=do_import_static,
runtime=runtime,
for entry in course.pdf_textbooks:
for chapter in entry.get('chapters', []):
if StaticContent.is_c4x_path(chapter.get('url', '')):
asset_key = StaticContent.get_location_from_path(chapter['url'])
chapter['url'] = StaticContent.get_static_path_from_location(asset_key)
# Original wiki_slugs had value location.course. To make them unique this was changed to 'org.course.name'.
# If we are importing into a course with a different course_id and wiki_slug is equal to either of these default
# values then remap it so that the wiki does not point to the old wiki.
if course_key != course.id:
original_unique_wiki_slug = u'{0}.{1}.{2}'.format(
course_key.org,
course_key.course,
course_key.run
)
if course.wiki_slug == original_unique_wiki_slug or course.wiki_slug == course_key.course:
course.wiki_slug = u'{0}.{1}.{2}'.format(
course.id.org,
course.id.course,
course.id.run,
)
for entry in course.pdf_textbooks:
for chapter in entry.get('chapters', []):
if StaticContent.is_c4x_path(chapter.get('url', '')):
asset_key = StaticContent.get_location_from_path(chapter['url'])
chapter['url'] = StaticContent.get_static_path_from_location(asset_key)
# Original wiki_slugs had value location.course. To make them unique this was changed to 'org.course.name'.
# If we are importing into a course with a different course_id and wiki_slug is equal to either of these default
# values then remap it so that the wiki does not point to the old wiki.
if course_key != course.id:
original_unique_wiki_slug = u'{0}.{1}.{2}'.format(
course_key.org,
course_key.course,
course_key.run
)
if course.wiki_slug == original_unique_wiki_slug or course.wiki_slug == course_key.course:
course.wiki_slug = u'{0}.{1}.{2}'.format(
course.id.org,
course.id.course,
course.id.run,
)
# cdodge: more hacks (what else). Seems like we have a
# problem when importing a course (like 6.002) which
# does not have any tabs defined in the policy file.
# The import goes fine and then displays fine in LMS,
# but if someone tries to add a new tab in the CMS, then
# the LMS barfs because it expects that -- if there are
# *any* tabs -- then there at least needs to be
# some predefined ones
if course.tabs is None or len(course.tabs) == 0:
CourseTabList.initialize_default(course)
store.update_item(course, user_id)
return course, course_data_path
# cdodge: more hacks (what else). Seems like we have a
# problem when importing a course (like 6.002) which
# does not have any tabs defined in the policy file.
# The import goes fine and then displays fine in LMS,
# but if someone tries to add a new tab in the CMS, then
# the LMS barfs because it expects that -- if there are
# *any* tabs -- then there at least needs to be
# some predefined ones
if course.tabs is None or len(course.tabs) == 0:
CourseTabList.initialize_default(course)
# raise an exception if the course wasn't found
raise Exception("Course module not found in imported modules")
store.update_item(course, user_id)
return course, course_data_path
def _import_static_content_wrapper(static_content_store, do_import_static, course_data_path, dest_course_id, verbose):
......@@ -524,8 +541,8 @@ def _import_course_draft(
# attributes (they are normally in the parent object,
# aka sequential), so we have to replace the location.name
# with the XML filename that is part of the pack
fn, __ = os.path.splitext(filename)
descriptor.location = descriptor.location.replace(name=fn)
filename, __ = os.path.splitext(filename)
descriptor.location = descriptor.location.replace(name=filename)
index = int(descriptor.xml_attributes['index_in_children_list'])
if index in drafts:
......@@ -566,7 +583,7 @@ def _import_course_draft(
sequential = store.get_item(seq_location, depth=0)
non_draft_location = module.location.map_into_course(target_course_id)
if non_draft_location not in sequential.children:
if not any(child.block_id == module.location.block_id for child in sequential.children):
sequential.children.insert(index, non_draft_location)
store.update_item(sequential, user_id)
......
common/lib/xmodule/xmodule/tests/__init__.py
+ 6
− 38
View file @ 9254cfa9
......@@ -213,9 +213,9 @@ class CourseComparisonTest(unittest.TestCase):
else:
revision = None
actual_items = actual_store.get_items(actual_course_key, revision=revision)
self._assertCoursesEqual(expected_items, actual_items, actual_course_key)
self._assertCoursesEqual(expected_items, actual_items, actual_course_key, expect_drafts=True)
def _assertCoursesEqual(self, expected_items, actual_items, actual_course_key):
def _assertCoursesEqual(self, expected_items, actual_items, actual_course_key, expect_drafts=False):
self.assertEqual(len(expected_items), len(actual_items))
actual_item_map = {
......@@ -264,7 +264,6 @@ class CourseComparisonTest(unittest.TestCase):
# compare children
self.assertEqual(expected_item.has_children, actual_item.has_children)
if expected_item.has_children:
expect_drafts = getattr(expected_item, 'is_draft', getattr(actual_item, 'is_draft', False))
expected_children = [
course1_item_child.location.map_into_course(actual_item.location.course_key)
for course1_item_child in expected_item.get_children()
......@@ -272,10 +271,10 @@ class CourseComparisonTest(unittest.TestCase):
if expect_drafts or not getattr(course1_item_child, 'is_draft', False)
]
actual_children = [
item_child.location
for item_child in actual_item.get_children()
# get_children was returning drafts for published parents :-(
if expect_drafts or not getattr(item_child, 'is_draft', False)
item_child.location
for item_child in actual_item.get_children()
# get_children was returning drafts for published parents :-(
if expect_drafts or not getattr(item_child, 'is_draft', False)
]
try:
self.assertEqual(expected_children, actual_children)
......@@ -331,34 +330,3 @@ class CourseComparisonTest(unittest.TestCase):
actual_thumbs = actual_store.get_all_content_thumbnails_for_course(actual_course_key)
self._assertAssetsEqual(expected_course_key, expected_thumbs, actual_course_key, actual_thumbs)
def compute_real_state(self, store, item):
"""
In draft mongo, compute_published_state can return draft when the draft == published, but in split,
it'll return public in that case
"""
supposed_state = store.compute_publish_state(item)
if supposed_state == PublishState.draft and isinstance(item.runtime.modulestore, DraftModuleStore):
# see if the draft differs from the published
published = store.get_item(item.location, revision=ModuleStoreEnum.RevisionOption.published_only)
if item.get_explicitly_set_fields_by_scope() != published.get_explicitly_set_fields_by_scope():
# checking content: if published differs from item, return draft
return supposed_state
if item.get_explicitly_set_fields_by_scope(Scope.settings) != published.get_explicitly_set_fields_by_scope(Scope.settings):
# checking settings: if published differs from item, return draft
return supposed_state
if item.has_children and item.children != published.children:
# checking children: if published differs from item, return draft
return supposed_state
# published == item in all respects, so return public
return PublishState.public
elif supposed_state == PublishState.public and item.location.category in DIRECT_ONLY_CATEGORIES:
if not all([
store.has_item(child_loc, revision=ModuleStoreEnum.RevisionOption.draft_only)
for child_loc in item.children
]):
return PublishState.draft
else:
return supposed_state
else:
return supposed_state
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment