def load_definition(cls, xml_object, system, location):
'''Load a descriptor from the specified xml_object:
If there is a filename attribute, load it as a string, and
log a warning if it is not parseable by etree.HTMLParser.
If there is not a filename attribute, the definition is the body
of the xml_object, without the root tag (do not want <html> in the
middle of a page)
'''
filename = xml_object.get('filename')
if filename is None:
definition_xml = copy.deepcopy(xml_object)
cls.clean_metadata_from_xml(definition_xml)
return {'data': stringify_children(definition_xml)}, []
else:
# html is special. cls.filename_extension is 'xml', but
# if 'filename' is in the definition, that means to load
# from .html
# 'filename' in html pointers is a relative path
# (not same as 'html/blah.html' when the pointer is in a directory itself)
pointer_path = "{category}/{url_path}".format(
category='html', url_path=name_to_pathname(location.name))
base = path(pointer_path).dirname()
# log.debug("base = {0}, base.dirname={1}, filename={2}".format(base, base.dirname(), filename))
filepath = "{base}/{name}.html".format(base=base, name=filename)
# log.debug("looking for html file for {0} at {1}".format(location, filepath))
# VS[compat]
# TODO (cpennington): If the file doesn't exist at the right path,
# give the class a chance to fix it up. The file will be written out
# again in the correct format. This should go away once the CMS is
# 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))
for candidate in candidates:
if system.resources_fs.exists(candidate):
filepath = candidate
break
try:
with system.resources_fs.open(filepath) as file:
html = file.read().decode('utf-8')
# Log a warning if we can't parse the file, but don't error
if not check_html(html) and len(html) > 0:
msg = "Couldn't parse html in {0}, content = {1}".format(
filepath, html)
log.warning(msg)
system.error_tracker("Warning: " + msg)
definition = {'data': html}
# TODO (ichuang): remove this after migration
# for Fall 2012 LMS migration: keep filename (and unmangled filename)
definition['filename'] = [filepath, filename]
return definition, []
except (ResourceNotFoundError) as err:
msg = 'Unable to load file contents at path {0}: {1} '.format(
filepath, err)
# add more info and re-raise
raise Exception(msg), None, sys.exc_info()[2]
def load_definition(cls, xml_object, system, location, id_generator):
'''Load a descriptor from the specified xml_object:
If there is a filename attribute, load it as a string, and
log a warning if it is not parseable by etree.HTMLParser.
If there is not a filename attribute, the definition is the body
of the xml_object, without the root tag (do not want <html> in the
middle of a page)
Args:
xml_object: an lxml.etree._Element containing the definition to load
system: the modulestore system or runtime which caches data
location: the usage id for the block--used to compute the filename if none in the xml_object
id_generator: used by other impls of this method to generate the usage_id
'''
filename = xml_object.get('filename')
if filename is None:
definition_xml = copy.deepcopy(xml_object)
cls.clean_metadata_from_xml(definition_xml)
return {'data': stringify_children(definition_xml)}, []
else:
# html is special. cls.filename_extension is 'xml', but
# if 'filename' is in the definition, that means to load
# from .html
# 'filename' in html pointers is a relative path
# (not same as 'html/blah.html' when the pointer is in a directory itself)
pointer_path = "{category}/{url_path}".format(
category='html',
url_path=name_to_pathname(location.name)
)
base = path(pointer_path).dirname()
# log.debug("base = {0}, base.dirname={1}, filename={2}".format(base, base.dirname(), filename))
filepath = "{base}/{name}.html".format(base=base, name=filename)
# log.debug("looking for html file for {0} at {1}".format(location, filepath))
# VS[compat]
# TODO (cpennington): If the file doesn't exist at the right path,
# give the class a chance to fix it up. The file will be written out
# again in the correct format. This should go away once the CMS is
# 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))
for candidate in candidates:
if system.resources_fs.exists(candidate):
filepath = candidate
break
try:
with system.resources_fs.open(filepath) as file:
html = file.read().decode('utf-8')
# Log a warning if we can't parse the file, but don't error
if not check_html(html) and len(html) > 0:
msg = "Couldn't parse html in {0}, content = {1}".format(filepath, html)
log.warning(msg)
system.error_tracker("Warning: " + msg)
definition = {'data': html}
# TODO (ichuang): remove this after migration
# for Fall 2012 LMS migration: keep filename (and unmangled filename)
definition['filename'] = [filepath, filename]
return definition, []
except (ResourceNotFoundError) as err:
msg = 'Unable to load file contents at path {0}: {1} '.format(
filepath, err)
# add more info and re-raise
raise Exception(msg), None, sys.exc_info()[2]