chore: add pointer tag OLX support for WordCloud XBlock

Author: ttqureshi

xblocks_contrib/word_cloud/word_cloud.py

@@ -5,16 +5,34 @@
 If student does not yet answered - `num_inputs` numbers of text inputs.
 If student have answered - words he entered and cloud.
 """
+import copy
+import json
+import logging
+import os
 import uuid
 
 from django.utils.translation import gettext_noop as _
+from lxml import etree
+from lxml.etree import ElementTree
 from web_fragments.fragment import Fragment
-from xblock.core import XBlock
-from xblock.fields import Boolean, Dict, Integer, List, Scope, String
+from xblock.core import XML_NAMESPACES, XBlock
+from xblock.fields import Boolean, Dict, Integer, List, Scope, ScopeIds, String
 from xblock.utils.resources import ResourceLoader
 from xblock.utils.studio_editable import StudioEditableXBlockMixin
 
+from xblocks_contrib.common.xml_utils import (
+    apply_pointer_attributes,
+    deserialize_field,
+    format_filepath,
+    is_pointer_tag,
+    load_definition_xml,
+    name_to_pathname,
+    own_metadata,
+    serialize_field,
+)
+
 resource_loader = ResourceLoader(__name__)
+log = logging.getLogger(__name__)
 
 
 def pretty_bool(value):
@@ -27,8 +45,270 @@ def pretty_bool(value):
     return value in bool_dict
 
 
+class WordCloudXmlMixin:
+    """
+    A mixin class to add XML parsing functionality to WordCloud.
+    """
+
+    metadata_to_strip = (
+        'data_dir',
+        'tabs',
+        'grading_policy',
+        'discussion_blackouts',
+        'course',
+        'org',
+        'url_name',
+        'filename',
+        'xml_attributes',
+        "x-is-pointer-node",
+    )
+
+    xml_attributes = Dict(
+        help="Map of unhandled xml attributes, used only for storage between import and export",
+        default={},
+        scope=Scope.settings
+    )
+
+    @classmethod
+    def apply_policy(cls, metadata, policy):
+        """
+        Add the keys in policy to metadata, after processing them
+        through the attrmap.  Updates the metadata dict in place.
+        """
+        for attr, value in policy.items():
+            if attr not in cls.fields:
+                # Store unknown attributes coming from policy.json
+                # in such a way that they will export to xml unchanged
+                metadata['xml_attributes'][attr] = value
+            else:
+                metadata[attr] = value
+
+    @staticmethod
+    def _get_metadata_from_xml(xml_object, remove=True):
+        """
+        Extract the metadata from the XML.
+        """
+        meta = xml_object.find('meta')
+        if meta is None:
+            return ''
+        dmdata = meta.text
+        if remove:
+            xml_object.remove(meta)
+        return dmdata
+
+    @classmethod
+    def clean_metadata_from_xml(cls, xml_object, excluded_fields=()):
+        """
+        Remove any attribute named for a field with scope Scope.settings from the supplied
+        xml_object
+        """
+        for field_name, field in cls.fields.items():
+            if (field.scope == Scope.settings
+                    and field_name not in excluded_fields
+                    and xml_object.get(field_name) is not None):
+                del xml_object.attrib[field_name]
+
+    @classmethod
+    def load_definition(cls, xml_object):
+        """
+        Load a block from the specified xml_object.
+
+        Args:
+            xml_object: an lxml.etree._Element containing the definition to load
+        """
+
+        filename = xml_object.get('filename')
+        definition_xml = copy.deepcopy(xml_object)
+        filepath = ''
+
+        definition_metadata = cls._get_metadata_from_xml(definition_xml)
+        cls.clean_metadata_from_xml(definition_xml)
+
+        if len(xml_object) == 0 and len(list(xml_object.items())) == 0:
+            definition, children = {'data': ''}, []
+        else:
+            definition, children = {'data': etree.tostring(xml_object, pretty_print=True, encoding='unicode')}, []
+
+        if definition_metadata:
+            definition['definition_metadata'] = definition_metadata
+        definition['filename'] = [filepath, filename]
+
+        return definition, children
+
+    @classmethod
+    def load_metadata(cls, xml_object):
+        """
+        Read the metadata attributes from this xml_object.
+
+        Returns a dictionary {key: value}.
+        """
+        metadata = {'xml_attributes': {}}
+        for attr, val in xml_object.attrib.items():
+
+            if attr in cls.metadata_to_strip:
+                # don't load these
+                continue
+
+            if attr not in cls.fields:
+                metadata['xml_attributes'][attr] = val
+            else:
+                metadata[attr] = deserialize_field(cls.fields[attr], val)
+        return metadata
+
+    @classmethod
+    def parse_xml(cls, node, runtime, keys):
+        """
+        Use `node` to construct a new block.
+
+        Arguments:
+            node (etree.Element): The xml node to parse into an xblock.
+
+            runtime (:class:`.Runtime`): The runtime to use while parsing.
+
+            keys (:class:`.ScopeIds`): The keys identifying where this block
+                will store its data.
+
+        Returns (XBlock): The newly parsed XBlock
+
+        """
+        if keys is None:
+            # Passing keys=None is against the XBlock API but some platform tests do it.
+            def_id = runtime.id_generator.create_definition(node.tag, node.get("url_name"))
+            keys = ScopeIds(None, node.tag, def_id, runtime.id_generator.create_usage(def_id))
+        aside_children = []
+
+        # Let the runtime construct the block. It will have a proper, inheritance-aware field data store.
+        block = runtime.construct_xblock_from_class(cls, keys)
+
+        # VS[compat]
+        # In 2012, when the platform didn't have CMS, and all courses were handwritten XML files, problem tags
+        # contained XML problem descriptions withing themselves. Later, when Studio has been created, and "pointer" tags
+        # became the preferred problem format, edX has to add this compatibility code to 1) support both pre- and
+        # post-Studio course formats simulteneously, and 2) be able to migrate 2012-fall courses to Studio. Old style
+        # support supposed to be removed, but the deprecation process have never been initiated, so this
+        # compatibility must stay, probably forever.
+        if is_pointer_tag(node):
+            # new style:
+            # read the actual definition file--named using url_name.replace(':','/')
+            definition_xml, filepath = load_definition_xml(node, runtime, keys.def_id)
+            aside_children = runtime.parse_asides(definition_xml, keys.def_id, keys.usage_id, runtime.id_generator)
+        else:
+            filepath = None
+            definition_xml = node
+
+        # Removes metadata
+        definition, children = cls.load_definition(definition_xml)
+
+        # VS[compat]
+        # Make Ike's github preview links work in both old and new file layouts.
+        if is_pointer_tag(node):
+            # new style -- contents actually at filepath
+            definition["filename"] = [filepath, filepath]
+
+        metadata = cls.load_metadata(definition_xml)
+
+        # move definition metadata into dict
+        dmdata = definition.get("definition_metadata", "")
+        if dmdata:
+            metadata["definition_metadata_raw"] = dmdata
+            try:
+                metadata.update(json.loads(dmdata))
+            except Exception as err:  # lint-amnesty, pylint: disable=broad-except
+                log.debug("Error in loading metadata %r", dmdata, exc_info=True)
+                metadata["definition_metadata_err"] = str(err)
+
+        definition_aside_children = definition.pop("aside_children", None)
+        if definition_aside_children:
+            aside_children.extend(definition_aside_children)
+
+        # Set/override any metadata specified by policy
+        cls.apply_policy(metadata, runtime.get_policy(keys.usage_id))
+
+        field_data = {**metadata, **definition}
+
+        for field_name, value in field_data.items():
+            # The 'xml_attributes' field has a special setter logic in its Field class,
+            # so we must handle it carefully to avoid duplicating data.
+            if field_name == "xml_attributes":
+                # The 'filename' attribute is specially handled for git links.
+                value["filename"] = definition.get("filename", ["", None])
+                block.xml_attributes.update(value)
+            elif field_name in block.fields:
+                setattr(block, field_name, value)
+
+        block.children = children
+
+        if aside_children:
+            cls.add_applicable_asides_to_block(block, runtime, aside_children)
+
+        return block
+
+    @classmethod
+    def add_applicable_asides_to_block(cls, block, runtime, aside_children):
+        """
+        Add asides to the block. Moved this out of the parse_xml method to use it in the VideoBlock.parse_xml
+        """
+        asides_tags = [aside_child.tag for aside_child in aside_children]
+        asides = runtime.get_asides(block)
+        for aside in asides:
+            if aside.scope_ids.block_type in asides_tags:
+                block.add_aside(aside)
+
+    def export_to_file(self):
+        """If this returns True, write the definition of this block to a separate
+        file.
+        """
+        return True
+
+    def add_xml_to_node(self, node):
+        """For exporting, set data on `node` from ourselves."""
+        xml_object = etree.Element(self.category)
+
+        if xml_object is None:
+            return
+
+        for aside in self.runtime.get_asides(self):
+            if aside.needs_serialization():
+                aside_node = etree.Element("unknown_root", nsmap=XML_NAMESPACES)
+                aside.add_xml_to_node(aside_node)
+                xml_object.append(aside_node)
+
+        self.clean_metadata_from_xml(xml_object)
+        xml_object.tag = self.category
+        node.tag = self.category
+
+        for attr in sorted(own_metadata(self)):
+            if attr not in self.metadata_to_strip:
+                # pylint: disable=unsubscriptable-object
+                val = serialize_field(self.fields[attr].to_json(getattr(self, attr)))
+                try:
+                    xml_object.set(attr, val)
+                except Exception:  # pylint: disable=broad-exception-caught
+                    logging.exception("Failed to serialize metadata attribute %s in module %s.", attr, self.url_name)
+
+        for key, value in self.xml_attributes.items():
+            if key not in self.metadata_to_strip:
+                xml_object.set(key, serialize_field(value))
+
+        if self.export_to_file():
+            url_path = name_to_pathname(self.url_name)
+            filepath = format_filepath(self.category, url_path)
+            self.runtime.export_fs.makedirs(os.path.dirname(filepath), recreate=True)
+            with self.runtime.export_fs.open(filepath, "wb") as fileobj:
+                ElementTree(xml_object).write(fileobj, pretty_print=True, encoding="utf-8")
+        else:
+            node.clear()
+            node.tag = xml_object.tag
+            node.text = xml_object.text
+            node.tail = xml_object.tail
+            node.attrib.update(xml_object.attrib)
+            node.extend(xml_object)
+
+        apply_pointer_attributes(node, self)
+
+
 @XBlock.needs("i18n")
-class WordCloudBlock(StudioEditableXBlockMixin, XBlock):
+class WordCloudBlock(StudioEditableXBlockMixin, WordCloudXmlMixin, XBlock):
     """
     Word Cloud XBlock.
     """