add wordpress migration

dnscms/dnscms/blocks.py

@@ -234,6 +234,7 @@ BASE_BLOCKS = [
     ("accordion", AccordionBlock()),
     ("fact_box", FactBoxBlock()),
     ("embed", EmbedBlock()),
+    ("raw_html", blocks.RawHTMLBlock()),
 ]
 
 

dnscms/dnscms/settings/base.py

@@ -12,6 +12,7 @@ https://docs.djangoproject.com/en/4.2/ref/settings/
 
 # Build paths inside the project like this: os.path.join(BASE_DIR, ...)
 import os
+import re
 
 PROJECT_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
 BASE_DIR = os.path.dirname(PROJECT_DIR)
@@ -36,6 +37,7 @@ INSTALLED_APPS = [
     "news",
     "openinghours",
     # end cms apps
+    "wagtail_wordpress_import",
     "grapple",
     "graphene_django",
     "wagtail.contrib.forms",
@@ -201,3 +203,74 @@ GRAPPLE = {
     ],
     "EXPOSE_GRAPHIQL": True,
 }
+
+# Wgtail WordPress import
+WAGTAIL_WORDPRESS_IMPORTER_SOURCE_DOMAIN = "https://studentersamfundet.no/"
+WAGTAIL_WORDPRESS_IMPORTER_CONVERT_HTML_TAGS_TO_BLOCKS = {
+    # "h1": "wagtail_wordpress_import.block_builder_defaults.build_heading_block",
+    "table": "wagtail_wordpress_import.block_builder_defaults.build_table_block",
+    # "iframe": "wagtail_wordpress_import.block_builder_defaults.build_iframe_block",
+    # "form": "wagtail_wordpress_import.block_builder_defaults.build_form_block",
+    # "img": "wagtail_wordpress_import.block_builder_defaults.build_image_block",
+    # "blockquote": "wagtail_wordpress_import.block_builder_defaults.build_block_quote_block",
+}
+WAGTAIL_WORDPRESS_IMPORTER_FALLBACK_BLOCK = (
+    "dnscms.wordpress.block_builder.build_richtext_block_content"
+)
+WORDPRESS_IMPORT_HOOKS_ITEMS_TO_CACHE = {
+    "attachment": {
+        "DATA_TAG": "thumbnail_id",
+        "FUNCTION": "dnscms.wordpress.import_hooks.header_image_processor",
+    }
+}
+# WORDPRESS_IMPORT_HOOKS_TAGS_TO_CACHE = {
+#     "wp:term": {
+#         "DATA_TAG": "category",
+#         "FUNCTION": "dnscms.wordpress.import_hooks.categories_processor",
+#     }
+# }
+WAGTAIL_WORDPRESS_IMPORTER_PROMOTE_CHILD_TAGS = {
+    "TAGS_TO_PROMOTE": [],
+    "PARENTS_TO_REMOVE": ["p", "div", "span"],
+}
+WAGTAIL_WORDPRESS_IMPORT_PREFILTERS = [
+    {
+        "FUNCTION": "wagtail_wordpress_import.prefilters.linebreaks_wp",
+    },
+    {
+        "FUNCTION": "wagtail_wordpress_import.prefilters.transform_shortcodes",
+    },
+    {
+        "FUNCTION": "wagtail_wordpress_import.prefilters.transform_inline_styles",
+        "OPTIONS": {
+            "TRANSFORM_STYLES_MAPPING": [
+                (
+                    re.compile(r"font-style:italic;font-weight:bold;", re.IGNORECASE),
+                    "wagtail_wordpress_import.prefilters.transform_styles_defaults.transform_style_bold_italic",
+                ),
+                (
+                    re.compile(r"font-weight:bold;", re.IGNORECASE),
+                    "wagtail_wordpress_import.prefilters.transform_styles_defaults.transform_style_bold",
+                ),
+                (
+                    re.compile(r"font-style:italic;", re.IGNORECASE),
+                    "wagtail_wordpress_import.prefilters.transform_styles_defaults.transform_style_italic",
+                ),
+                # (
+                #     re.compile(
+                #         r"text-align:center;",
+                #         re.IGNORECASE,
+                #     ),
+                #     transform_style_center,
+                # ),
+                # (re.compile(r"text-align:left;", re.IGNORECASE), transform_style_left),
+                # (re.compile(r"text-align:right;", re.IGNORECASE), transform_style_right),
+                # (re.compile(r"float:left;", re.IGNORECASE), transform_float_left),
+                # (re.compile(r"float:right;", re.IGNORECASE), transform_float_right),
+            ],
+        },
+    },
+    {
+        "FUNCTION": "wagtail_wordpress_import.prefilters.bleach_clean",
+    },
+]

dnscms/dnscms/wordpress/block_builder.py

@@ -0,0 +1,24 @@
+from django.conf import settings
+from wagtail_wordpress_import.block_builder_defaults import (
+    document_linker,
+    image_linker,
+    import_string,
+)
+
+
+def build_richtext_block_content(html, blocks):
+    """
+    image_linker is called to link up and retrive the remote image
+    document_linker is called to link up and retrive the remote documents
+    filters are called to replace inline shortcodes
+    """
+    html = image_linker(html)
+    html = document_linker(html)
+    for inline_shortcode_handler in getattr(
+        settings, "WAGTAIL_WORDPRESS_IMPORTER_INLINE_SHORTCODE_HANDLERS", []
+    ):
+        function = import_string(inline_shortcode_handler).construct_html_tag
+        html = function(html)
+    blocks.append({"type": "paragraph", "value": html})
+    html = ""
+    return html

dnscms/dnscms/wordpress/import_hooks.py

@@ -0,0 +1,51 @@
+from wagtail_wordpress_import.block_builder_defaults import get_or_save_image
+
+
+def header_image_processor(imported_pages, data_tag, items_cache):
+    """
+    imported_pages:
+        Is a specific() page model queryset of all imported pages.
+    data_tag:
+        Is the value of the `DATA_TAG` key from the configuration above.
+    items_cache:
+        Is a list of dictionaries, one for each item in the XML file.
+    """
+
+    # See note above about leading _ and : characters in the XML value
+    lookup = f"wp_post_meta__{data_tag}"
+
+    for attachment in items_cache:
+        # The id of the cached item used in the filter
+        thumbnail_id = attachment.get("wp:post_id")
+
+        # Filter the imported_pages for only pages that include the
+        # matching thumbnail_id in the wp_post_meta field
+        pages = imported_pages.filter(**{lookup: thumbnail_id})
+
+        if pages.exists():
+            # guid is the url of the image to fetch, the get_or_save_image()
+            # function will fetch the image if it doesn't exist
+            image_url = attachment.get("guid")
+            # fix cases where the /wp prefix is missing from the image url
+            if image_url.startswith("https://studentersamfundet.no/wp-content/uploads/"):
+                image_url = image_url.replace(
+                    "https://studentersamfundet.no/wp-content/uploads/",
+                    "https://studentersamfundet.no/wp/wp-content/uploads/",
+                )
+            try:
+                image = get_or_save_image(image_url)
+            except Exception as e:
+                print("Error with image", image_url, "associated with pages:", pages)
+                print(e)
+                continue
+
+            print("Attaching header images to pages:", pages)
+            try:
+                pages.update(featured_image=image)
+            except Exception:
+                pass
+
+            try:
+                pages.update(logo=image)
+            except Exception:
+                pass