Refine keep_unmapped_properties + fix audit test URL drift

README.md

@@ -53,7 +53,7 @@ The export configuration items may contain the following keys:
 | notion_filter | A [notion filter object](https://developers.notion.com/reference/post-database-query-filter) to be applied to the database. |
 | notion_sorts | A [notion sorts object](https://developers.notion.com/reference/post-database-query-sort) to be applied to the database. |
 | property_map | A mapping between the name of properties in Notion, and the name of the properties in the exported files. Set the new value to `null` to discard the property.|
-| keep_unmapped_properties | Defaults to `true`. When set to `false`, the `property_map` acts as an allowlist: any property not in the map is dropped from the export. This prevents properties added or renamed in Notion from leaking into the exported files. The `id_property` and `url_property` are always kept. |
+| keep_unmapped_properties | Defaults to `true`. When set to `false`, the `property_map` acts as an allowlist: any Notion property not in the map is dropped from the export. This prevents properties added or renamed in Notion from leaking into the exported files. The `id_property`, the `url_property`, and properties injected by plugins are always kept. |
 
 ## Example Configuration Files
 

n2y/export.py

@@ -54,13 +54,16 @@ def _page_properties(
             page.client.logger.warning(msg, original, page.notion_url, original, new)
     if not keep_unmapped_properties:
         # Treat the property map as an allowlist so that properties added or
-        # renamed in Notion don't leak into the exported files.
+        # renamed in Notion don't leak into the exported files. Only properties
+        # that exist in Notion are dropped; properties injected by plugins
+        # (which won't appear in the page's notion_data) are always kept.
+        notion_property_names = set(page.notion_data["properties"].keys())
         kept_properties = {new for new in property_map.values() if new}
         kept_properties.update(p for p in (id_property, url_property) if p)
         properties = {
             name: value
             for name, value in properties.items()
-            if name in kept_properties
+            if name in kept_properties or name not in notion_property_names
         }
     return properties
 

tests/test_audit_end_to_end.py

@@ -25,19 +25,22 @@ def test_audit():
     status, stdoutput = run_n2yaudit([object_id])
     assert status == 3
 
-    external_mention_in_top_page = "https://www.notion.so/Audited-cfa8ff07bba244c8b967c9b6a7a954c1#aa4fa886f8244c818de8018bb3491806"  # noqa: E501
-    external_mention_in_child_page = "https://www.notion.so/Child-f3e3628fc80c470ea68994fa7ec0ff17#d1d32ff6f0cb4c71a2f1c4ec55e00086"  # noqa: E501
-    internal_mention_in_child_page = "https://www.notion.so/Child-f3e3628fc80c470ea68994fa7ec0ff17#eab91ccc32924221ac3f0a74225a33dd"  # noqa: E501
-    external_mention_in_child_database = "https://www.notion.so/B-4412005dcec24ff2827abbc367c90b29#6373a0b5c2804fbe9dfac167ce6948a0"  # noqa: E501
-    internal_mention_in_database_in_column = "https://www.notion.so/Audited-cfa8ff07bba244c8b967c9b6a7a954c1#21a13c06ef86462e882a181c6cb52a64"  # noqa: E501
+    # Only the page-slug#block-id fragment is asserted (not the full URL)
+    # because Notion has changed the URL host/path over time (e.g. from
+    # www.notion.so/<slug> to app.notion.com/p/<slug>).
+    external_mention_in_top_page = "Audited-cfa8ff07bba244c8b967c9b6a7a954c1#aa4fa886f8244c818de8018bb3491806"  # noqa: E501
+    external_mention_in_child_page = "Child-f3e3628fc80c470ea68994fa7ec0ff17#d1d32ff6f0cb4c71a2f1c4ec55e00086"  # noqa: E501
+    internal_mention_in_child_page = "Child-f3e3628fc80c470ea68994fa7ec0ff17#eab91ccc32924221ac3f0a74225a33dd"  # noqa: E501
+    external_mention_in_child_database = "B-4412005dcec24ff2827abbc367c90b29#6373a0b5c2804fbe9dfac167ce6948a0"  # noqa: E501
+    internal_mention_in_database_in_column = "Audited-cfa8ff07bba244c8b967c9b6a7a954c1#21a13c06ef86462e882a181c6cb52a64"  # noqa: E501
 
     # NOTE: When you try to get a link for a "LinkToPageBlock" in the Notion UI,
     # it appears to give you the URL for the linked page or database, and not to
     # the block itself. Thus, these two links were extracted using a debugger
     # when running this test.
 
-    external_link_in_top_page = "https://www.notion.so/Audited-cfa8ff07bba244c8b967c9b6a7a954c1#c22d76d50c704761b0e729531e6cc24b"  # noqa: E501
-    internal_link_in_top_page = "https://www.notion.so/Audited-cfa8ff07bba244c8b967c9b6a7a954c1#3e2e2fb2e9bf4a1f8b00bbb18a9d97e9"  # noqa: E501
+    external_link_in_top_page = "Audited-cfa8ff07bba244c8b967c9b6a7a954c1#c22d76d50c704761b0e729531e6cc24b"  # noqa: E501
+    internal_link_in_top_page = "Audited-cfa8ff07bba244c8b967c9b6a7a954c1#3e2e2fb2e9bf4a1f8b00bbb18a9d97e9"  # noqa: E501
 
     assert external_mention_in_top_page in stdoutput
     assert external_mention_in_child_page in stdoutput

tests/test_export.py

@@ -56,6 +56,16 @@ def test_page_properties_drop_unmapped_keeps_id_and_url(page):
     assert set(properties) == {"id", "url"}
 
 
+def test_page_properties_drop_unmapped_keeps_plugin_injected(page):
+    # Plugins may inject extra properties (e.g. embedded mention data) that
+    # don't exist in the page's notion_data; the allowlist must keep them.
+    page.properties["injected"] = page.properties["property"]
+    properties = _page_properties(
+        page, property_map={"property": "p"}, keep_unmapped_properties=False
+    )
+    assert properties == {"p": "P\n", "injected": "P\n"}
+
+
 def test_page_properties_drop_unmapped_keeps_mapped_title(page):
     properties = _page_properties(
         page, property_map={"title": "title"}, keep_unmapped_properties=False