feat(property-converter): unique_id

notion_exporter/exporter.py

@@ -1,16 +1,20 @@
-from typing import Optional
 import asyncio
 import logging
+from typing import Optional
 
-from notion_client import AsyncClient as NotionClient, APIResponseError
+from notion_client import APIResponseError
+from notion_client import AsyncClient as NotionClient
 from notion_client import Client
 from notion_client.helpers import async_collect_paginated_api
 from tenacity import retry, retry_if_exception, stop_after_attempt, wait_exponential
 
 from notion_exporter.block_converter import BlockConverter
 from notion_exporter.property_converter import PropertyConverter
-from notion_exporter.retry_utils import is_rate_limit_exception, wait_for_retry_after_header, is_unavailable_exception
-
+from notion_exporter.retry_utils import (
+    is_rate_limit_exception,
+    is_unavailable_exception,
+    wait_for_retry_after_header,
+)
 
 logging.basicConfig(level=logging.INFO, format="%(asctime)s %(message)s", datefmt="%m/%d/%Y %I:%M:%S %p")
 logger = logging.getLogger(__name__)
@@ -26,19 +30,22 @@ def __init__(
         notion_token: str,
         export_child_pages: bool = False,
         extract_page_metadata: bool = False,
+        extract_page_properties: bool = False,
         exclude_title_containing: Optional[str] = None,
     ):
         """
         :param notion_token: Notion API token.
         :param export_child_pages: Whether to export child pages. Default: True.
         :param extract_page_metadata: Whether to extract page metadata. Default: False.
         :param exclude_title_containing: If specified, pages with titles containing this string will be excluded.
+        :param extract_page_properties: Whether to extract page properties. Default: False.
 
         """
         self.notion = NotionClient(auth=notion_token)
         self.sync_notion = Client(auth=notion_token)
         self.export_child_pages = export_child_pages
         self.extract_page_metadata = extract_page_metadata
+        self.extract_page_properties = extract_page_properties
         self.exclude_title_containing = exclude_title_containing
         self.block_converter = BlockConverter()
         self.property_converter = PropertyConverter(self)
@@ -293,7 +300,10 @@ async def _get_child_blocks(self, page_id: str) -> list[dict]:
         stop=stop_after_attempt(3),
     )
     async def _get_page_meta(self, page_id: str) -> dict:
-        page_object = await self.notion.pages.retrieve(page_id)
+        try:
+            page_object = await self.notion.pages.retrieve(page_id)
+        except APIResponseError:
+            return {"page_id": page_id, "title": "Untitled", "url": "", "created_by": "", "last_edited_by": ""}
         created_by, last_edited_by = await asyncio.gather(
             self._get_user(page_object["created_by"]["id"]), self._get_user(page_object["last_edited_by"]["id"])
         )
@@ -401,7 +411,7 @@ def _get_page_front_matter(self, page_meta: dict, page_paths: dict, parent_page_
         front_matter = ""
         if self.extract_page_metadata:
             front_matter += "---\n"
-            front_matter += f"title: {page_meta['title']}\n"
+            front_matter += f'title: "{page_meta['title']}"\n'
             # Add quotation marks to avoid issues with colons in page titles
             front_matter += f'path: "{current_page_path}"\n'
             front_matter += f"url: {page_meta['url']}\n"
@@ -411,7 +421,7 @@ def _get_page_front_matter(self, page_meta: dict, page_paths: dict, parent_page_
             front_matter += "---\n\n"
 
         # Add properties of database entries as key-value pairs
-        if "properties" in page_meta:
+        if self.extract_page_properties and "properties" in page_meta:
             for prop_name, prop in page_meta["properties"].items():
                 front_matter += f"{prop_name}: {prop}\n"
             front_matter += "\n"

notion_exporter/property_converter.py

@@ -34,6 +34,7 @@ def __init__(self, notion_exporter: "NotionExporter"):
             "status": self.status,
             "title": self.title,
             "url": self.url,
+            "unique_id": self.unique_id,
         }
         self.notion_exporter = notion_exporter
 
@@ -206,3 +207,13 @@ def url(property_item: dict) -> str:
         """
         url = property_item["url"] if property_item["url"] else ""
         return url
+
+    @staticmethod
+    def unique_id(property_item: dict) -> str:
+        """
+        Converts a unique_id property to a Markdown string.
+        """
+        uid = property_item["unique_id"]
+        prefix = uid.get('prefix', '')
+        value = uid.get('number', '')
+        return f"{prefix}-{value}"