Implement get_document() API

Author: krassowski

docs/source/developer/python_api.rst

@@ -4,6 +4,25 @@
 Python API
 ==========
 
+``jupyter_collaboration`` instantiates :any:`YDocExtension` and stores it under ``serverapp.settings`` dictionary, under the ``"jupyter_collaboration"`` key.
+This instance can be used in other extensions to access the public API methods.
+
+For example, to access a read-only view of the shared notebook model in your jupyter-server extension, you can use the :any:`get_document` method:
+
+.. code-block::
+
+   collaboration = serverapp.settings["jupyter_collaboration"]
+   document = collaboration.get_document(
+       path='Untitled.ipynb',
+       content_type="notebook",
+       file_format="json"
+   )
+   content = document.get()
+
+
+API Reference
+-------------
+
 .. automodule:: jupyter_collaboration.app
   :members:
   :inherited-members:

jupyter_collaboration/app.py

@@ -3,16 +3,26 @@
 from __future__ import annotations
 
 import asyncio
+from typing import Literal
 
 from jupyter_server.extension.application import ExtensionApp
+from jupyter_ydoc import ydocs as YDOCS
+from jupyter_ydoc.ybasedoc import YBaseDoc
+from pycrdt import Doc
 from pycrdt_websocket.ystore import BaseYStore
 from traitlets import Bool, Float, Type
 
 from .handlers import DocSessionHandler, YDocWebSocketHandler
 from .loaders import FileLoaderMapping
+from .rooms import DocumentRoom
 from .stores import SQLiteYStore
-from .utils import AWARENESS_EVENTS_SCHEMA_PATH, EVENTS_SCHEMA_PATH
-from .websocketserver import JupyterWebsocketServer
+from .utils import (
+    AWARENESS_EVENTS_SCHEMA_PATH,
+    EVENTS_SCHEMA_PATH,
+    encode_file_path,
+    room_id_from_encoded_path,
+)
+from .websocketserver import JupyterWebsocketServer, RoomNotFound
 
 
 class YDocExtension(ExtensionApp):
@@ -124,6 +134,39 @@ def initialize_handlers(self):
             ]
         )
 
+    async def get_document(
+        self: YDocExtension,
+        *,
+        path: str,
+        content_type: Literal["notebook", "file"],
+        file_format: Literal["json", "text"],
+    ) -> YBaseDoc | None:
+        """Get a read-only view of the shared model for the matching document.
+
+        The returned shared model is a fork, meaning that any changes made to it
+        will not be propagated to the shared model used by the application.
+        """
+        file_id_manager = self.serverapp.web_app.settings["file_id_manager"]
+        file_id = file_id_manager.index(path)
+
+        encoded_path = encode_file_path(file_format, content_type, file_id)
+        room_id = room_id_from_encoded_path(encoded_path)
+
+        try:
+            room = await self.ywebsocket_server.get_room(room_id)
+        except RoomNotFound:
+            return None
+
+        if isinstance(room, DocumentRoom):
+            update = room.ydoc.get_update()
+
+            fork_ydoc = Doc()
+            fork_ydoc.apply_update(update)
+
+            return YDOCS.get(content_type, YDOCS["file"])(fork_ydoc)
+
+        return None
+
     async def stop_extension(self):
         # Cancel tasks and clean up
         await asyncio.wait(

jupyter_collaboration/handlers.py

@@ -26,6 +26,7 @@
     LogLevel,
     MessageType,
     decode_file_path,
+    room_id_from_encoded_path,
 )
 from .websocketserver import JupyterWebsocketServer
 
@@ -74,7 +75,7 @@ async def prepare(self):
             await self._websocket_server.started.wait()
 
         # Get room
-        self._room_id: str = self.request.path.split("/")[-1]
+        self._room_id: str = room_id_from_encoded_path(self.request.path)
 
         async with self._room_lock(self._room_id):
             if self._websocket_server.room_exists(self._room_id):

jupyter_collaboration/utils.py

@@ -70,3 +70,8 @@ def encode_file_path(format: str, file_type: str, file_id: str) -> str:
             path (str): File path.
     """
     return f"{format}:{file_type}:{file_id}"
+
+
+def room_id_from_encoded_path(encoded_path: str) -> str:
+    """Transforms the encoded path into a stable room identifier."""
+    return encoded_path.split("/")[-1]

tests/test_app.py

@@ -59,3 +59,23 @@ def test_settings_should_change_ystore_class(jp_configurable_serverapp):
     settings = app.web_app.settings["jupyter_collaboration_config"]
 
     assert settings["ystore_class"] == TempFileYStore
+
+
+async def test_get_document_file(rtc_create_file, jp_serverapp):
+    path, content = await rtc_create_file("test.txt", "test", store=True)
+    collaboration = jp_serverapp.web_app.settings["jupyter_collaboration"]
+    document = await collaboration.get_document(path=path, content_type="file", file_format="text")
+    assert document.get() == content == "test"
+    await collaboration.stop_extension()
+
+
+async def test_get_document_file_is_a_fork(rtc_create_file, jp_serverapp, rtc_fetch_session):
+    path, content = await rtc_create_file("test.txt", "test", store=True)
+    collaboration = jp_serverapp.web_app.settings["jupyter_collaboration"]
+    document = await collaboration.get_document(path=path, content_type="file", file_format="text")
+    document.set("other")
+    fresh_copy = await collaboration.get_document(
+        path=path, content_type="file", file_format="text"
+    )
+    assert fresh_copy.get() == "test"
+    await collaboration.stop_extension()