Support get file(notebook) md5

docs/source/developers/contents.rst

@@ -63,6 +63,10 @@ Models may contain the following entries:
 |                    |``None``   |if any. (:ref:`See            |
 |                    |           |Below<modelcontent>`)         |
 +--------------------+-----------+------------------------------+
+|**md5**             |unicode or |The md5 of the contents.      |
+|                    |``None``   |                              |
+|                    |           |                              |
++--------------------+-----------+------------------------------+
 
 .. _modelcontent:
 
@@ -76,6 +80,8 @@ model. There are three model types: **notebook**, **file**, and **directory**.
       :class:`nbformat.notebooknode.NotebookNode` representing the .ipynb file
       represented by the model.  See the `NBFormat`_ documentation for a full
       description.
+    - The ``md5`` field a hexdigest string of the md5 value of the notebook
+      file.
 
 - ``file`` models
     - The ``format`` field is either ``"text"`` or ``"base64"``.
@@ -85,12 +91,14 @@ model. There are three model types: **notebook**, **file**, and **directory**.
       file models, ``content`` simply contains the file's bytes after decoding
       as UTF-8.  Non-text (``base64``) files are read as bytes, base64 encoded,
       and then decoded as UTF-8.
+    - The ``md5`` field a hexdigest string of the md5 value of the file.
 
 - ``directory`` models
     - The ``format`` field is always ``"json"``.
     - The ``mimetype`` field is always ``None``.
     - The ``content`` field contains a list of :ref:`content-free<contentfree>`
       models representing the entities in the directory.
+    - The ``md5`` field is always ``None``.
 
 .. note::
 
@@ -129,6 +137,7 @@ model. There are three model types: **notebook**, **file**, and **directory**.
         "path": "foo/a.ipynb",
         "type": "notebook",
         "writable": True,
+        "md5": "7e47382b370c05a1b14706a2a8aff91a",
     }
 
     # Notebook Model without Content

jupyter_server/services/contents/fileio.py

@@ -4,6 +4,7 @@
 # Copyright (c) Jupyter Development Team.
 # Distributed under the terms of the Modified BSD License.
 import errno
+import hashlib
 import os
 import shutil
 from base64 import decodebytes, encodebytes
@@ -268,7 +269,9 @@ def _read_notebook(self, os_path, as_version=4, capture_validation_error=None):
         with self.open(os_path, "r", encoding="utf-8") as f:
             try:
                 return nbformat.read(
-                    f, as_version=as_version, capture_validation_error=capture_validation_error
+                    f,
+                    as_version=as_version,
+                    capture_validation_error=capture_validation_error,
                 )
             except Exception as e:
                 e_orig = e
@@ -309,13 +312,17 @@ def _read_file(self, os_path, format):
         format:
           If 'text', the contents will be decoded as UTF-8.
           If 'base64', the raw bytes contents will be encoded as base64.
+          If 'byte', the raw bytes contents will be returned.
           If not specified, try to decode as UTF-8, and fall back to base64
         """
         if not os.path.isfile(os_path):
             raise HTTPError(400, "Cannot read non-file %s" % os_path)
 
         with self.open(os_path, "rb") as f:
             bcontent = f.read()
+        if format == "byte":
+            # Not for http response but internal use
+            return bcontent, "byte"
 
         if format is None or format == "text":
             # Try to interpret as unicode if format is unknown or if unicode
@@ -350,6 +357,12 @@ def _save_file(self, os_path, content, format):
         with self.atomic_writing(os_path, text=False) as f:
             f.write(bcontent)
 
+    def _get_md5(self, os_path):
+        c, _ = self._read_file(os_path, "byte")
+        md5 = hashlib.md5()  # noqa: S324
+        md5.update(c)
+        return md5.hexdigest()
+
 
 class AsyncFileManagerMixin(FileManagerMixin):
     """
@@ -417,13 +430,17 @@ async def _read_file(self, os_path, format):
         format:
           If 'text', the contents will be decoded as UTF-8.
           If 'base64', the raw bytes contents will be encoded as base64.
+          If 'byte', the raw bytes contents will be returned.
           If not specified, try to decode as UTF-8, and fall back to base64
         """
         if not os.path.isfile(os_path):
             raise HTTPError(400, "Cannot read non-file %s" % os_path)
 
         with self.open(os_path, "rb") as f:
             bcontent = await run_sync(f.read)
+        if format == "byte":
+            # Not for http response but internal use
+            return bcontent, "byte"
 
         if format is None or format == "text":
             # Try to interpret as unicode if format is unknown or if unicode
@@ -457,3 +474,9 @@ async def _save_file(self, os_path, content, format):
 
         with self.atomic_writing(os_path, text=False) as f:
             await run_sync(f.write, bcontent)
+
+    async def _get_md5(self, os_path):
+        c, _ = await self._read_file(os_path, "byte")
+        md5 = hashlib.md5()  # noqa: S324
+        await run_sync(md5.update, c)
+        return md5.hexdigest()

jupyter_server/services/contents/filemanager.py

@@ -268,6 +268,7 @@ def _base_model(self, path):
         model["mimetype"] = None
         model["size"] = size
         model["writable"] = self.is_writable(path)
+        model["md5"] = None
 
         return model
 
@@ -335,7 +336,7 @@ def _dir_model(self, path, content=True):
 
         return model
 
-    def _file_model(self, path, content=True, format=None):
+    def _file_model(self, path, content=True, format=None, md5=False):
         """Build a model for a file
 
         if content is requested, include the file contents.
@@ -364,10 +365,13 @@ def _file_model(self, path, content=True, format=None):
                 content=content,
                 format=format,
             )
+        if md5:
+            md5 = self._get_md5(os_path)
+            model.update(md5=md5)
 
         return model
 
-    def _notebook_model(self, path, content=True):
+    def _notebook_model(self, path, content=True, md5=False):
         """Build a notebook model
 
         if content is requested, the notebook content will be populated
@@ -386,10 +390,12 @@ def _notebook_model(self, path, content=True):
             model["content"] = nb
             model["format"] = "json"
             self.validate_notebook_model(model, validation_error)
+        if md5:
+            model["md5"] = self._get_md5(os_path)
 
         return model
 
-    def get(self, path, content=True, type=None, format=None):
+    def get(self, path, content=True, type=None, format=None, md5=None):
         """Takes a path for an entity and returns its model
 
         Parameters
@@ -404,6 +410,8 @@ def get(self, path, content=True, type=None, format=None):
         format : str, optional
             The requested format for file contents. 'text' or 'base64'.
             Ignored if this returns a notebook or directory model.
+        md5: bool, optional
+            Whether to include the md5 of the file contents.
 
         Returns
         -------
@@ -431,11 +439,11 @@ def get(self, path, content=True, type=None, format=None):
                 )
             model = self._dir_model(path, content=content)
         elif type == "notebook" or (type is None and path.endswith(".ipynb")):
-            model = self._notebook_model(path, content=content)
+            model = self._notebook_model(path, content=content, md5=md5)
         else:
             if type == "directory":
                 raise web.HTTPError(400, "%s is not a directory" % path, reason="bad type")
-            model = self._file_model(path, content=content, format=format)
+            model = self._file_model(path, content=content, format=format, md5=md5)
         self.emit(data={"action": "get", "path": path})
         return model
 
@@ -686,7 +694,9 @@ def _get_dir_size(self, path="."):
                 ).stdout.split()
             else:
                 result = subprocess.run(
-                    ["du", "-s", "--block-size=1", path], capture_output=True, check=True
+                    ["du", "-s", "--block-size=1", path],
+                    capture_output=True,
+                    check=True,
                 ).stdout.split()
 
             self.log.info(f"current status of du command {result}")
@@ -784,7 +794,7 @@ async def _dir_model(self, path, content=True):
 
         return model
 
-    async def _file_model(self, path, content=True, format=None):
+    async def _file_model(self, path, content=True, format=None, md5=False):
         """Build a model for a file
 
         if content is requested, include the file contents.
@@ -813,10 +823,13 @@ async def _file_model(self, path, content=True, format=None):
                 content=content,
                 format=format,
             )
+        if md5:
+            md5 = await self._get_md5(os_path)
+            model.update(md5=md5)
 
         return model
 
-    async def _notebook_model(self, path, content=True):
+    async def _notebook_model(self, path, content=True, md5=False):
         """Build a notebook model
 
         if content is requested, the notebook content will be populated
@@ -835,10 +848,12 @@ async def _notebook_model(self, path, content=True):
             model["content"] = nb
             model["format"] = "json"
             self.validate_notebook_model(model, validation_error)
+        if md5:
+            model["md5"] = await self._get_md5(os_path)
 
         return model
 
-    async def get(self, path, content=True, type=None, format=None):
+    async def get(self, path, content=True, type=None, format=None, md5=False):
         """Takes a path for an entity and returns its model
 
         Parameters
@@ -853,6 +868,8 @@ async def get(self, path, content=True, type=None, format=None):
         format : str, optional
             The requested format for file contents. 'text' or 'base64'.
             Ignored if this returns a notebook or directory model.
+        md5: bool, optional
+            Whether to include the md5 of the file contents.
 
         Returns
         -------
@@ -875,11 +892,11 @@ async def get(self, path, content=True, type=None, format=None):
                 )
             model = await self._dir_model(path, content=content)
         elif type == "notebook" or (type is None and path.endswith(".ipynb")):
-            model = await self._notebook_model(path, content=content)
+            model = await self._notebook_model(path, content=content, md5=md5)
         else:
             if type == "directory":
                 raise web.HTTPError(400, "%s is not a directory" % path, reason="bad type")
-            model = await self._file_model(path, content=content, format=format)
+            model = await self._file_model(path, content=content, format=format, md5=md5)
         self.emit(data={"action": "get", "path": path})
         return model
 
@@ -1147,7 +1164,9 @@ async def _get_dir_size(self, path: str = ".") -> str:
                 ).stdout.split()
             else:
                 result = subprocess.run(
-                    ["du", "-s", "--block-size=1", path], capture_output=True, check=True
+                    ["du", "-s", "--block-size=1", path],
+                    capture_output=True,
+                    check=True,
                 ).stdout.split()
 
             self.log.info(f"current status of du command {result}")