Temporarily disable proto compilation for orbax/experimental/model in doc serving script.

PiperOrigin-RevId: 838982402

Author: lyglst

docs/conf.py

@@ -24,6 +24,7 @@
 add these directories to sys.path here. If the directory is relative to the
 documentation root, use os.path.abspath to make it absolute, like shown here.
 """
+import datetime
 import inspect
 import os
 import sys
@@ -37,7 +38,7 @@
 # -- Project information -----------------------------------------------------
 
 project = 'Orbax'
-copyright = '2024, Google'  # pylint: disable=redefined-builtin
+copyright = f'{datetime.datetime.now().year}, Google'  # pylint: disable=redefined-builtin
 author = 'Orbax Contributors'
 
 # -- General configuration ---------------------------------------------------

docs/guides/checkpoint/orbax_checkpoint_101.ipynb

@@ -15,7 +15,7 @@
         "id": "1Ik2ARq4JaL3"
       },
       "source": [
-        "This page serves as a simple overview of common tasks that you may wish to accomplish with Orbax. For more in-depth documentation of the API's, see [API Overview](https://orbax.readthedocs.io/en/latest/guides/checkpoint/orbax_checkpoint_api_overview.html)."
+        "This page serves as a simple overview of common tasks that you may wish to accomplish with Orbax. For more in-depth documentation of the API's, see {doc}`API Overview</guides/checkpoint/orbax_checkpoint_api_overview>`."
       ]
     },
     {
@@ -33,7 +33,7 @@
         "id": "VyfEdKvwswys"
       },
       "source": [
-        "The following example shows how you can synchronously save and restore a PyTree. See [Checkpointing PyTrees](https://orbax.readthedocs.io/en/latest/guides/checkpoint/checkpointing_pytrees.html) for more detail.\n",
+        "The following example shows how you can synchronously save and restore a PyTree. See {doc}`Checkpointing PyTrees</guides/checkpoint/checkpointing_pytrees>` for more detail.\n",
         "\n"
       ]
     },
@@ -106,7 +106,7 @@
       "source": [
         "To save and restore, we create a `Checkpointer` object. The `Checkpointer` must be constructed with a `CheckpointHandler` - essentially as a configuration providing the `Checkpointer` with the logic needed to save and restore your object.\n",
         "\n",
-        "For PyTrees, the most common checkpointable object, we can use the convenient shorthand of `StandardCheckpointer`, which is the same as `Checkpointer(StandardCheckpointHandler())` (see [docs](https://orbax.readthedocs.io/en/latest/api_reference/checkpoint.checkpoint_handlers.html#standardcheckpointhandler) for more info)."
+        "For PyTrees, the most common checkpointable object, we can use the convenient shorthand of `StandardCheckpointer`, which is the same as `Checkpointer(StandardCheckpointHandler())` (see {py:class}`StandardCheckpointHandler <orbax.checkpoint.StandardCheckpointHandler>` for more info)."
       ]
     },
     {
@@ -161,7 +161,7 @@
         "id": "LI-3bCyMEUCC"
       },
       "source": [
-        "It is often necessary to deal with multiple distinct checkpointable objects at once, often with different types. A `Checkpointer` combined with a `CompositeCheckpointHandler` ([docs](https://orbax.readthedocs.io/en/latest/api_reference/checkpoint.checkpoint_handlers.html#compositecheckpointhandler)) can be used to represent a single checkpoint consisting of multiple sub-items, each represented by a sub-directory within the checkpoint."
+        "It is often necessary to deal with multiple distinct checkpointable objects at once, often with different types. A `Checkpointer` combined with a {py:class}`CompositeCheckpointHandler <orbax.checkpoint.CompositeCheckpointHandler>` can be used to represent a single checkpoint consisting of multiple sub-items, each represented by a sub-directory within the checkpoint."
       ]
     },
     {
@@ -274,9 +274,9 @@
         "id": "sYxNt2Zrw_Yt"
       },
       "source": [
-        "In the context of training a model, it is helpful to deal with a series of steps. The `CheckpointManager` allows you to save steps sequentially, according to a given period, cleaning up after a certain number are stored, and many other [functionalities](https://orbax.readthedocs.io/en/latest/api_reference/checkpoint.checkpoint_manager.html#checkpointmanageroptions).\n",
+        "In the context of training a model, it is helpful to deal with a series of steps. The `CheckpointManager` allows you to save steps sequentially, according to a given period, cleaning up after a certain number are stored, and many other functionalities(check {py:class}`CheckpointManagerOptions <orbax.checkpoint.CheckpointManagerOptions>` for more info).\n",
         "\n",
-        "**Beware: `CheckpointManager.save(...)` happens in a background thread by default. See [Asynchronous Checkpointing](https://orbax.readthedocs.io/en/latest/guides/checkpoint/async_checkpointing.html) for more details.**\n",
+        "**Beware: `CheckpointManager.save(...)` happens in a background thread by default. See {doc}`Asynchronous Checkpointing</guides/checkpoint/async_checkpointing>` for more details.**\n",
         "\n"
       ]
     },

docs/index.rst

@@ -211,7 +211,7 @@ Model
          :class-card: sd-text-black sd-bg-light
          :link: api_reference/model.html
 
-    .. grid-item::
+   .. grid-item::
       :columns: 6 6 6 4
 
       .. card:: Roundtripping between Orbax Model and JAX Model