Revert "[Doc] Refactor DDPG and DQN tutos to narrow the scope (#979)"

This reverts commit c3765cf.

Author: vmoens

docs/source/_static/img/replaybuffer_traj.png

@@ -218,7 +218,7 @@ Check the :obj:`torchrl.envs.utils.check_env_specs` method for a sanity check.
 Utils
 -----
 
-.. currentmodule:: torchrl.data
+.. currentmodule:: torchrl.data.datasets
 
 .. autosummary::
     :toctree: generated/

docs/source/_static/js/theme.js

@@ -114,6 +114,7 @@ provides more information on how to design a custom environment from scratch.
     EnvBase
     GymLikeEnv
     EnvMetaData
+    Specs
 
 Vectorized envs
 ---------------

docs/source/reference/data.rst

@@ -32,7 +32,7 @@ TensorDict modules
 
 Hooks
 -----
-.. currentmodule:: torchrl.modules
+.. currentmodule:: torchrl.modules.tensordict_module.actors
 
 .. autosummary::
     :toctree: generated/

docs/source/reference/envs.rst

@@ -16,15 +16,13 @@ The main characteristics of TorchRL losses are:
   method will receive a tensordict as input that contains all the necessary
   information to return a loss value.
 - They output a :class:`tensordict.TensorDict` instance with the loss values
-  written under a ``"loss_<smth>"`` where ``smth`` is a string describing the
+  written under a ``"loss_<smth>`` where ``smth`` is a string describing the
   loss. Additional keys in the tensordict may be useful metrics to log during
   training time.
   .. note::
     The reason we return independent losses is to let the user use a different
     optimizer for different sets of parameters for instance. Summing the losses
-    can be simply done via
-
-        >>> loss_val = sum(loss for key, loss in loss_vals.items() if key.startswith("loss_"))
+    can be simply done via ``sum(loss for key, loss in loss_vals.items() if key.startswith("loss_")``.
 
 Training value functions
 ------------------------
@@ -218,5 +216,5 @@ Utils
     next_state_value
     SoftUpdate
     HardUpdate
-    ValueEstimators
+    ValueFunctions
     default_value_kwargs

docs/source/reference/modules.rst

@@ -73,7 +73,7 @@ Hooks can be split into 3 categories: **data processing** (:obj:`"batch_process"
 - **Data processing** hooks update a tensordict of data. Hooks :obj:`__call__` method should accept
   a :obj:`TensorDict` object as input and update it given some strategy.
   Examples of such hooks include Replay Buffer extension (:obj:`ReplayBufferTrainer.extend`), data normalization (including normalization
-  constants update), data subsampling (:class:`torchrl.trainers.BatchSubSampler`) and such.
+  constants update), data subsampling (:doc:`BatchSubSampler`) and such.
 
 - **Logging** hooks take a batch of data presented as a :obj:`TensorDict` and write in the logger
   some information retrieved from that data. Examples include the :obj:`Recorder` hook, the reward

docs/source/reference/objectives.rst

@@ -89,10 +89,11 @@ class MockingLossModule(nn.Module):
 
 def mocking_trainer(file=None, optimizer=_mocking_optim) -> Trainer:
     trainer = Trainer(
-        collector=MockingCollector(),
-        total_frames=None,
-        frame_skip=None,
-        optim_steps_per_batch=None,
+        MockingCollector(),
+        *[
+            None,
+        ]
+        * 2,
         loss_module=MockingLossModule(),
         optimizer=optimizer,
         save_trainer_file=file,
@@ -861,7 +862,7 @@ def test_recorder(self, N=8):
         with tempfile.TemporaryDirectory() as folder:
             logger = TensorboardLogger(exp_name=folder)
 
-            environment = transformed_env_constructor(
+            recorder = transformed_env_constructor(
                 args,
                 video_tag="tmp",
                 norm_obs_only=True,
@@ -873,7 +874,7 @@ def test_recorder(self, N=8):
                 record_frames=args.record_frames,
                 frame_skip=args.frame_skip,
                 policy_exploration=None,
-                environment=environment,
+                recorder=recorder,
                 record_interval=args.record_interval,
             )
             trainer = mocking_trainer()
@@ -935,7 +936,7 @@ def _make_recorder_and_trainer(tmpdirname):
                 raise NotImplementedError
             trainer = mocking_trainer(file)
 
-            environment = transformed_env_constructor(
+            recorder = transformed_env_constructor(
                 args,
                 video_tag="tmp",
                 norm_obs_only=True,
@@ -947,7 +948,7 @@ def _make_recorder_and_trainer(tmpdirname):
                 record_frames=args.record_frames,
                 frame_skip=args.frame_skip,
                 policy_exploration=None,
-                environment=environment,
+                recorder=recorder,
                 record_interval=args.record_interval,
             )
             recorder.register(trainer)

docs/source/reference/trainers.rst

@@ -3,7 +3,6 @@
 # This source code is licensed under the MIT license found in the
 # LICENSE file in the root directory of this source tree.
 
-from . import datasets
 from .postprocs import MultiStep
 from .replay_buffers import (
     LazyMemmapStorage,

test/test_trainer.py

@@ -1,2 +1 @@
 from .d4rl import D4RLExperienceReplay
-from .openml import OpenMLExperienceReplay