Add structure for mkdocs instead of Sphinx

README.md

@@ -88,7 +88,7 @@ The framework consists of four main parts, as demonstrated in the flowchart belo
 - ActiveLearningLoop
 
 <p align="center">
-  <img src="./docs/literature/images/Baalscheme.svg">
+  <img src="docs/research/literature/images/Baalscheme.svg">
 </p>
 
 To get started, wrap your dataset in our _[**ActiveLearningDataset**](baal/active/dataset.py)_ class. This will ensure

docs/Makefile

@@ -1,23 +1,10 @@
-# Minimal makefile for Sphinx documentation
+# Minimal makefile for Mkdocs documentation
 #
 
-# You can set these variables from the command line.
-SPHINXOPTS    =
-SPHINXBUILD   = sphinx-build
-SOURCEDIR     = .
-BUILDDIR      = _build
-
 
 # Put it first so that "make" without argument is like "make help".
-help:
-	@$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
-
-.PHONY: help Makefile
-
-# Catch-all target: route all unknown targets to Sphinx using the new
-# "make mode" option.  $(O) is meant as a shortcut for $(SPHINXOPTS).
-%: Makefile
-	@$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
+build:
+	mkdocs build
 
 server:
-	open _build/html/index.html
+	mkdocs watch

docs/api/bayesian.md

@@ -22,11 +22,10 @@ model = MCDropoutConnectModule(model, layers=["Linear"], weight_dropout=0.5)
 
 ## API
 
-```eval_rst
-.. autoclass:: baal.bayesian.dropout.MCDropoutModule
-    :members: __init__
+### baal.bayesian.dropout.MCDropoutModule
 
-..autoclass:: baal.bayesian.weight_drop.MCDropoutConnectModule
-    :members: __init__
+::: baal.bayesian.dropout.MCDropoutModule
 
-```
+### baal.bayesian.weight_drop.MCDropoutConnectModule
+
+::: baal.bayesian.weight_drop.MCDropoutConnectModule

docs/api/calibration.md

@@ -1,6 +1,5 @@
 # Calibration Wrapper
 
-```eval_rst
-.. autoclass:: baal.calibration.DirichletCalibrator
-    :members:
-```
+### baal.calibration.DirichletCalibrator
+
+::: baal.calibration.DirichletCalibrator

docs/api/compatibility/huggingface.md

@@ -1,10 +1,7 @@
 ## HuggingFace Compatibility
 
- ```eval_rst
-.. autoclass:: baal.transformers_trainer_wrapper.BaalTransformersTrainer
-    :members: predict_on_dataset, predict_on_dataset_generator
+**baal.transformers_trainer_wrapper.BaalTransformersTrainer**
+::: baal.transformers_trainer_wrapper.BaalTransformersTrainer
 
-.. autoclass:: baal.active.nlp_datasets.HuggingFaceDatasets
-    :members:
-
-```
+**baal.active.dataset.nlp_datasets.HuggingFaceDatasets**
+::: baal.active.dataset.nlp_datasets.HuggingFaceDatasets

docs/api/compatibility/pytorch-lightning.md

@@ -1,12 +1,10 @@
 ## Pytorch Lightning Compatibility
 
- ```eval_rst
-.. autoclass:: baal.utils.pytorch_lightning.ResetCallback
-    :members: on_train_start
+**baal.utils.pytorch_lightning.ResetCallback**
+::: baal.utils.pytorch_lightning.ResetCallback
 
-.. autoclass:: baal.utils.pytorch_lightning.BaalTrainer
-    :members: predict_on_dataset, predict_on_dataset_generator
+**baal.utils.pytorch_lightning.BaalTrainer**
+::: baal.utils.pytorch_lightning.BaalTrainer
 
-.. autoclass:: baal.utils.pytorch_lightning.BaaLDataModule
-    :members: pool_dataloader
-```
+**baal.utils.pytorch_lightning.BaaLDataModule**
+::: baal.utils.pytorch_lightning.BaaLDataModule

docs/api/dataset_management.md

@@ -36,13 +36,11 @@ assert al_dataset.pool.transform is None
 
 ### API
 
-```eval_rst
-.. autoclass:: baal.active.ActiveLearningDataset
-    :members:
+### baal.active.ActiveLearningDataset
+::: baal.active.ActiveLearningDataset
 
-.. autoclass:: baal.active.ActiveLearningLoop
-    :members:
+### baal.active.ActiveLearningLoop
+::: baal.active.ActiveLearningLoop
 
-.. autoclass:: baal.active.FileDataset
-    :members:
-```
+### baal.active.FileDataset
+::: baal.active.FileDataset

docs/api/heuristics.md

@@ -33,13 +33,14 @@ BALD(reduction="mean")
 
 ### API
 
-```eval_rst
-.. autoclass:: baal.active.heuristics.AbstractHeuristic
-    :members:
+### baal.active.heuristics.AbstractHeuristic
+::: baal.active.heuristics.AbstractHeuristic
 
-.. autoclass:: baal.active.heuristics.BALD
+### baal.active.heuristics.BALD
+::: baal.active.heuristics.BALD
 
-.. autoclass:: baal.active.heuristics.Random
+### baal.active.heuristics.Random
+::: baal.active.heuristics.Random
 
-.. autoclass:: baal.active.heuristics.Entropy
-```
+### baal.active.heuristics.Entropy
+::: baal.active.heuristics.Entropy

docs/api/index.md

@@ -1,25 +1,18 @@
 # API Reference
 
-```eval_rst
-.. toctree::
-    :caption: API Definition
-    :maxdepth: 1
-
-    baal.modelwrapper.ModelWrapper <./modelwrapper>
-    baal.bayesian <./bayesian>
-    baal.active <./dataset_management>
-    baal.active.heuristics <./heuristics>
-    baal.calibration <./calibration>
-    baal.utils <./utils>
-
-.. toctree::
-    :caption: Compatibility
-    :maxdepth: 1
-
-    baal.utils.pytorch_lightning <./compatibility/pytorch-lightning>
-    baal.transformers_trainer_wrapper <./compatibility/huggingface>
-
-```
+### :material-file-tree: API Definition
+
+* [baal.modelwrapper.ModelWrapper](./modelwrapper.md)
+* [baal.bayesian](./bayesian.md)
+* [baal.active](./dataset_management.md)
+* [baal.active.heuristics](./heuristics.md)
+* [baal.calibration](./calibration.md)
+* [baal.utils](./utils.md)
+
+### :material-file-tree: Compatibility
+
+* [baal.utils.pytorch_lightning] (./compatibility/pytorch-lightning)
+* [baal.transformers_trainer_wrapper](./compatibility/huggingface)
 
 
 

docs/api/modelwrapper.md

@@ -32,7 +32,6 @@ predictions.shape
 
 ### API
 
-```eval_rst
-.. autoclass:: baal.ModelWrapper
-    :members:
-```
+### baal.ModelWrapper
+
+::: baal.ModelWrapper

docs/api/utils.md

@@ -5,7 +5,6 @@
 
 To work with `baal.modelwrapper.ModelWrapper`, we provide `Metrics`.
 
-
 ### Examples
 
 ```python
@@ -41,7 +40,6 @@ print(wrapper.active_learning_metrics)
 """
 ```
 
-```eval_rst
-.. automodule:: baal.utils.metrics
-    :members:
-```
+### baal.utils.metrics
+
+::: baal.utils.metrics