Make the doc prettier

docs/basics/comparison.md

@@ -131,3 +131,5 @@ print(compare_agents(agents))
 The results of `compare_agents(agents)` show the p-values and significance level if the method is tukey_hsd and  it shows the decision accept or reject of the test with Family-wise error controlled by $0.05$. In our case, we see that DQN is worse than A2C and PPO but the difference between PPO and A2C is not statistically significant. Remark that no significance (which is to say, decision to accept $H_0$) does not necessarily mean that the algorithms perform the same, it can be that there is not enough data (and it is likely that it is the case here).
 
 *Remark*: the comparison we do here is a black-box comparison in the sense that we don't care how the algorithms were tuned or how many training steps are used, we suppose that the user already tuned these parameters adequately for a fair comparison.
+
+*Remark 2*: the comparison we do here is non-adaptive. To go further and use as little seeds as possible, look at [AdaStop](adastop_userguide).

docs/conf.py

@@ -36,6 +36,7 @@
 # ones.
 extensions = [
     "sphinx.ext.doctest",
+    "sphinx_design",
     "sphinx.ext.todo",
     "sphinx.ext.viewcode",
     "sphinx.ext.autodoc",

docs/index.md

@@ -18,6 +18,57 @@ We provide you a number of tools to help you achieve **reproducibility**, **stat
  If you begin with [rlberry](https://github.com/rlberry-py/rlberry), **check our** [RL quickstart](quick_start) **and our** [Deep RL quickstart](TutorialDeepRL).
 
 
+``````{grid} 3
+:gutter: 3 3 3 3
+```{grid-item-card} Comprehensive workflow for RL research
+:link: quick_start
+:link-type: ref
+:img-bottom: thumbnails/experiment_manager_diagram.png
+:width: 90
+
+Tools designed to facilitate RL experimentation.
+```
+```{grid-item-card} Statistical tests
+:link: comparison_page
+:link-type: ref
+:img-bottom: thumbnails/adastop_boxplots.png
+:width: 90
+
+Statistical tools for agent comparison.
+
+```
+```{grid-item-card} Visualization
+:link: visu_userguide
+:link-type: ref
+:img-bottom: thumbnails/output_9_3.png
+:width: 90
+
+Proper smoothing and plot of confidence intervals.
+```
+```{grid-item-card} Teaching tools
+:link: https://github.com/rlberry-py/rlberry-scool
+:img-bottom: thumbnails/chain_thumb.jpg
+:width: 90
+
+Simple environments and tools for use in teaching RL.
+```
+```{grid-item-card} Hackable library
+:link: https://rlberry-py.github.io/rlberry/basics/userguide/agent.html#create-your-own-agent
+:width: 90
+:img-bottom: thumbnails/code.png
+
+
+Code is easy to read and easy to change.
+```
+```{grid-item-card} Compatible
+:link: user_guide
+:link-type: ref
+:width: 90
+
+Uses *stable-baselines*, *tensorboard*, *gymnasium*, and you can easily write your wrapper to use your own RL library. Visualization and statistical tool can be used with dataframes as input/output in order to be library-independent.
+```
+``````
+
 ## Documentation Contents
 You can find main documentation here :
 - [Installation](installation)

docs/requirements.txt

@@ -6,3 +6,4 @@ myst-parser
 git+https://github.com/sphinx-contrib/video
 matplotlib
 sphinx-copybutton
+sphinx-design

pyproject.toml

@@ -47,6 +47,7 @@ sphinx-gallery = { version= "^0.14.0", optional=true}
 sphinx-math-dollar = {version="^1.2.1", optional=true}
 sphinxcontrib-video = {version="0.2.0", optional=true}
 sphinx-copybutton = "*"
+sphinx-design = "*"
 numpydoc = {version="^1.6.0", optional=true}
 myst-parser = {version="^2.0.0", optional=true}
 
@@ -81,6 +82,7 @@ sphinx-math-dollar = {version="^1.2.1", optional=true}
 numpydoc = {version="^1.6.0", optional=true}
 myst-parser = {version="^2.0.0", optional=true}
 sphinx-copybutton = "*"
+sphinx-design = "*"
 
 
 [build-system]