Add more introductory pages (and update "create custom environment")

[pseudo-rnd-thoughts]

Description

In reviewing the website, I realised that we currently have a pretty bad set of introductory pages for new users.
Therefore, this PR adds several key pages that I believe will be very helpful for new users:

• basic usage - Same as before just updated and addresses [Bug Report] "Basic usage" documentation incorrect/incomplete #735
• training an agent - Most users will be interested in using a built-in gymnasium environment to train a basic RL agents to begin with. The easiest solution is to move the "Solving blackjack using Q-learning" tutorial with minor changes. We should look to add tutorials for training agents to play Atari and Mujoco environments.
• create a custom environment - The second most common thing for users should be creating their own custom environment. The easiest solution again is to move the "Creating a custom environment" tutorial but this tutorial needs large scale changes (@kir0ul is working on this). We should look to add a section at the end linking to third-party environments.
• Recording agents - While uncommon, I think having a page about recording data on agents during training and evaluation is helpful. In particular how to use the RecordVideo and RecordEpisodeStatistics.
• Speeding up environments - A bit more niche but critical is using vector environments. I think "speeding up environment" is more helpful to new users than "how to use vector environments" and we can note other techniques that can be used.
• Compatibility with OpenAI Gym - Same as before (show be updated for v1.0.0)
• Migration guide - Same as before but should include updating to v1.0.0

Todo

• Add compiling from python to rst (we support this for tutorials but we don't want the gallery aspect)
• Finish "Recording agents"
• Modify "Speeding up the environment" to add an explanation of vector environments
• Update training an agent to be more introductory

[pseudo-rnd-thoughts]

closed for #791