Revisit docs contents

[BalzaniEdoardo]

Overview of the documentation refactoring plan:

• Home: hyperlinks to some other references.
• Quickstart: See issue Move Overview Chapter of index.md to quickstart.md #220
• Install
• Background: Theory focus, introduce GLM, convolution, basis. See Refactor Background #176
• How-to-Guide: Assume background knowledge, show in depth what you can do with our functions
  • Model temporal effects with basis
  • Model fix non linearities with basis
  • Model multiple inputs
  • Take full advantage of pynapple & NeMoS
  • Regularization: types and when to use them
  • Model selection with sklearn
  • Fit large datasets
• Tutorials: Fully worked out neural examples. Goal: teach good data analysis practices
• API Guide: Assume background knowledge. Explanation of all functions and methods, and their parameters with minimal examples. See issue Add examples to all user-facing docstrings #177
• For Developers: note facilitating new contributors explaining the internals of NeMoS machinery

[billbrod]

Also, see #199

[sjvenditto]

• For Developers:
  • Update page navigation order:
    • "The base_class Module" first
    • "The glm Module" after "The Abstract Class BaseRegressor" module
  • The base_class Module updates
    • Remove Concrete Sublass RecurrentGLM from tree
    • Add Abstract Class Basis to tree
    • Either update concrete subclasses (add in Basis subclasses, add PopulationGLM, include additional Regularizer subclasses) or remove concrete subclasses and only include on individual pages
  • The Abstract Class BaseRegressor updates
    • Add subclass tree
    • Link directly to to developer notes for glm Module in Example card
  • The glm Module updates
    • Add subclass tree
    • Fix typo in figure legend: "A is a subclass if B" -> "A is a subclass of B"
    • (optional) Flip arrow direction in figure to point towards subclasses (e.g. have arrows indicate "A is a parent of B") (this directionality is more intuitive to me)
  • The observation_models Module updates
    • Add subclass tree
    • Solvers
      • Add an how to use SVRG guide
      • Modify batching not comparing gradient descent and SVRG