Define a simpler PenguinDataset in Exercise 1 notebook

[ma595]

This PR creates a simpler PenguinDataset within the worked-solutions/01_penguin_classification_solutions.ipynb notebook.

Within this easier PenguinDataset we:

• Move ontents of _load_penguin_data(...) and _split_data(...) into __init__()
• Remove lambdas and Compose object, making it easier to follow.
• Remove docstrings (I am aware that this isn't ideal, but I think it is fine for this purpose). Any objections, please comment.

Other notebook changes include:

• The unchanged _penguins.py has been referenced in the notebook as an example of good software practice.
• Markdown has been modified to make the notebook easier to follow in light of the above changes.
  • Mention of why data reading code is encapsulated in a class
• All discussion on Compose has been removed.

Before merging, the following tasks must be completed:
Propagate changes to:

• exercise notebook
• colab - (is NOT done automatically - This is done by rebasing or merging into the colab branch - consider doing this after merging to main.)
• ex 2+? Consider leaving this as is.

Closes #71

[ma595]

@jatkinson1000 would you prefer the jupyter notebook outputs to be retained?

[ma595]

Jack confirmed that we should retain the cell outputs of the solutions notebook for reference.

[jatkinson1000]

This all looks sensible to me in theory @ma595

Caveat that I have not run the code and will leave this (and approval) to @surbhigoel77 as the co-presenter.

I agree that it would be easier to move changes to the colab branch after merging to main.

Also consider how other exercises might need adapting.

[surbhigoel77]

Runs fine for me,