[functions.md] Update np.random → Generator API

[Chihiro2000GitHub]

This PR updates lectures/functions.md as part of the Numpy random API migration described in QuantEcon/meta#299.

Approach

A single rng = np.random.default_rng() is introduced where random draws first appear in the main lecture flow (the "4.4.1. Random Draws" section) and reused across subsequent code cells, following the natural notebook execution order. This avoids redundant redefinition while keeping the changes minimal.

For exercise solution blocks, which are intended to be self-contained, rng = np.random.default_rng() is defined locally at the top of each relevant solution cell. This aligns with the strategy I took in the previous pull request on python_by_example.md: #538

Point for review

In Section 4.4.2. ("Adding Conditions"), the lecture demonstrates passing a callable as an argument to generate_data(). After this migration, the object passed is rng.uniform, which is technically a method rather than a function. As a result, a few sentences related to this - such as "by just passing the desired generator type as a function," "This object is a function", and "bound to the function rng.uniform" - are no longer strictly accurate.

I chose to leave these sentences unchanged for now, as this is an introductory programming lecture and I was uncertain whether introducing terms like "method," "callable," or "function-like object" would help or confuse readers who are likely encountering these concepts for the first time.

I would appreciate the reviewer's guidance on how to handle this wording. I am happy to revise the sentences based on your feedback and will request a re-review after making any requested changes. I will only merge once you are satisfied with the result.

[github-actions]

🚀 Deployed on https://69edbf1bf0d20095cc2b5f37--epic-agnesi-957267.netlify.app