Novice notebooks updates

[matt-graham]

Updates to the notebooks performed before last run of 'novice' course in February 2022

• In 000Exemplar.ipynb notebook changes incorrect references to PNG image to JPEG image and removes use of BytesIO in reading images from bytes.
• In 010Internet.ipynb notebook updates Google Maps URLs due to deprecation of previous API, adds more detail to description of URI scheme, minor updates to example usages of Requests library.
• In 020TabularData.ipynb notebook adds extra description of tabular data formats and delimter-separated value files specifically and switches to using in-built csv module to read CSV rather than NumPy to avoid having to explain NumPy concepts before main NumPy introduction.
• In 030StructuredData.ipynb removes !pip install pyyaml shell command in notebook to avoid OS compatibility issues and added exemplar maze structure from beginners course exercise for final exercise to allow for students who haven't done beginners course and so don't have own definition.
• In 070Classes.ipynb small tweaks to remove explicit subclassing from object, use format strings rather than string concatenation and improve formatting of code / Markdown text.
• In Makefile updates to list of files to clean to reflect new names.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[dpshelio]

Way removed the only docstring we had? Yeah, weird we have only one... but it may be the useful one.

Also, on step the two prints were to distinguish before from after. Maybe the print in the middle could be added with a separator, like a ------ or something.

Reply via ReviewNB

[matt-graham]

I partly removed as it the previous docstring was using a plain string "..." rather than a multiline string """...""" , which while allowed and a valid docstring, is non-conventional . We also don't introduce docstrings previously at all I think so having one single example here without explanation I feel is more hindrance than help as I don't think it's clear this is a docstring and its not obvious why we would choose to add only this method. I've now updated to add docstrings to all the methods.

[dpshelio]

Pretty good!! Some small comments left on the reviewNB. But approving it so you can merge it straight away if you use these suggestions.

[matt-graham]

Pretty good!! Some small comments left on the reviewNB. But approving it so you can merge it straight away if you use these suggestions.

Thanks for the review and suggestions! Now made changes to address all of the comments.

[matt-graham]

Build website job is failing with error

ImportError: cannot import name 'contextfilter' from 'jinja2'

which seems to be as the current version of nbconvert pinned in the requirements.txt file (nbconvert==5.6.1) doesn't include this fix for changes to the jinja2 API (see for example this related issue).

I will try pinning to a newer version with the fixed included.

[matt-graham]

In the end trying to use an updated nbconvert version ended up hitting up against #19 and after going down a rabbit hole trying to fix that I've decided for now to just pin jinja2==3.0 which appears to fix the immediate issue, and delay sorting the broader issue of updating the build system dependencies to another PR...