Fix ruff compressions and performance issues

[cclauss]

% ruff check --select=C4,PERF --statistics .

3	C408   	[*] unnecessary-collection-call
2	PERF401	[ ] manual-list-comprehension
1	C400   	[*] unnecessary-generator-list
1	PERF203	[ ] try-except-in-loop
1	PERF403	[ ] manual-dict-comprehension

% ruff rule PERF401

manual-list-comprehension (PERF401)

Derived from the Perflint linter.

What it does

Checks for for loops that can be replaced by a list comprehension.

Why is this bad?

When creating a transformed list from an existing list using a for-loop,
prefer a list comprehension. List comprehensions are more readable and
more performant.

Using the below as an example, the list comprehension is ~10% faster on
Python 3.11, and ~25% faster on Python 3.10.

Note that, as with all perflint rules, this is only intended as a
micro-optimization, and will have a negligible impact on performance in
most cases.

Example

original = list(range(10000))
filtered = []
for i in original:
    if i % 2:
        filtered.append(i)

Use instead:

original = list(range(10000))
filtered = [x for x in original if x % 2]

If you're appending to an existing list, use the extend method instead:

original = list(range(10000))
filtered.extend(x for x in original if x % 2)

[AWhetter]

Please can we leave this on. It seems wasteful to run everything if something has failed.

[cclauss]

Wasteful of compute time or developer time? fail-fast: false saves developers time because they see all failing jobs in a single run and can understand the root cause faster and be more efficient in fixing. If this helps them make fewer commits then it also reduces compute time.

fail-fast: true is a good default for GitHub because it reduces compute time however it forces the developer to go thru multiple commits when debugging multiple problems.,

[AWhetter]

I'm not a fan of multi-line list comprehensions. I think we sacrifice readability for a minimal performance benefit. Please could this stay as-is.

[AWhetter]

The sphinx issue has been addressed and the classifiers have been added since this was written. Please could you correct these.