gh-159675: Use lazy urllib imports in saxutils

[jelly]

Both urllib imports are only used for prepare_input_source, normally not used when parsing or writing XML files.
A real life example of the benefit of lazy loading is the "firewalld" daemon on Fedora/RHEL Linux systems where lazy loading urllib shaves off 3MB memory.

---

Alternatively I can just move the imports to prepare_input_source if desired and drop the test case.

• Issue: Lazy load urllib modules in saxutils #159675

[python-cla-bot]

All commit authors signed the Contributor License Agreement.

[bedevere-app]

Most changes to Python require a NEWS entry. Add one using the blurb_it web app or the blurb command-line tool.

If this change has little impact on Python users, wait for a maintainer to apply the skip news label instead.

[jelly]

"Benchmarks":

Before:

[jelle@natrium][~/projects/cpython]%hyperfine -w 5 "./python -c 'import xml.sax.saxutils'"
Benchmark 1: ./python -c 'import xml.sax.saxutils' (85 runs)
                mean     ±       σ          min     …     max
  Wall Time     34.3 ms  ±     1.4 ms      32.6 ms  …    42.0 ms
  Memory        20.3 MiB ±     0.2 MiB     19.7 MiB …    20.6 MiB

After

[jelle@natrium][~/projects/cpython]%hyperfine -w 5 "./python -c 'import xml.sax.saxutils'"
Benchmark 1: ./python -c 'import xml.sax.saxutils' (271 runs)
                mean     ±       σ          min     …     max
  Wall Time     10.6 ms  ±     0.5 ms       9.1 ms  …    11.6 ms
  Memory        10.2 MiB ±     0.1 MiB      9.8 MiB …    10.4 MiB

[aisk]

The class name is a little bit misleading, it doesn't test the import time at all. Maybe we can put the test function in TestModule in this file.

[jelly]

This makes sense, adjusted and fixed the PEP8 issue

[bedevere-app]

Most changes to Python require a NEWS entry. Add one using the blurb_it web app or the blurb command-line tool.

If this change has little impact on Python users, wait for a maintainer to apply the skip news label instead.