[echoix]

I tested out GitHub’s Code Quality tool based on CodeQL, from my second fork (an org) in order to enable it.
It also allowed some AI findings on changed files, and also was able to create the suggestions and open a PR, however, only one change per PR. So, I queued up most of the suggestions I agreed on for one single problem
(I wasn’t 100% sure for 3 of the 34-36 instances).

The PRs opened from the dashboard of issues essentially contained:

Use a specific comparison assertion instead of a boolean assertion for relational checks.

Best fix (no functionality change): in imagery/i.gensig/testsuite/test_i_gensig.py, replace line 142:

• from: self.assertTrue(abs(Sn.sig[0].var[1][1] - 0.6) < 0.1)
• to: self.assertLess(abs(Sn.sig[0].var[1][1] - 0.6), 0.1)

This preserves behavior while improving failure messages and aligns with nearby assertions already using assertLess. No imports, helpers, or additional definitions are needed.

Suggested fixes powered by Copilot Autofix. Review carefully before merging.

So, I’m noting that this was made with AI, but the usage was checked against the unittest docs by me, and the changes are trivial. I excluded the 3 suggestions where I wasn’t sure if it was the equality operation itself being tested.