Fix alpha handling when flattening images for vector backends - #32107

PR summary

This PR fixes a bug when flattening images for vector backends and a non-unity scalar alpha is specified. Image flattening is when the Agg resampler is used to flatten a sequence of images into a single image so that only a single image needs to be drawn by the backend, and is enabled by default. When an image is added with a scalar alpha (i.e., imshow(..., alpha=...)), that scalar alpha is mistakenly applied twice during flattening: (1) when preparing the image to be resampled, and (2) when resampling the prepared image into the output buffer. For example, an image that is intended to have an alpha of 0.6 ends up in the flattened image as if it had an alpha of 0.36 (0.6 squared). This is the root cause of #8847.

I also added a clarifying comment for future maintainers because I got tripped up with inspecting the code. The Agg resampler typically assumes that the input image is in premultiplied alpha space (cf. #31021 (comment)), except when the input image is uint8 (code). I forgot about that exception.

Fixes #8847

AI Disclosure

PR checklist

• "closes #0000" is in the body of the PR description to link the related issue
• new and changed code is tested
• [N/A] Plotting related features are demonstrated in an example
• [N/A] New Features and API Changes are noted with a directive and release note
• [N/A] Documentation complies with general and docstring guidelines