Add complete list of supported bioformats filetypes

[GenevieveBuckley]

I was trying to open a deltavision .dv image file the other day and found that the bioformats reader doesn't recognise all bioformats filetypes. This means that the pims.open() command won't call out to the bioformats reader.

Even ahough you can force it to open correctly by calling pims.Bioformats() instead of pims.open(), this isn't ideal for downstream libraries/users (eg: dask-image uses pims.opne() and we wouldn't want to implement any extra logic around trying to pick the pims reader ourselves, since pims does that much better).

I got the list of bioformats supported file extensions from https://docs.openmicroscopy.org/bio-formats/5.7.1/supported-formats.html

There is some overlap with file extensions recognized by other readers, but that was true of the last set too & it shouldn't matter since I think the Bioformats reader is tried last.

[nkeim]

This is a good idea! I would love to see it in v0.6.

[GenevieveBuckley]

Note: this PR needs to specify a relative path to bioformats_filetypes.txt, then the CI should pass

[tacaswell]

"power-cycled" to get new CI.

@GenevieveBuckley Is this list still accurate (2 years later)?

[nkeim]

Looks like this code needs to find bioformats_filetypes.txt via the module path. Once that's done I'm for merging; even if the list is a bit outdated it's still a big enhancement.

[GenevieveBuckley]

"power-cycled" to get new CI.

@GenevieveBuckley Is this list still accurate (2 years later)?

I don't know, I haven't looked at it since then.

You could cross check it against the latest (or whichever version pims currently pins to, which is not the one I originally looked at anymore I don't think) https://docs.openmicroscopy.org/bio-formats/latest/supported-formats.html

Alternatively, perhaps it would be better to link directly to that page, either in addition to or instead of maintaining a separate copy of the list here. That would reduce the maintenance burden, although I guess dead links are always a possibility.

[tacaswell]

Given that this is tied to a version of bioformats and #403 is changing that, I think we should do both together.

I do not think we want have a network hit on import, it is probably best to just embed this list as a (very long) literal in the source (the test failures are failing to open the txt file, not sure if it is cwd related or packaging related...) .