Fix flux diagram naming of species images

[sevyharris]

There can be multiple species that end the same way, so using endswith is not guaranteed to get you the right images. This change makes sure the flux diagrams will map the correct species image onto the flux diagram.

Motivation or Problem

The flux diagram tool sometimes grabs the wrong image for a species. See #2904

The problem is that it's relying on the filename ending the same as the species name, instead of matching exactly. In the above example, you get RMG using the image of C4H71-O2 for O2.

Description of Changes

Instead of searching for a file name by checking that it ends with the right key, it now forces the file name to match exactly.

Testing

I tested this using the example in ReactionMechanismGenerator/RMG-Py/issues/2904

The bad "before" image is shown in the bug, and the fixed "after" image is shown below. You can see O2 is now displaying properly.

I will also test this for the case of nested directories. If you put the O2 image inside another folder, does it still work?
Yes. I tested and this works fine.

[rwest]

I was curious how this came to be. Why had someone deliberately chosen endswith instead of == ?

It's because when Josh first wrote it in 1307f71, species_index was just the number at the end like (42) concatenated with .png so endswith() would correctly match anything(42).png.

Then, in order to fix something else, in daab0fb Connie made species_index equal to the species label instead of the number. This bug hid silently.

This fix is clearly correct. I'll rebase (as it's a single commit) and merge.