BUG: Properly validate FASTA files

[Oddant1]

closes #187

[ebolyen]

Minor nitpicks, and a potential strategy for testing the buffer boundaries (which would be good to demonstrate we 100% understand the way that is working).

[ebolyen]

This is indeed much better! Maybe we should think about how we could generalize this for other formats in the back of our heads (obviously not for this PR).

[ebolyen]

Two small updates and this is ready I think! Thanks @Oddant1!

[ebolyen]

Nice work!

[thermokarst]

Hey @Oddant1, can you take a look at this forum post: https://forum.qiime2.org/t/training-feature-classifier-2019-10-error/13025/

It looks like some of the new validation you implemented in this PR is being hit by these users, although it looks like maybe they don't have quite enough information in the error message to fix anything. Also, it looks like maybe the validation isn't quite working as expected? I'm not sure, I would need to take a closer look at this code to fully understand, but wanted to ping you first to see if you have any ideas. Thanks!

[Oddant1]

@thermokarst, apologies for the late reply, I have a theory I will be posting on the forum. I'm not 100% sure it's correct as I was strangely unable to exactly replicate the errors either user experienced, but I got something similar. Also yes I think it's a good idea to make it so the error also prints what character is erroneous and ideally also where said character is within the line.

[Oddant1]

Apologies for forgetting to update this asap. Based on the responses of the forum users in the forum thread, I believe my theory was correct, and they had trailing whitespace causing validation errors. The user then ran into what (I think) was a largely unrelated issue that @nbokulich helped them with.