Errors in `highquality_clust30`

[valentynbez]

The database contains empty structures which lead to segfault if one tries to extract them.
The list of empty structures was obtained by running:

foldcomp check highquality_clust30 --threads 32 &> error_ids.txt

error_ids.txt

[khb7840]

Thank you for notifying this. I'm checking this database along with other issues.

[valentynbez]

Encountered the same error when creating my own database, it's not clear what is the feature of the structures that fail to compress.

[valentynbez]

In both cases there are sequences split on weird alphabet letter, supposedly X:

>phrog_433:protein63813
M
--
>phrog_433:protein63813
KCRKKIFLYREDGTEDIKVIKYKDNVNEVYSLTGAHFSDEKKIMTDSDLKRFKGAHGLLYEQELGLQATIFDI

>MGYP003343806611
MLRIKITDADRAGRAGEWCQANLGRDDWNLYGHNLFTGTPYYEFEFTDSETAMMFALRWA
--
>MGYP003343806611
YY

It appears to be related to #53 and the way AA_UNK_CHAR is processed.