Repeated sequence not being captured entirely, intermittent data issues

[wlymanambry]

Describe the bug
I'm using seqrepo to load many small protein sequences, on the order of millions. What we are finding is that the data loaded does not always match the input data that was loaded. We are still investigating but it appears to be missing repeated sequence (of *) at the beginning of the sequence. As an example, this file was loaded:

but the returned sequence only contains 7 asterisks instead of the expected 36:

This doesn't appear to be a one off. We are seeing a good amount of this in our testing.

To Reproduce
Whatever is causing this doesn't appear to be reproducible. Reloading the sequence a second time appears to be resolve the issue.

Expected behavior
The loaded sequences + reported sequence would match the input sequence.

[wlymanambry]

Sorry, I should have added, if I check the compressed file that was written, it has also omitted some of the repeating sequence:
Preloaded sequence:

Compressed post loaded sequence: