Invalid indices in make_index_files.R assignment of non-HLA transcript sequences

[jfass]

I'm working in a conda virtual environment, using R 4.0.5. dplyr gets updated to 1.0.9 when installing hlaseqlib, and gives a warning msg about "group_by" and a change as of dplyr 1.0.0 ... which may or may not be related to this issue (I'm not well versed in the tidyverse). When stepping through the make_index_files.R to troubleshoot, it's the assignment of transcript_no_hla that fails, with a message about invalid indices (sorry I don't have the exact text in front of me). I solved with classic R:

transcripts_no_hla <- transcripts[ which( names( transcripts ) %in% transcripts_db$transcript_id[ which( !( transcripts_db$gene_name %in% hladb_genes ) ) ] ) ]

... hopefully that gives you an idea of what the problem could be? I'm using a gencode v21 protein_coding trancripts fasta and annotation for main chromosomes. Before isolating the problematic step, I tried truncating the ENS[GT] id's to get rid of the ".#" (version) in both annotation and fasta, and removing all non-protein_coding annotations from the gtf file, but neither changed the error. So I think the problem is in some changed dplyr syntax, maybe, and the versions I'm using.

[jfass]

Here's the warning (doesn't seem to cause a problem once I've substituted the command above):

Processing locus DRB4
Processing locus DRB5
Warning message:
Problem while computing `data = map(locus, ~hla_compile_index(., imgt_db))`.
ℹ The `...` argument of `group_indices()` is deprecated as of dplyr 1.0.0.
  Please `group_by()` first 
Reading transcript annotations...
writing index files...
Done!

[youknow16]

I'm getting a similar error except it just exits the program:

  Warning message:
  Problem while computing `data = map(locus, ~hla_compile_index(., imgt_db))`.
  ℹ The `...` argument of `group_indices()` is deprecated as of dplyr 1.0.0. Please `group_by()` first 
  Reading transcript annotations...
  Error: subscript contains invalid names
  Execution halted

[jfass]

Yah I think I was seeing execution die as well, with the "invalid indices" message. Have you had a chance to try my fix above? There's probably a better, (older) tidyverse way to do it, but ...