www.lesswrong.com/posts/3pdyxFi6JS389nptu/is-programbench-impossible
1 correction found
The program has two sub-commands hrum and kfreq
The seqtk subcommand is spelled `hrun`, not `hrum`. ProgramBench’s own seqtk tests use `TestHrunCommand`, and Heng Li’s examples invoke `seqtk hrun ...`.
Full reasoning
This appears to be a typo in the command name.
ProgramBench's own seqtk task tests refer to TestHrunCommand and test_hrun_basic, which indicates the benchmarked subcommand is hrun, not hrum.
Separately, an example gist by Heng Li (the author of seqtk) uses the command seqtk hrun hs37d5.fa, again confirming the command name is hrun.
Because command names are exact strings in a CLI, this misspelling could mislead readers trying to inspect or reproduce the behavior discussed in the post.
2 sources
- ProgramBench seqtk task tests.json
... "tests.test_merge_split.TestHrunCommand.test_hrun_basic", "tests.test_merge_split.TestHrunCommand.test_hrun_min_length", "tests.test_merge_split.TestHrunCommand.test_hrun_output_format" ...
- Scripts and command lines to create universal mask for hs37d5 · GitHub
... # long homopolymer runs seqtk hrun hs37d5.fa \ | bgzip > 01compositional/hs37d5.hrun.bed.gz ...