For this package all the CSV files are relatively small I think.
CSV is a heavy dependency, on startup. It pays off if you load large CSV files. I was thinking maybe you could rather use the really fast (built into Base), it seems even with compressed files:
julia> @time using DelimitedFiles
0.004993 seconds (1.15 k allocations: 82.234 KiB, 78.42% compilation time)
A downside could be: If in all real-world code you would load CSV anyway, then not effective (but neither slower this way). Another option could be, this package if often (not always) used with RCall (what I'm looking into now). And plausibly then at least you could rather use CSV reading from R than Julia's CSV. Or Python's tool where appropriate.
For this package all the CSV files are relatively small I think.
CSV is a heavy dependency, on startup. It pays off if you load large CSV files. I was thinking maybe you could rather use the really fast (built into Base), it seems even with compressed files:
https://docs.julialang.org/en/v1/stdlib/DelimitedFiles/
A downside could be: If in all real-world code you would load CSV anyway, then not effective (but neither slower this way). Another option could be, this package if often (not always) used with RCall (what I'm looking into now). And plausibly then at least you could rather use CSV reading from R than Julia's CSV. Or Python's tool where appropriate.