Should there be a warning for improperly formatted CAS numbers and/or should they be automatically reformatted?

[ethanbass]

It seems that most webchem functions do not currently check the formatting of CAS numbers. For example, cts_convert("1333740", from='CAS', to="ChemSpider") returns NA, but if you reformat the CAS number cts_convert(webchem::as.cas("1333740"), from='CAS', to="ChemSpider") the function returns several chemspider IDs. However, some functions (like nist_ri) internally reformat the CAS number by running it through the as.cas function.

While CAS numbers are normally written with hyphens, it is also fairly common to see CAS numbers "in the wild" written as one long string of numbers without hyphens. I am thinking it would be helpful to throw a warning for improperly formatted CAS numbers and/or automatically run them through as.cas to ensure they are formatted correctly, rather than returning NA for CAS numbers that are "improperly" formatted.

[stitam]

Thanks @ethanbass, I like the idea to automatically format everywhere, the impact would be less characters of code to reach the end result and less frustration. Maybe as.cas() should throw a message (if verbose) if it cannot convert a string to CAS? Wondering if there is anything against automatic formatting?

[Aariq]

is.cas() already provides messages giving details of why something isn't a CAS number. as.cas() already calls is.cas()—it would just need a verbose argument added to pass down to is.cas() probably, or vebose=TRUE could be hard-coded inside of as.cas()

e.g.:

library(webchem)
is.cas("12-3-12", verbose = TRUE)
#> 12-3-12: Second part should have two digits!
#> 12-3-12 
#>   FALSE

^{Created on 2023-05-03 with reprex v2.0.2}

[ethanbass]

I took a crack at this. Working on a pull request

[ethanbass]

#398