Input either a file path or character vector—how to implement?

[Aariq]

The proposed refactoring to allow input of other chemical representations besides molfiles means that we have two categories of possible inputs that are pretty different. The mol format exists either as a text file on disk or as a "SDFset" S4 object in R when it gets read in by ChemmineR::read.SDFset(). SMILES, InChI, and other identifiers are going to be character vectors.

I don't think I've ever encountered an R function that accepts either a character vector or an R object or one that takes a character vector that could be a literal input but might be a file path. I'm not sure what the best way to handle this is. Some options:

1. Use a from argument to specify what format the input is in. This is how a lot of functions in webchem work, although we've never mixed file paths and IDs before.

calc_vol(input = "Clc(c(Cl)c(Cl)c1C(=O)O)c(Cl)c1Cl", from = "smiles")
calc_vol(input = "data/map00361/C16181.mol", from = "mol_path")

2. Two arguments, exactly one of which is required

calc_vol(id_string = "Clc(c(Cl)c(Cl)c1C(=O)O)c(Cl)c1Cl",  from = "smiles")
calc_vol(mol_path = "data/map00361/C16181.mol")

We would still need from= to distinguish between SMILES and InChI, for example

3. Two separate functions

calc_vol_file("data/map00361/C16181.mol")
calc_vol_id("Clc(c(Cl)c(Cl)c1C(=O)O)c(Cl)c1Cl", from = "smiles")

I think I like the first option best, personally, but I want to make sure I get user input on this.

[Aariq]

We could also make the user read in the .mol file and supply an SDFset object as input rather than a file path. This would be advantageous if people work with SDFset objects a lot in-memory so they wouldn't have to write them to a file before passing to calc_vol(). I don't know how common of a workflow that is though.

[Aariq]

Because the only options in ChemmineR are to read .mol or convert smiles (i.e. no InChI option), better to do this for the user and go with method 1.