toml issue on df_basis

[MilitaoLucas]

Hello, I am having the following issue on toml when reading the following file:

threads = 30
verbosity = 2

[scf]
input = "cucomplex.xyz"
method = "r2scan"
basis = "def2-tzvp"
charge = 0
multiplicity = 2
output = "fchk"
spherical = true
df_basis = "def2-universal-jkfit"

This gives out:
INI was not able to parse scf.df_basis

threads = 30
verbosity = 2

[scf]
input = "cucomplex.xyz"
method = "r2scan"
basis = "def2-tzvp"
charge = 0
multiplicity = 2
output = "fchk"
spherical = true
density_fitting_basis = "def2-universal-jkfit"

this seems to work fine.

[peterspackman]

Pretty sure it needs to be df-basis, it's based on the command line arguments (as literal strings). I can add --df_basis as an argument to address this in a future version.

Longer term my preference would be to migrate to something more robust/flexible than the CLI/toml input files eventually to be honest. Embedding lua is my current preference. Though supporting both a python API/interface as well as an embedded one will be a bit of a nightmare.

[MilitaoLucas]

This example uses df_basis
https://github.com/peterspackman/occ/wiki/Command-line-interface#molecular-scf-calculations
I think maybe standardizing using either _ or - would be better.

[peterspackman]

I should definitely add it or update the documentation then! In general I would always write - for command line arguments, but for the toml file that felt a bit odd for some reason to me hence why I added so many extra _ options (with generally longer names)

[MilitaoLucas]

Fair enough. But I think some parity would be good. Maybe not one letter abbreviations, but the others? I feel like it would be nice if it was possible to easily turn a command into a TOML file just by looking at it. This mostly works already. I can maybe help with the documentation tomorrow. It is already almost 1 AM here...