Bug in the basis set parser

[q-posev]

The regex that is currently used to parse the basis set files (const prim_pat = r"([+-]?\d*?\.\d+[D+-]{0,2}\d\d)\s+?([+-]?\d*?\.\d+[D+-]{0,2}+\d\d)" in BasisParser.jl) is not applicable to all available basis sets.
For example, the 6-311+g file contains floating point values that are not in scientific notation (e.g. 1.0000 instead of 1.0D+00). So the parser crashes because it cannot parse these values. Perhaps this can be handled via the try/catch mechanism and using an alternative regex.
Let me know if you need help fixing this. I am new to Julia but can work on the PR in my free time.

[gustavojra]

Hey @q-posev, thank you for pointing that out.

The original regex could handle numbers that are not in scientific notation. The quantifier {0,2} before [D+-] indicates that this bit was optional.

The problem was this particular piece after the period: \d+[D+-]{0,2}\d\d). Written like this, the regex expects at least one decimal after the period followed by the optional scientific notation marker (D+/D-) and then two more obligatory decimal digits... Consequently, numbers not in scientific notation would not be parsed unless they had at least 3 digits after the period.

This broke down as soon as I tried fluorine using the basis set you suggested. Moreover, it also breaks dramatically for potassium because its first entry:

K     0
S    6   1.00
 182594.                     0.000227747

No decimal digits at all...

I updated the regex to be more flexible, in case you're interested here is what I came up with

const prim_pat = r"([+-]?\d*?\.\d*(?:D[+-]\d+)?)\s+?([+-]?\d*?\.\d*(?:D[+-]\d+)?)"

Note that \.\d+ not became \.\d*. So I dropped the obligatory digits after the decimal mark. Also, I put the scientific notation bit inside a non-capturing group (?:X) where X = D[+-]\d+. The ? just after the group indicates that this part is optional. However, if D+ or D- is matched, there has to be at least one digit after it... Because a number like 1.0D+ does not make any sense to me.

Anyways, you can check out the PR right here #15 . I just wanted to give you some feedback.

[q-posev]

Hi @gustavojra, thank you for the prompt fix and for the explanation! I will have a look at the PR.
Indeed, I forgot to mention that it failed when I tried to run molecules containing F and Cl atoms. This fix is also important since many basis sets from BasisSetExchange contain values in the decimal notation. I also noticed that the 6-311g basis set is missing in the lib folder. Was it intentional?

[gustavojra]

Not, it is not intentional, back when I created the lib folder I just downloaded all basis set in the BSE. Maybe this one wasn't downloaded for some reason? Either way, for standard basis set files like that we can always create a simple PR including new files.

[q-posev]

OK, thanks @gustavojra ! I can prepare a PR.