Launch Binder Not working

[Sulstice]

Hello,

Your Launch Binder feature is not working but it is working on google colab. Can we switch the demo to there?

[VHchavez]

Hello Sul, thank you for stopping by.
The project is in a big need of an overhaul. I will have some time in the next week for that.
There were some changes done in Psi4 that prevent the Binder to work... I will be fixing that as well.
I have never used Google Colab. You're more than welcome to help me set that up :) . I'd wait for all the other updates to complete though.

[Sulstice]

Hey @VHchavez Sure I actually figured it out for google colab. I'll have a solid demo for you soon, give me a couple of days to wrap it up and I also used your software into mine, GlobalChem. Hope that is okay, I am extending the functionality of other stuff and moly is great for its visualization.

[Sulstice]

Here's the demo I made using some of your software, it has psi4 installation instructions in the first cell. Just hit the button play or play all.

[VHchavez]

Looks great thanks.
Now that I graduated, I have some time to work on this 🤣. Let me do the updates that I had in mind and then we can add your work to the main repo.
Thanks again!

[Sulstice]

Yeah! haha.

It is actually part of a bigger project of mine if you want to somehow merge it in? It's going to be a paper soon, might be cool to rope it all in and have you be part of it. I read your code and it's really useful.

https://github.com/Sulstice/global-chem

What do you think?

[VHchavez]

Thanks for your kind words.
I think it is a good practice to keep projects as separate entities. Each of them with a single thing to do. So that if anyone else is interested in using it does not have to first decouple it from a larger project.
But I am more than happy to optimize and enhance moly in any way that GlobalChem requires.

[Sulstice]

Works for me :)

I actually recently added this function and got it from forte to get cube files of occupied and virtual orbitals from psi4. What are your thoughts on it. Otherwise it's kind of a mess you get out for larger molecules:

def psi4_cubeprop(
            self,
            path,
            orbitals = [],
            number_occupied = 0,
            number_virtual = 0,
            density=False,
    ):

        """
        Run a psi4 cubeprop computation to generate cube files from a given Wavefunction object
        By default this function plots from the HOMO -2 to the LUMO + 2
        Arguments:
            path (String): file path to the output directory of the cube files
            orbitals (List): Any specific orbitals you want included,
            number_occupied (Int): Number of Molecule Orbitals Occupied by the Electron
            number_virtual (Int): Number of Virtual Orbitals Occupied by the Electron
            density (Bool): Density of the cubeprop
        """

        import os.path

        cubeprop_tasks = []

        if isinstance(orbitals, str):
            if orbitals == 'frontier_orbitals':
                cubeprop_tasks.append('FRONTIER_ORBITALS')
        else:
            cubeprop_tasks.append('ORBITALS')

            if number_occupied + number_virtual > 0:
                alpha_electrons = self.wave_function.nalpha()
                molecule = self.wave_function.nmo()
                min_orb = max(1, alpha_electrons + 1 - number_occupied)
                max_orb = min(molecule, alpha_electrons + number_virtual)
                orbitals = [k for k in range(min_orb, max_orb + 1)]

            print(f'Preparing cube files for orbitals: {", ".join([str(orbital) for orbital in orbitals])}')

        if density:
            cubeprop_tasks.append('DENSITY')

        if not os.path.exists(path):
            os.makedirs(path)

        psi4.set_options(
            {
                'CUBEPROP_TASKS': cubeprop_tasks,
                'CUBEPROP_ORBITALS': orbitals,
                'CUBEPROP_FILEPATH': path
            }
        )

        psi4.cubeprop(self.wave_function)

Could be worthwhile for additional mods?

[VHchavez]

I mean that's just how you get cube files using psi4 :)
I have tried moving sections like that to Python, but they've just been very inefficient. Cube files can get intractable very quickly, but that's just the nature of discretizing a massive chunk of space.