Problem with Avogadro Reading Input Generators

rossi · June 11, 2020, 10:21pm

This is my second time asking a question about this.

I downloaded the avogadro2 input generators.

This is where I put them:

/home/rossi/.local/share/OpenChemistry/Avogadro/generators

The only instructions provided are that they should be placed in a directory where avogadro2 can read them.

This is the contents of the directory:

LICENSE dalton.py* gaussian.py* inputgeneratortest.py* mopac.py* orca.py* psi4.py* qchem.py*
README.md gamessuk.py* generators.cmake molpro.py* nwchem.py* plugin.json pyscf.py* terachem.py*

There are no instructions on what to do with generators.cmake.
There are no instructions on what to do with plugin.json.
There are no instructions to know if the input generators have been read properly or read at all.

avogadro2 ----> Extensions —> Plugin Downloader provides text.
But, I don’t see anything about the input generators.

Don’t you think some more information could be provided to make the install process a little smoother.

The bottom line is that I think the input generator files are in the right directory, but avogadro2 doesn’t seem to recognize them.

But there are no directions to install them or what to look for.

ghutchis · June 12, 2020, 8:29pm

Anything that’s not a script will get ignored. On launch, it will go through the files and track which ones it can run. We decided not to restrict extensions because someone might want to write in Shell or Ruby or Julia or whatever.

The point of downloading is that you can update those scripts without needing a new release of Avogadro.

The versions you downloaded are likely slightly newer than in the last released version. I don’t remember offhand if there’s something new, but I think PySCF was added recently.

IIRC the GitHub repository also as “solvation” as an option in orca.py

If you want to know if they’ve loaded correctly, go in and change something, like the basis sets:

    userOptions['Theory'] = {}
    userOptions['Theory']['type'] = 'stringList'
    userOptions['Theory']['default'] = 7
    userOptions['Theory']['toolTip'] = 'Hamiltonian or DFT method to use'
    userOptions['Theory']['values'] = \
        ['HF', 'MP2', 'CCSD', 'BLYP', 'PBE', 'B3LYP', 'B97', 'wB97X' ]

    userOptions['Basis'] = {}
    userOptions['Basis']['type'] = 'stringList'
    userOptions['Basis']['default'] = 3
    userOptions['Basis']['toolTip'] = 'Gaussian basis set'
    userOptions['Basis']['values'] = \
        ['6-31G(d)', 'cc-pVDZ', 'aug-cc-pVTZ', 'def2-SVP', 'ma-def2-SVP',
        'def2-SVPD', 'def2-TZVP', 'def2-QZVP', 'pc-2', 'aug-pc-2']

They’re just Python scripts - add some items to the list, change the default (e.g., 3 = def2-SVP)… if you like, you can submit a pull request or patch to update the scripts.

If you have a different program that you’d like to support, you can copy a script.

rossi · June 13, 2020, 2:01pm

Hello,

Thank you for responding to my question, but I am really not sure if you answered it.

This is what I see when I start avogadro2 and the GUI starts to appear:

Sat Jun 13 09:36:25 [rossi@debian ~]$avogadro2
Locale: “C”
“/usr/share/qt5/translations”
Extension plugins dynamically found… 24
Checking for “commands” scripts in “/home/rossi/.local/share/OpenChemistry/Avogadro/commands”
Checking for “commands” scripts in “/usr/share/xfce4/OpenChemistry/Avogadro/commands”
Checking for “commands” scripts in “/usr/local/share/OpenChemistry/Avogadro/commands”
Checking for “commands” scripts in “/usr/share/OpenChemistry/Avogadro/commands”
Checking for “commands” scripts in “/usr/bin/…/lib/x86_64-linux-gnu/avogadro2/scripts/commands”
OBProcess::executeObabel: Running “obabel” “-L formats read”
OBProcess::executeObabel: Running “obabel” “-L formats write”
OBProcess::executeObabel: Running “obabel” “-L forcefields”
OBProcess::executeObabel: Running “obabel” “-V”
“obabel” found: “obabel: Open Babel 3.1.0 – May 31 2020 – 04:31:34”
“Navigator” added
“Editor” added
“Selection” added
“Manipulator” added
“BondCentric” added
“MeasureTool” added
“PlayerTool” added

Why doesn’t a message appear that the input generators are read? The messages say that Avogadro2 is looking for them, but has not said it found them.

Again, here is where the generators are:

Sat Jun 13 09:43:35 [rossi@debian generators]$pwd
/home/rossi/.local/share/OpenChemistry/Avogadro/generators

I thought it was stated “where Avodagro2 can find them”. Is this not a good directory.

Again, Extensions ===> Plugin Downloader shows no signs of being read in.

Also, this is an aside. The File —> Open and the Molecule readers seem to be inconsistently different. The file readers are different. Should they be more consistent?

The interface shown on the documentation is somewhat different than what I see in my version. I haven’t been able to perform reading an output molecule from a Gaussian calculation from the File command (It’s a Gaussian log file.) that Avogadro2 doesn’t ask me how to convert is using babel or suggests GAMESS or NWChem. Why is that? Why can’t Avogadro2 just read the Gaussian file like Chemcraft, like JSMOL?

Let me end this reponse by asking you another question? I am using Debian Linux, the latest and greatest version. Have you tested Avogadro2 on Debian based systems?

Gosh, Avogadro2 has become so non-intuitive. It’s disappointing.

Again, thanks for your help.

Kind regards,

Angelo

ghutchis · June 13, 2020, 5:36pm

Avo2 development is occurring in my not-so-abundant free time and as it indicates, it’s beta. I’ve been working hard on improving the interface - it’s still rough. Which is also why the documentation is for Avo 1.x, not 2.x and there are differences - the 2.x interface is still in flux and I have a long list of changes planned. (Much like the many differences and rough edges in 0.7, 0.8., 0.9 before 1.0)

Can you please separate your questions into separate threads? It’d be much easier to respond to them separately.

I’m not sure what you mean by the File => Open and molecule readers.
I can read Gaussian, GAMESS and other log files. If you’re having problem with a particular file, please post a question or bug report about that with a link to the file.
I don’t use Debian, but other users do.
I mentioned several times to you that the plugin downloader has a clear bug - I have not had time to fix it.

I’m sorry you feel that software that indicates strongly that it’s a work in progress is disappointing. Submit bug reports, indicate what you feel is non-intuitive, etc.

At the moment, I’ve been the main contributor to Avo2 for months. Since I’m also running a research group, writing papers, grants, etc. i do not currently have a lot of time to code.

As an open source project, help is always welcome

rossi · June 13, 2020, 6:45pm

Hello,

Thanks again for your response and explanation.

OK. I wasn’t aware that there was a bug with the input generators. I apologize for that.

I am not able to read output from a previously executed input Gaussian file

I have a folder with the optimized geometry of acrolein: acrolein_vib.com, acrolein_vib.log, and acrolein_vib.chk
Using avogadro2: File —> Open —> I click on the file acrolein_vib.log

The response from Avogadro2:
Multiple readers found that can read this format.
The choices are OpenBabel, NWChem, and GAMESS; (If I choose OpenBabel, I get an error.)

WHY? Why doesn’t Avogadro2 recognize the acrolein_vib.log and just read it?

Why isn’t one of the choices Gaussian?

OK. How do I read in a Gaussian log file?
I am at a loss of what to do.
3. What I was trying to say and did not say it well was this:
Open --> and then search for file …

Molecule --> and then search for file

have different formats for searching or scanning through subdirectories to the target file. They behave differently.
4. I guess I am whining. But I am frustrated.

I would like to visualize the IR and UV spectra of acrolein, and also visualize the orbitals. But I cannot seem to even read in the log file.

I don’t think I have the expertise to be a contributor to Avogadro2.

My expertise is in High Performance Computing (http://rossi.chemistry.uconn.edu/)

Thanks fo all your patience.

Regards,

A. R. Rossi

ghutchis · June 13, 2020, 8:30pm

Avogadro v1 didn’t handle reading files - it did that through Open Babel. Some readers are built into Avo2, but most are still handled through OB. if you can post that file, I can see what the problem is. (I use OB and Avo2 to handle Gaussian output all the time.)

I would like to visualize the IR and UV spectra of acrolein, and also visualize the orbitals. But I cannot seem to even read in the log file.

The most useful Gaussian format is the .fchk file - Avo2 will definitely read these internally and allow orbital display.

I’ll be honest that I don’t remember the state of spectra in Avo2 - my group’s work has been on other things, so that may need to get ported this summer.

My expertise is in High Performance Computing (http://rossi.chemistry.uconn.edu/)

Yes, I looked you up.

I desperately want to get Avo2 in a polished state by the fall - I want to use it more myself and for teaching. One thing I’ve managed to do at Pitt is hire some undergrads for documentation, tutorials, and occasionally someone with coding experience.

So far, we’ve been singularly unsuccessful in getting NSF or NIH funding for Avogadro development. My hope is to come up with some sort of ‘consortium’ where departments put in some $$ or hire their own students to help with various tasks.

In any case, send me the output file, and try out an fchk.