Features to improve

imitrichev · December 22, 2023, 6:27pm

Dear developers,
sorry, but I cant’ work with 1.98.

Atom labelling in Avogadro 1.98.
Quantum chemists (working e.g. in Orca) do need atom numbering from zero. It really saves, and this is the feature why I chose Avogadro.
It was in 1.96-1.97 (Label → Index), but some non-programmer (I guess…) decided to make atom numbering (Labels → Numbers) from unity in 1.98.
Please, return atom numbering from zero = Index. If both regimes are needed, save both of them (numbering from zero was called Unique ID in Avo1).
1.98 simply does not Open anything on Windows on my computer
Open - File - choose C:\Orca\1.hess.v006.xyz
“the file name contains invalid characters”

Dr. Ivan Mitrichev

ghutchis · December 23, 2023, 4:07pm

We released version 1.98.1, which includes this bug fix.

As far as numbering atoms from zero or one … certainly it can be an option. The vast majority of users prefer numbering atoms from 1. Most programs also output atoms numbering from one, even if they’re handled internally in a zero-based array.

In the meantime, you certainly can change the code yourself as a developer:

github.com

OpenChemistry/avogadrolibs/blob/b45dcfd9061619dd4e67be4510abe45d03fd9d94/avogadro/qtplugins/label/label.cpp#L319


      
          if (!m_layerManager.atomEnabled(layer, i)) {
            continue;
          }
          
          auto& interface = m_layerManager.getSetting<LayerLabel>(layer);
          std::string text = "";
          if (interface.atomOptions & LayerLabel::LabelOptions::Custom) {
            text += (text == "" ? "" : " / ") + atom.label();
          }
          if (interface.atomOptions & LayerLabel::LabelOptions::Index) {
            text += (text == "" ? "" : " / ") + std::to_string(atom.index() + 1);
          }
          if (interface.atomOptions & LayerLabel::LabelOptions::Name) {
            text +=
              (text == "" ? "" : " / ") + std::string(Elements::symbol(atomicNumber));
          }
          if (interface.atomOptions & LayerLabel::LabelOptions::Ordinal) {
            text += (text == "" ? "" : " / ") +
                    std::string(Elements::symbol(atomicNumber) +
                                std::to_string(atomCount[atomicNumber]));
          }

imitrichev · December 23, 2023, 6:36pm

Thank you very much!
Nice to hear that, I’ll propse the changes.

Thomas · December 24, 2023, 3:57pm

@imitrichev Would it suit your needs if there were a variable in Avogadro’s GUI (similar to e.g., select which force field to be used to optimize the geometry) if atom indexing starts by 1 (as in Lua, Julia, Fortran, etc.), or 0 (as the family around C/C++/C#, or Python)? I can imagine the very first start of Avogadro displays a widget asking the user to opt-in for either choice (similar to the first time’s setup in gabedit) – in any subsequent use, Avogadro would stick to this selection until one intentionally selects the other level.

ghutchis · December 24, 2023, 4:00pm

@imitrichev helpfully already started a pull request: add UniqueID label type by imitrichev · Pull Request #1548 · OpenChemistry/avogadrolibs · GitHub

matterhorn103 · December 24, 2023, 9:03pm

Would it make sense to have indexing from 0 or 1 changeable from the options in the “View Configuration” pane for the “Labels” view option?

ghutchis · December 24, 2023, 9:05pm

That’s essentially the pull request - it adds a separate option for that.

matterhorn103 · December 24, 2023, 9:15pm

Personally I think sensible defaults are better than a “first run” dialog. The way Microsoft Edge currently requires several user choices before Windows 11 can be used at all is extremely irritating in a Domain environment, so it should be possible to use Avogadro immediately after starting in my opinion.

As long as it is easy to find and switch the setting I think assuming a default is fine.

And I agree with @ghutchis that for non-programmer chemists indexing starting at 1 is overwhelmingly the expected behaviour and should be the default.

But having the option for 0 is definitely important, makes working with ORCA much easier. Relatedly, I think the indexing for MOs and frequencies needs also to either be user-defineable as 0 or 1, or they should be read automatically from e.g. ORCA files as currently the displayed MO/mode number and that in the ORCA output files is mismatched.