Many nodes contain internal state that the user might like to either chart or export. For example, analysis nodes will contain histograms and configuration nodes will have the atom positions. For a better user experience, we would like for these to be handled directly by the nodes themselves, rather than having the node output the data and then forcing the user to pipe it into another node to enable the plotting.
All of this will require a framework for allowing nodes to export chart data. From some stand-up discussions, it is clear that we want each node to also be able to host its won charts in an "expanded" view of the entire graph, as well as seeing the node in their original, condensed forms.
Given some of the parity with the Parameters setup, it might be worth making the charts a form of hidden output. Alternately, it could just be that each node exports an object and a QML file that handles its own charting. This part of the design needs to be figured out.
Many nodes contain internal state that the user might like to either chart or export. For example, analysis nodes will contain histograms and configuration nodes will have the atom positions. For a better user experience, we would like for these to be handled directly by the nodes themselves, rather than having the node output the data and then forcing the user to pipe it into another node to enable the plotting.
All of this will require a framework for allowing nodes to export chart data. From some stand-up discussions, it is clear that we want each node to also be able to host its won charts in an "expanded" view of the entire graph, as well as seeing the node in their original, condensed forms.
Given some of the parity with the Parameters setup, it might be worth making the charts a form of hidden output. Alternately, it could just be that each node exports an object and a QML file that handles its own charting. This part of the design needs to be figured out.