Concept map issues

[cashaffer]

1) It should be possible to back up to the last state before the current click. 2) We need more context for "leaf" nodes. Currently, when you click on a "leaf" node, you get a map with just a single node on it (and nowhere to go). 3) There are actually two types of nodes in the concept map graph. There are the terms, and there are the labels for the edge. For example, from "graph" there is a link to "contains" (which is an edge label, not a term), with links to "vertices" and "edges" (which are terms). Nothing should happen when one clicks on an "edge label" node. Right now it thinks this is a term that is not in the glossary. 4) We need to make a distinction between the display text in a term node and the actual term as it appears in the glossary. For example, I want "edges" to appear as a label in the concept map, but the actual term is "edge". This will require some sort of change to the :to-term: directive. 5) The graph layout is still problematic. For example, see the graph that results from clicking on "graph". There are a lot of overlapping nodes. I also notice that there is some randomness involved. If I repeatedly click on "graph" in the glossary, the resulting concept map layout changes each time.

[efouh]

https://github.com/OpenDSA/OpenDSA/commit/7bc5193876c0ae9b06ed48c4f2d073756ca9f66b

[cashaffer]

The changes that you made have stopped the concept map window from displaying. There is what appears to be an error message in the console, but it flies by too fast when I click the link and the page changes to the glossary.

[cashaffer]

Turns out the issue with no concept map displayed comes from changes in the update to how it handles data in localStorage, combined with old data in localStorage.

With the changes, the issue of "edge labels" is fixed, and the backup works in principle. But here is a new round of issues:

• Perhaps should better handle localStorage issue. I know this is somewhat legacy, but I can see this being an issue. If its just a line or two of code, might be worth checking if the value in localStorage is an item instead of a list. Or more likely, when opening a new concept map, should perhaps reinitialize the stack for backup? See points below.
• When I open a new concept map, I should be at some initial state. Specifically, I mean that there should not be the ability to back up. Ideally, the back button won't appear, or at worst it will be (obviously) grey'ed out. Right now, there appears to be history from prior uses of the map.
• Re-clicking on the current node should not add to the stack. (Right now, it does.)
• The "current" node of the graph should be highlighted.
• Need better handling of terms that do not have links. Right now, I get a popup that says "the term is not in the glossary". I then get a blank graph. But none of this is true. I should at least get the node with an empty definition box, or better the statement that there is no definition should be in the definition box, not in a popup window.
• When the book is compiled, if there is a link to a missing glossary term, it should complain.
• If I click on "graph" then on "minimal-cost spanning tree" and then on "Kruskal's algorithm", it claims that "Kruskal's algorithm" is not in the glossary. This is not true.

• If I click on "graph" then on "edges" it says that "the term edges is not in the glossary". This also happens if I click on the term "edges" in the glossary (instead of in the concept map). I think the issue here relates to the display anchor text vs. the actual term. "edge" is the glossary term, while "edges" is the display label. Note that I went back and recompiled the book with the :to term: directive for "graph" to read:

  :to-term: edges :label: contains

So it has the information to know (even in the concept map) that the glossary term should be "edge". This compiles correctly, and displays "edges" as the node text, but does not behave correctly on the click.

• The edge/arrows of the graph are nearly impossible to see which way the arrow head goes. The style needs to change to make the arrow head bigger. Possibly the edge should be stroke width 2 instead of 1 as well.

[cashaffer]

I am getting feedback that we need greater visual contrast between the "term" nodes and the "label" nodes. I think that it might work to remove the white background and edge shading around the "label" nodes. In this way, the will look more like labels and less like nodes.