tylercecil / chomtree Goto Github PK

Need to actually set up the npm package / put some structure in. I think once I see the project set up, with tests, I'll feel like I understand what's going on.

See example [T [NP [D This] [N example]] [T [Aux_1 /][Aux* this is some data yo!]] [VP* has a really big node.]]

I'm picturing just a table, with a number of stress tests. That way, if any changes are made to rendering, we can just open up playground.html and spot check a number of cases. In the future, once we are stable, that table can include the "considered correct" trees as well, so the changes can be seen. If that happens, though, we'll need to be sure to update them as rendering changes (can make a hook for that).

[X* data] is currently a tree with one node, which contains leaf data. However, "data" will be rendered as its own node, with a "triangle link" between X and data. Therefor, the real data model should build data as a child node with leaf data.

When rendering the nodeType, there's no way of knowing the order that sub and sup should be rendered. I think, honestly, this is an issue with the way WORD tokens are tokenized, and I've been thinking of what I'd prefer.

sub and sup tokens should be made first class, and WORDS should be chains of normal, sub, and sup tokens. What's more, while WORD is being turned into a more complicated structure, it would make a lot of sense to allow for quote-delineated words. This would allow for multi-word nodeTypes, as well as a mechanism to "escape" special characters (by treating everything in the quotes as a literal.

See #16 to see the current issue.

I don't want to get into this, though, until after rendering is totally finished.