👩🎓 JATS citation / affiliation fixes #783
Conversation
| // This could also be an output, etc. | ||
| (selectAll('[visibility=remove],[visibility=hide]', tree) as Block[]).forEach((node) => { | ||
| if (node.visibility === 'remove' || node.visibility === 'hide') { | ||
| (node as any).type = '__delete__'; | ||
| } | ||
| }); | ||
| // Blocks are converted to sections - avoid doing this for part blocks | ||
| doNotDelete.forEach((node) => { | ||
| node.type = 'block-part' as any; |
There was a problem hiding this comment.
This is a little sketchy - we want to include referenced parts in the transformations, but we do not want to include other hidden blocks.
So, first we find all the blocks that will be extracted as parts in blocksToKeep, then we make sure those are not marked for deletion here...
There was a problem hiding this comment.
However, we must give them a type different from 'block' since in another transform, all blocks are converted to "sections"
|
|
||
| export function restoreBlockPartTypeTransform(tree: GenericParent) { | ||
| selectAll('block-part', tree).forEach((node) => { | ||
| node.type = 'block'; |
There was a problem hiding this comment.
Finally, once transformations are done, we return them to their original 'block' type, so the extractParts function will behave correctly. Pretty gross.
There was a problem hiding this comment.
I am leaving this grossness in, however, since I think the real fix is bulid JATS in unist (there's already some utility for this in jats-xml) - then, we will not need to do these extra jats-specific transforms on the input MyST AST; we instead can apply simpler, more sensible transforms to the output JATS AST.
This PR addresses some of the tasks here - #347:
partsare left in the tree for final myst-to-jats mdast transforms, then plucked off immediately beforerender