Closed StevenClontz closed 9 months ago
I checked the first part of this yesterday, just the effect of changes to extract-pg.xsl
. They are as expected with regard to new XML containers.
With a switch from text output to XML output, the ampersand and less-than are being excaped, as they would need to be. And I guess that is fine since once consumed by lxml
they behave properly.
I'll check the whole pipeline here shortly. I'd also like to have input from @Alex-Jordan, or at least have him say he can't get to this one.
I'd like to implement Alex's idea that there's no need to write a file to disk: use this XSL in memory to create an lxml ElementTree, then read it directly rather than writing to disk. But that will come in a later PR.
I won't have time to directly test this out. I get the high level description of the change and it sounds good. All I would suggest is what Rob is already planning to do: compare before and after with the representations file on the WW sample chapter. There should be no changes, I think. Or if there are, they should be inconsequential.
Reorganized as two logical commits, but no change to the overall code.
Zero appreciable change to the representations file for the sample chapter.
This PR modifies
extract-pg.xsl
to output an XML file, which Python reads with lxml to obtain the data it requires to create webwork-representations.xml.I am able to successfully build webwork-representations.xml for both the minimal and sample-chapter webwork examples using this script, and they are identical to the current result of the head of the master branch.