Whitespace in Lowfat

[jonathanrobie]

Whitespace in Lowfat has changed in ways that make it harder to read in some editors.

Could we please restore whitespace to the previous form? This is probably as simple as running the transform originally used to create it, making sure we keep all the attributes.

[ryderwishart]

Can you describe the issue a little more specifically? Do you mean you would like every attribute on its own line, regardless of line length?

[klosoter]

This probably happened because of me. The Python library I use (lxml) puts all attributes on the same line as the element. As far as I know, there are no easy ways (with Python) to change the formatting of an XML file. So, it will probably happen again when I make new commits.

Any ideas on how to avoid this problem?

[themikejr]

@klosoter lxml has a pretty print option that will improve the formatting (see this gist). There are also other python libraries that may offer more customization for xml formatting (e.g. xmlformatter). Ideally though, the lowfat data pipeline for hebrew and greek would have some final formatting step that is automated and consistent across repos.

[jonathanrobie]

@klosoter Up to now, Lowfat has always been generated from Nodes using an XQuery that does the right thing.

I don't think it's good to update the two in parallel, I think one should be generated from the other.

[klosoter]

Let's generate lowfat from the other trees, then. Indenting attributes on a new line is not something pretty print or xmlformatter can do.

[ryderwishart]

I can regenerate the lowfat trees right now on your branch. However, did we resolve the nodes that were getting dropped?

[klosoter]

I did

[klosoter]

Also, I made some changes in lowfat (there were duplicate xml:id) and some Node/m that should have been wg/w.

[klosoter]

This is about macula-hebrew right? @ryderwishart

[ryderwishart]

Yes that's right @klosoter. Didn't notice it was the wrong repo. 🤦‍♂️