Webonary now imports xhtml files by default in the background. This allows you to close the browser after the file is uploaded to the server. You can go back to the import page to view how far the import has progressed:
Importing in the background is faster than displaying a progress counter and is therefore less likely to time out. If it does time out however, you can now resume the import without having to start from the beginning.
Also, you will now receive an email once the import has completed.
Data can be exported from Flex using the Configured XHTML export and imported into the Webonary site. These steps clarify the process for doing this manual to produce sites like those that are currently available.
The xhtml can be post processed. This is often done to replace the homograph numbers with Unicode sub-scripted numbers since the style syntax output from flex does not produce a sub-scripted number (but rather a full size number) -> this is no longer necessary, the importer does it for you!. Another reason for doing this would be to combine the lexical relations of the same type under a single heading. In any case, there are XSLT transforms available including CombineRelations.xsl that performs these operations. Follow the instructions for applying an XSLT.
In addition, rather than uploading with the style sheet from Flex, you may want to use the style sheets that have been custom modified for one of the other sites. One style sheet makes the abbreviations into small caps (pacoh).
I use Oxygen which has a nominal fee. I like this option because it comes with a development environment which allows me to step through the execution and set break points to jump to a particular line in the input or a particular template in the style sheet.
A free command line options is Saxon. (See: https://saxon.sourceforge.net https://saxonica.com))
Microsoft provides a command line tool. Remember to use the -xe switch.
Microsoft C# (and other languages) include XSLT processing. The express version of these languages can be downloaded for free.
Python also will process XSLT. A script can embed the transformation if desired. It relies on installing lxml (which is a Python add on module for processing XML files.)
You can learn using the online tutorial.