Using XBRL platform Arelle as AWS lambda function

Question

I am trying to call Arelle in an AWS lambda function. I fail because Arelle wants to read/write the input XBRL file and the taxonomy to the read-only file sytem.

I want to request the facts. The following is an example for my request, if I would use the Arelle REST webserver. http://localhost:8080/rest/xbrl/view?file=C:/path/to/xbrl/file/test.xbrl&media=xml&packages=http:/url/to/taxonomy/some-taxonomy&view=facts&factListCols=Label,unitRef,Dec,Value,EntityScheme,EntityIdentifier,Period,Dimensions

To call it from AWS lambda function, I use the Arelle Python API, as described here: https://arelle.org/arelle/documentation/api/ The input file should be provided as a base64 encoded string. The result should be returned as base64 as well.

I already set the 'disablePersistentConfig' option.

However, I could not prevent Arelle from reading the input file and wrinting the taxonomy package to file system. For the result file, I used the FileNamedStringIO, which worked well.

factsFile = FileNamedStringIO(resultMedia)
ViewFileFactList.viewFacts(modelXbrl, factsFile, labelrole=None, lang=labelLang, cols=factListCols)
result = factsFile.getvalue()

I tried to use FileNamedStringIO for the input data was well, but this fails.

Code:

modelXbrl = self.modelManager.load(FileNamedStringIO(_f), "views loading", entrypoint=_entrypoint)

Error message:

 File "...\...\Python311\site-packages\arelle\FileSource.py", line 146, in __init__
    self.url = str(url)  # allow either string or FileNamedStringIO
               ^^^^^^^^
TypeError: __str__ returned non-string (type _io.StringIO)

How can I persuade Arelle to not expect a real file from the file system?

And is there a possibility to deactivate the Webcache that wants to write the taxonomy to file system?