Add tabIngest field and 500 error workaround

[JR-1991]

According to IQSS/dataverse#11265 (comment) and #18, direct uploads of multiple tabular files trigger an OptimisticLock exception, resulting in a 500 status response when trying to register the uploaded files. This issue can be avoided by disabling tabular file ingestion.

This PR enhances the file.py model by introducing the tabIngest field. This field is set to True by default but can be disabled if the error occurs. The README.md has been updated to include information about this new field, as well as a troubleshooting section to guide users on how to resolve the issue.

Update: This PR includes a fix to prevent empty requests to replaceFiles or addFiles as pointed out by @landreev in IQSS/dataverse#11265 (comment)

• Closes 500 error during "registering files" step #18
• Closes Handle empty registration cases #25

Update 2: This PR includes a fix to handle ZIP files correctly in two ways:

• Metadata updates do not apply to ZIP files, since these will disappear on the server side once unzipped.
• Double-zipped files are handled as well.
• If a ZIP contains too many files and the server throws an error, this one will be propagated to the user.
• Added test cases to validate the behaviour

[JR-1991]

The PR has now been updated to accommodate intermediate locks by utilizing tenacity as a general retry solution. The library is applied to both the upload and registration steps to prevent timeouts and lock problems. It has been used previously, but the retry strategy was very tight and has been loosened now.

Configuring the retry behavior

Furthermore, this PR adds a documentation and a dedicated function to control the retry behavior. Since instances can and will differ, users might run into timeout issues or others, and this provides a way to accommodate these cases. However, there is no one-fits-all solution, and values may need to be determined manually. In most cases, the defaults should work, though.

Extensions of test cases

Finally, the PR adds multiple test cases for the native upload, which uploads a set of large tabular files to check whether ingest errors are occurring. For local testing, the size of the tabular files can be configured via TEST_ROWS. On GitHub runners, this is defaulted to 100_000 for mild stress tests.