make replace parameter work in edit_dataset_metadata

[pallinger]

• do not delete params in put_request
• fix replace documentation and parametrization in edit_dataset_metadata

The main problem was that params was overwritten by {} in put_request at the beginning, effectively deleting any parameters passed. This caused the replace parameter in edit_dataset_metadata to not work.
There was also some confusion with is_replace in the documentation of edit_dataset_metadata, and also passing boolean instead of string "true" (although that may work, this is more in-line with https://guides.dataverse.org/en/latest/api/native-api.html#edit-dataset-metadata)

• closes edit_dataset_metadata replace bug #156

[pallinger]

Could someone review this?

[skasberger]

Update: I left AUSSDA, so my funding for pyDataverse development has stopped.

I want to get some basic funding to implement the most urgent updates (PRs, Bug fixes, maintenance work). If you can support this, please reach out to me. (www.stefankasberger.at). If you have feature requests, the same.

Another option would be, that someone else helps with the development and / or maintenance. For this, also get in touch with me (or comment here).

[pdurbin]

@pallinger hi! Is this a fix for the following issue?

• edit_dataset_metadata replace bug #156

[pallinger]

@pdurbin Yes, it seems to be.

[pdurbin]

Ok, thanks, I added "closes" to the description.

[sonarcloud]

Kudos, SonarCloud Quality Gate passed!   

0 Bugs
0 Vulnerabilities
0 Security Hotspots
0 Code Smells

No Coverage information
No Duplication information

[shoeffner]

Thanks a lot for the PR!
The documentation strings are a good catch and should definitely be adjusted, however, I disagree with some of the code changes and added a few comments why.

[shoeffner]

This should stay None.
It is almost always better to use None instead of {} or [], the common pattern being:

def fun(arg=None):
    if arg is None:
        arg = {}

The reason is, roughly speaking, that python will process the def-line at its first pass through the code, initializing all arguments given there once and storing the references.
If the underlying argument would then, e.g., be appended to, this would modify the same object for each call. Consider:

def fun(arg=[]):
    arg.append(1)
    print(arg)
fun()
fun()
fun()

This would print:

[1]
[1, 1]
[1, 1, 1]

However, if the expected output is [1], [1], [1], the code should be:

def fun(arg=None):
    if arg is None:
        arg = []
    arg.append(1)
    print(arg)
fun()
fun()
fun()

While it is not so important here, I would still recommend sticking with the convention unless there's good reason not to.

For more information, see https://docs.python.org/3/faq/programming.html#why-are-default-values-shared-between-objects.

[shoeffner]

This can stay as it is. Maybe with requests this was necessary, but with httpx I tried this:

>>> import httpx
>>> httpx.get('https://example.org', params={'bool': True, 'string': 'true'}).url
URL('https://example.org?bool=true&string=true')

So both seem to result in the same outcome. One could even consider using params = {"replace": replace} instead of the ternary. But I am not sure if the API supports it, so I would keep it as-is.