upsert dataset methods and new backend

[dylanmcreynolds]

@abigailalexander recently added upsert functionality into pyscicat. I'm trying to add support for the new backend into pyscicat. One issue that upsert has not been implemented. I'm curious how much we should push to get it in, or if the use cases could be accomplished with a patch?

I've submitted an issue to track this in the new backend project SciCatProject/scicat-backend-next#93

[dylanmcreynolds]

At the developer meeting, we discussed whether the new backend would support upsert, and the answer was no, it will not. Therefore, we decided to remove it from the pyscicat code. Some code at RFI will be have to be rewritten to accommodate this change.

[abigailalexander]

I see, thank you for the update on the developer meeting and apologies I couldn't make it. I had a quick look at the new scicat backend and it looks like it still makes use of an update dataset method. For us, an update/patch dataset method would most likely cover the use cases that upsert is used for and would hopefully be compatible with both current and next scicat backends.

[toqduj]

is there a convenience method perhaps to check if a dataset exists (creating a new one if not and updating if yes)?

[dylanmcreynolds]

Right now there is no method in the server that just checks for the existence. If you know the PID of the dataset, you can use a GET on /datasets/{id}. That will return the dataset with a 200 or a 404 if not there.