Minor updates

[tasdomas]

Import scikit-learn directly.
Use updated cml cli.

[tasdomas]

@yathomasi PTAL - doing some house keeping on iterative's example repos. Though it looks like there's a problem with the dvc data sources for this one.

[yathomasi]

Though it looks like there's a problem with the dvc data sources for this one.

Yeah, I also got the following error on dvc pull. @tasdomas I am not the expert here 😃.

gcsfs.retry.HttpError: Anonymous caller does not have storage.objects.list access to the Google Cloud Storage bucket., 401
ERROR: unexpected error - Anonymous caller does not have storage.objects.list access to the Google Cloud Storage bucket., 401  

cc: @iterative/dvc

[daavoo]

Not sure what this line is supposed to do as the downloaded folder is not used by the pipeline in dvc.yaml?

[daavoo]

If you want to run dvc pull you need to provide credentials for the DVC remote defined in https://github.com/iterative/stale-model-example/blob/main/.dvc/config .

I have no idea who has access to those buckets

[daavoo]

I don't believe this workflow has ever run successfully as it is impossible for DVC to access the data given the current and previous setup

[pmrowla]

gcsfs.retry.HttpError: Anonymous caller does not have storage.objects.list access to the Google Cloud Storage bucket., 401
ERROR: unexpected error - Anonymous caller does not have storage.objects.list access to the Google Cloud Storage bucket., 401  

There's no credentials configured for the default gcs remote.

stale-model-example/.dvc/config

Lines 5 to 6 in 4597725

	['remote "gcpbikes"']
	url = gs://updatedbikedata

can you double check that this bucket is actually publically accessible and that it's configured to allow public (anonymous) users to have the required permissions?

[casperdcl]

Is this required because of dvc get?

[pmrowla]

It's required for dvc exp run

[casperdcl]

Is it needed for dvc repro too? Can't find any mention of this requirement in https://dvc.org/doc. Asking because we may need to update/fix https://github.com/iterative/cml_dvc_case/blob/master/.github/workflows/cml.yaml

[pmrowla]

For dvc repro normally the answer is no, but if the user's pipeline is doing things with the local git repo (like pipeline stages that run dvc get) then it may be needed

[casperdcl]

@tasdomas closed by accident?

[tasdomas]

@casperdcl I think this repo needs to be archived altogether.

[casperdcl]

/CC @jendefig ^

[jendefig]

/CC @jendefig ^

Thanks @casperdcl @tasdomas! @RCdeWit had put an admonition on the blog post. But it seems that this may be too broken to keep up, especially since the blog post starts from the repo. Looks like will need to pull this down and put it into the Content Creation Dashboard for an eventual redo.