Skip to content

Hydro data updates - #62

Merged
rabernat merged 9 commits into
pangeo-data:masterfrom
jhamman:hydro_data_updates
Feb 7, 2020
Merged

Hydro data updates#62
rabernat merged 9 commits into
pangeo-data:masterfrom
jhamman:hydro_data_updates

Conversation

@jhamman

@jhamman jhamman commented Nov 28, 2019

Copy link
Copy Markdown
Member

Adding two new datasets to the Pangeo catalog. Do not merge until I move these to their new bucket.

args:
urlpath: "gs://pangeo-data/soilgrids/{{ variable }}_250m.tif"
chunks: {'y': 5120, 'x': 5120}
storage_options: {'anon': True} # TODO: remove, see GH#61

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@rabernat and @davidbrochart, wondering if you have suggestions on how to do this. This catalog loads fine in ocean.pangeo.io and fails on travis with the following error:

E   rasterio._err.CPLE_AppDefinedError: Fetching OAuth2 access code from auth code failed.

The bucket is currently public so I'm not sure why I need to do any auth here. Thoughts?

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We have a similar issue in #59.

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actually this is a different issue. This one is more similar to intake/intake-xarray#31. I made a PR here: https://github.com/jhamman/pangeo-datastore/pull/1

@jhamman
jhamman marked this pull request as ready for review December 2, 2019 17:58
@jhamman

jhamman commented Feb 7, 2020

Copy link
Copy Markdown
Member Author

Hi @charlesbluca - any chance you could look into this? I'm curious how we will go about testing things like this now that http://catalog.pangeo.io/ is up. I'm pretty sure the failures here are just related to the CI system we're using in this repo but it would be good to get your thoughts on them.

@charlesbluca

Copy link
Copy Markdown
Member

It looks like the failures are rooted to a bad response from http://metadata.google.internal/computeMetadata/v1/ - are the requests made to this URL being done by our own service account? The error page doesn't seem to mention anything about lack of permissions at this URL - just a 404 error. Is this form of refreshing service account info deprecated?

@rabernat

rabernat commented Feb 7, 2020

Copy link
Copy Markdown
Member

The CI is broken here because we don’t have credentials in Travis to open requester pays bukets.

Our options are:

  • fix this (by providing credentials somehow)
  • disable the test that actually opens the dataset with xarray and just focus on testing the intake catalog itself

I favor temporarily disabling the xarray opening, also because it’s slow. It doesn’t scale to open every dataset in the catalog every time CI is run. An even better option would be to look at the diff and only open the new datasets in each PR. That would require some git ninja skills that I don’t have.

@jhamman

jhamman commented Feb 7, 2020

Copy link
Copy Markdown
Member Author

FWIW: I am in favor of disabling the dataset open here and just testing / linting the catalog.

@charlesbluca

Copy link
Copy Markdown
Member

If we want to leverage a solution that allows for additional credentials, we could consider a Cloud Function that tests updated catalogs (or just all catalogs on a trigger). This is currently how I've been updating the build badges for the catalog.

@rabernat rabernat mentioned this pull request Feb 7, 2020
@rabernat

rabernat commented Feb 7, 2020

Copy link
Copy Markdown
Member

@jhamman could you just rebase this on master? The tests should go green.

@rabernat
rabernat merged commit e8396f5 into pangeo-data:master Feb 7, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants