You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Looks like the REST interface drops optimization_parameters data on the floor. The dict
TODO:
I think moe.views.schemas.OptimizationInfo class needs to have optimization_parameters field added as a schemanode--just a dumb 'dict' node that doesn't do any validation.
Also, the comments for OptimizationInfo are out of date.
Since the the webserver is mocked (self.testapp), we should have access to all the data we pass through the POST. So maybe btwn these lines
We should test that the information deserialized from the post request is what we expect. This might not be the right place for this test but something along these lines should be checked.
Related Issue: GH-97
Maybe the deserializers should complain about any unexpected/non-understood data?
NOTE:
When you make this change, the gp_next_points test will start using very different gradient descent parameters (TEST_ instead of DEFAULT_). This may cause the tests to fail or do nothing useful--TEST_ defaults may need to be adjusted.
The text was updated successfully, but these errors were encountered:
BLOCKING: GH-97
Looks like the REST interface drops
optimization_parameters
data on the floor. The dictTODO:
I think
moe.views.schemas.OptimizationInfo
class needs to haveoptimization_parameters
field added as a schemanode--just a dumb 'dict' node that doesn't do any validation.Also, the comments for
OptimizationInfo
are out of date.Since the the webserver is mocked (
self.testapp
), we should have access to all the data we pass through the POST. So maybe btwn these linesWe should test that the information deserialized from the post request is what we expect. This might not be the right place for this test but something along these lines should be checked.
Related Issue: GH-97
NOTE:
When you make this change, the gp_next_points test will start using very different gradient descent parameters (TEST_ instead of DEFAULT_). This may cause the tests to fail or do nothing useful--TEST_ defaults may need to be adjusted.
The text was updated successfully, but these errors were encountered: