Éric Dupuis

Results 95 comments of Éric Dupuis

No worries! Space dimensions are not an issue. I will look into it, but if you're using `group = "time.dayofyear"`, you will need a regular calendar yes. I think this...

Can you share more information on the error? If you could share 3 .nc files for `dref, dhist,dsim` with a single lon/lat point it would be helpful so I can...

The python script in your folder is for the old way of working things with `NpdfTransform`? Do you have the script you used for this above?

Thanks! The problem is that you have different time coordinates: ``` dref.time >>> '1986-01-01T00:00:00.000000000', '1986-01-02T00:00:00.000000000', '1986-01-03T00:00:00.000000000', ..., dhist.time >>> '1986-01-01T12:00:00.000000000', '1986-01-02T12:00:00.000000000', '1986-01-03T12:00:00.000000000', ..., ``` for instance if you try `sdba.QuantileDeltaMapping.train(ref=dref,...

For xclim devs: We could have a similar function to `_harmonize_units` which simply checks if times match? This would look more like `_check_inputs`, but it could not really fit in...

#1995 and #2013 close this.

> * I am not sure about the APP for lognorm It was exactly the MLE with saddle point estimators, good job

* I can't seem to resolve the coveralls issue * still no groundwater testdata * I can't be the reviewer on this anymore

Sorry, I didn't see you had reviewed the PR @huard . I will address your comments now.

> I just want to make sure this PR is not impeding the use of PWM for genextreme and gamma. On the contrary, PWM is allowed with any distribution found...