Gerrit Holl

See also [question on Stack Overflow](http://stackoverflow.com/q/31839047/974555).

Related (closed) issue: https://github.com/pytroll/pyorbital/issues/28

It's not enough to manage `warnings.catch_warnings`, one would also need to manage `np.errstate`, since either can be used to silence a `RuntimeWarning`. Perhaps not in the example at the OP,...

From a user point of view, ideally I would like something that mimics the numpy `errstate` context manager closely, so that a user could write: ``` with dask.errstate(invalid="ignore"): da =...

In principle, interested yes, but I can't allocate time to this right now (also because I would need to learn a lot more about dask first).

Unless there is a public place with FCI data served over S3, I can't test the remote reading functionality. However, for local processing, there is a severe regression in performance....

> @djhoese @mraspaud just looking through some PR's to clean up some open stuff. Is there something else than merging main again which needs to be done here or can...

What is the current/previous way to get the correct reflectance where this is not built in to the reader's calibration routine? There does not appear to be a modifier `sun_earth_distance_corrected`...

> the vis_08 channel is too bright, so that a factor of 0.8 should be applied to get correct radiances/reflectances Although I don't think tuning based on buggy test data...

This fails only for the experimental run now, which, as I understand it, is allowed. So the PR is ready as far as I'm concerned.