Facilitate retrieval of CTP even if cloud shadow is not computed

[heptaflar]

This pull request would enable the retrieval of CTP on request, even if cloud shadow is not requested. This makes sense in cases where one needs CTP in subsequent processing steps but does not need cloud shadow or is going to compute cloud shadow by different means, which need CTP as input.

[dolaf]

looks ok for me

[TonioF]

In general, this seems to work fine (and it makes more sense than the existing implementation). Three comments:

1. The parameter outputCtp is not well named anymore. The import part is that it is computed, so it should be "computeCtp". The renaming probably would cause backwards incompatibility, though, so I don't insist on this.
2. I guess it would be okay to output the ctp band everytime it is computed, either because "outputCtp" or "computeCloudShadow" is true.
3. I'd love if in the label and the description you could use "cloud top pressure" instead of CTP.

[marpet]

In the @parameter annotation alias="computeCtp" could be added. Then outputCtp will still work, but computeCtp has precedence.

[marpet]

Is this applicable to other sensors besides OLCI? Just to harmonise things.

[heptaflar]

I agree with the first and the last point Tonio raised. Renaming the parameter to "computeCtp" while introducing an alias "outputCtp" ensures compatibility with previous versions. Thanks Marco for pointing this out.

I am not sure about Tonio's second point, however. For me, cloud shadow and CTP are separate concerns. I prefer the present behaviour, i.e., CTP is only put out, if requested explicitly. Anyway, Tonio's suggestion would work for me, too.

I modified the original pull request to address point 1 and 3 on Tonio's list.

[TonioF]

I am not sure about Tonio's second point, however. For me, cloud shadow and CTP are separate concerns. I prefer the present behaviour, i.e., CTP is only put out, if requested explicitly. Anyway, Tonio's suggestion would work for me, too.

Any more thoughts on this? Then we can have a majority vote. I'd be fine with accepting now.

[marpet]

I would vote for Ralfs solution and leaving CTP and cloud shadow independent.

[dolaf]

I would also vote for keeping them separate. Also, the user should still have the choice not to compute CTP if not needed. Every bit of performance is useful :-) Olaf Am 29.01.2021 um 11:45 schrieb Marco Peters:

…

I would vote for Ralfs solution and leaving CTP and cloud shadow independent. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#33 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABUGBR3IJSCUOLVB4P26QTS4KGVDANCNFSM4WXMDFVA>.

--

____________________________________________________________ Dr. Olaf Danne Brockmann Consult GmbH phone: +49 (0)40 696389 313 email: olaf.danne@brockmann-consult.de skype-id: olaf.danne.bc

____________________________________________________________ Brockmann Consult GmbH Chrysanderstr. 1 D-21029 Hamburg, Germany Amtsgericht Hamburg HRB 157689 Geschäftsführer Dr. Carsten Brockmann Web: www.brockmann-consult.de Twitter: @BrockmannCon

____________________________________________________________

[heptaflar]

Accepted by me. Who merges?

[marpet]

Probably Olaf, when he finds the time to do it

[marpet]

Not needed anymore. CTP calculation is not the purpose of IdePix. The output is only included for debugging purpose.