pyRadiance integration

[cdeline]

LBNL has a new python wrapper around RADIANCE binaries. This may improve integration and eliminate issues users are having with properly installing RADIANCE and setting RAYPATH, etc.

It appears that the underlying C binaries are still being used, just with a more pythonic wrapper, so there should be no performance issues.

Repo: https://github.com/LBNL-ETA/pyradiance
Reference pages: https://lbnl-eta.github.io/pyradiance/reference/

[fredrbl]

Ah, @cdeline this would be amazing!
I am trying to set up a docker image with bifacial_radiance, and this is a pain because of the Radiance installation..! If I understand this PR correctly, this could make my Dockerfile (and the image) a whole lot simpler.

(If you already have some pointers or Dockerfiles that installs Radiance (I understand it would have to be the full Radiance (core+extra) in docker, I would greatly appreciate that as well :D )

[shirubana]

I agree :D

But, also, there is a docker file in bifacial_radiance that does the radiance install...

[fredrbl]

Thanks for response @shirubana !

From what I understood from this Dockerfile (I assume it is this: https://github.com/NREL/bifacial_radiance/blob/main/docker/Dockerfile), it does not take the entire Radiance code (i.e both core and extra). And this would be required to run certain commands in bifacial_radiance, such as gencumsky. Please correct me if Ive misunderstood. (however, I can maybe use the gendaymtx from pyradiance instead of gencumsky).
But I will give it a go and test! Thanks again for the tip.

[shirubana]

Kinda. That dockerfile uses an older version of Radiance; it adds the Jaloxa executables as part of the process that are now part of the newer Radiance releases. Otherwise it's the same. The docker file also adds the gencumsky function as part of the internal workings, so you don't need to switch to gencumsky yet. S. From: fredrbl ***@***.***> Sent: Wednesday, September 25, 2024 11:17 AM To: NREL/bifacial_radiance ***@***.***> Cc: Ovaitt, Silvana ***@***.***>; Mention ***@***.***> Subject: Re: [NREL/bifacial_radiance] pyRadiance integration (Issue #533) CAUTION: This email originated from outside of NREL. Do not click links or open attachments unless you recognize the sender and know the content is safe. Thanks for response @shirubana<https://github.com/shirubana> ! From what I understood from this Dockerfile (I assume it is this: https://github.com/NREL/bifacial_radiance/blob/main/docker/Dockerfile), it does not take the entire Radiance code (i.e both core and extra). And this would be required to run certain commands in bifacial_radiance, such as gencumsky. (however, I can maybe use the gendaymtx from pyradiance in stead). But I will give it a go and test! Thanks again for the tip. - Reply to this email directly, view it on GitHub<#533 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AFOEOZWGGXXAZT4OYKS7MNDZYLVZ5AVCNFSM6AAAAABNG4RFXGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNZUGY4TMOJYGY>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[cdeline]

I've started some tentative investigation into pyRadiance integration. At the moment there are still a few Radiance binary calls that are unsupported, but maybe there is a way around this.

• xform: this is now supported starting in pyRadiance 1.0.0, which has a requirement for python >=3.10. This is a mandatory feature, so this would require us to bump our minimum supported python from 3.8 to 3.10. Not a deal-breaker.
• objview: this is used in module.showModule and main.showScene. There might be a way around this, although the documentation says that it's a Perl script that calls rad which is not present in pyRadiance to my knowledge..
• main.saveImage and module.saveImage also uses rad which is not in pyRadiance. I'm sure there is a different workflow to save .hdr images and falsecolor images in pyRadiance
• pextrem function is used to normalize falsecolor image by finding the maximum value, but I'll bet that this can be substituted with some other non-Radiance (like numpy) function.
• gencumulativesky isn't in pyRadiance but gendaymtx is - we would have to switch to this process, per gendaymtx -- explore new radiance sky option ~  #184 and gendaymtx #549

See some discussion at the Radiance message boards here.
See some worked notebook examples from 2024 radiance workshop here.

[shirubana]

gendaymtx has some very nice spectral features now if I recall from the 2024 workshop, so yes double thumbs up to move that way