Skip to content

RTE refactor#955

Merged
unbohn merged 99 commits into
devfrom
rte_refactor
May 27, 2026
Merged

RTE refactor#955
unbohn merged 99 commits into
devfrom
rte_refactor

Conversation

@pgbrodrick

Copy link
Copy Markdown
Collaborator

Draft refactor for tracking overall diff.

brentwilder and others added 30 commits April 24, 2026 11:37
Redefine RT Engine config to always be dict (no longer list)

Initialize new rt engine setup

Initialize new rt engine setup

Comment out ThermalIR test that will need to be reworked in the future

Correct to single engine in multipart check

Correct typo to config order which is forward_model, radiative_transfer, then engine

Mark skip for IR pytest and reduce redundancy in radiative_transfer.py
Backwards Compatibility for Configs
@pgbrodrick

Copy link
Copy Markdown
Collaborator Author

Seeing fairly noteable discrepancies in one-to-one tests with 3.x dev, looking for causes:

Screenshot 2026-05-18 at 11 35 44 AM

@pgbrodrick

Copy link
Copy Markdown
Collaborator Author

Seeing fairly noteable discrepancies in one-to-one tests with 3.x dev, looking for causes:

Uploading Screenshot 2026-05-18 at 11.35.44 AM.png…

@github-actions

github-actions Bot commented May 19, 2026

Copy link
Copy Markdown
📊 Generated results:

URL: isofit/isofit-test-results#32
SHA: e702b2c

@pgbrodrick

Copy link
Copy Markdown
Collaborator Author

Good news! This gives the same result as the above (IE, the new lut piece was seemeless) - still seems like we have an issue with externally provided wavelength files to hunt down though.

@unbohn

unbohn commented May 21, 2026

Copy link
Copy Markdown
Collaborator

More good news! Results from running an EnMAP scene (both per pixel and AOE) look pretty decent.

Per pixel:
image

AOE:
image

@evan-greenbrg

Copy link
Copy Markdown
Collaborator

Here are my suite of 1:1 comparisons from basic 6c Apply OE runs.

Summary: Some small deltas driven by differences in H2O lookup grids and atmospheric solutions.

Reflectance:

image image

Across the cliff:

image

Atmosphere:

image

@unbohn

unbohn commented May 22, 2026

Copy link
Copy Markdown
Collaborator

@evan-greenbrg Thanks! Overall, looking pretty good. Any idea what happened to the refactor AOD retrieval? I don't see any variation in values in the plot...

@evan-greenbrg

evan-greenbrg commented May 22, 2026

Copy link
Copy Markdown
Collaborator

@evan-greenbrg Thanks! Overall, looking pretty good. Any idea what happened to the refactor AOD retrieval? I don't see any variation in values in the plot...

@unbohn The apparent lack of variation is a function of the color stretch. Same figure without color stretch:

image

However, I'm not quite sure why the AOD is considerably more consistent. It does look a bit locked to the prior, but the prior variance should be the same between runs.

@evan-greenbrg evan-greenbrg self-requested a review May 27, 2026 21:50
@unbohn

unbohn commented May 27, 2026

Copy link
Copy Markdown
Collaborator

Thanks to the whole team for this major effort! Merging now for preparing the EARSeL tutorial v4.0.0 pre release.

@unbohn unbohn merged commit 39753f1 into dev May 27, 2026
41 of 42 checks passed
@unbohn unbohn deleted the rte_refactor branch May 27, 2026 21:55
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants