Problem with dataset 'geo_color'

[B. Schitter]

Hi,

I tried with a very simple script (see attached file) to calculate the product 'geo_color'.

I can load the dataset 'geo_color' but when I do a resample I get the following message:

/opt/met-mtg/lib64/python3.12/site-packages/satpy/readers/generic_image.py:83: UserWarning: The specified chunks separate the stored chunks along dimension "x" starting at index 4096. This could degrade performance. Instead, consider rechunking after loading.
  data = xr.open_dataset(self.finfo["filename"], engine="rasterio",
cannot convert float NaN to integer
rsr data may not be up to date: /opt/met-mtg/share/pyspectral/rsr_fci_Meteosat-12.h5
The following datasets were not created and may require resampling to be generated: DataID(name='geo_color')

How can I calculate the dataset 'geo_color' correctly?

When I do the same with 'dust', 'airmass', ... I don't get this message.

Regards,

Bernhard

[B. Schitter]

Here the script.

[Andrea Meraner]

Hi Bernhard, 

you are likely encountering the issue as described here: https://github.com/pytroll/satpy/issues/2733 

The solution is to use generate=True, or to resample twice the composite - check which one is faster for your use case.

Cheers, 

Andrea

[Holl Gerrit]

Hi Bernhard,

as the message says, you need to resample after the load call:

ls = sc.resample(resampler="native")

You get this message because geo color combines channels with different resolutions, and it needs to resample them onto a common grid in order to calculate the composite.

Regards,
Gerrit.

--
Remote Sensing Application Development
Research and development
Deutscher Wetterdienst (DWD)
+49 (69) 8062-3171

________________________________________
From: pyt...@googlegroups.com <pyt...@googlegroups.com> on behalf of B. Schitter <bernhard...@gmail.com>
Sent: 14 July 2025 11:21:53
To: pytroll
Subject: [pytroll] Problem with dataset 'geo_color'

--
You received this message because you are subscribed to the Google Groups "pytroll" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pytroll+u...@googlegroups.com<mailto:pytroll+u...@googlegroups.com>.
To view this discussion, visit https://groups.google.com/d/msgid/pytroll/15ae93b0-323d-4e26-a25b-7f169c0a0f8bn%40googlegroups.com<https://groups.google.com/d/msgid/pytroll/15ae93b0-323d-4e26-a25b-7f169c0a0f8bn%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Johan Strandgren]

Hi Bernard,

The message you get is expected for geo_color since it uses two static global files (land water mask and NASA black marble with citylights) in the background that have to be resampled. However, you do resample after the loading so in principle it should work. Did you try to save the image and actually get an error? Because the error message you get is from the load call, but should be fixed by the following resample call.

Cheers,
Johan

--
You received this message because you are subscribed to the Google Groups "pytroll" group.

To unsubscribe from this group and stop receiving emails from it, send an email to pytroll+u...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/pytroll/9eb22356-1b7a-46e6-aeda-5482bd91f860n%40googlegroups.com.

--

Johan Strandgren

johan.st...@gmail.com

Tel: +49 1523 611 3407

[B. Schitter]

Dear Andrea, Gerrit and Johan,

thank you very much for your help - now I can create the image.

Kind regards,

Bernhard

[B. Schitter]

Hi again,

I am experimenting a little bit with all available datasets.

During processing "true_color_reproduction" I get the following error:

KeyError: 'No conversion matrix found for platform Meteosat-12'

What does this mean?

FYI: I am working on an offline machine with no connection to the internet. I set the environment variables PSP_CONFIG_FILE and SATPY_DATA_DIR pointing to local directories.

Kind regards,

Bernhard

[Andrea Meraner]

Hi Bernhard, 

I believe you found a bug: in this dictionary we still use `mtg-i1` even though the platform is now `meteosat-12`  https://github.com/pytroll/satpy/blob/main/satpy/enhancements/__init__.py#L640

I will try to do a PR for the fix. In the meanwhile, as a hotfix, you can try to modify that file in your environment manually if you want to test it.

Cheers, 

Andrea

[B. Schitter]

Dear Andrea,

thank you for the answer. The modification in my environment works for me at the moment :-)

The last dataset on which I have "problems" is the "simple_fci_fire_mask". I can load and resample, but when I try to save the image (png) it I get a TypeError (see attachment).

Kind regards,

Bernhard

[Andrea Meraner]

Hi Bernhard, 

good to hear! 

Interesting, I never tried to save that composite by itself - it is a binary mask that is used to identify fires during the generation of the "true_color_flames_with_night_ir105" composite. The fact that it is boolean is probably causing the issue - I will look into it. However it is not meant to be used by itself but it's rather only a component of the other dataset, so you can just ignore it.

Cheers, 

Andrea