Scene.load error on conflicting 'y' values with MSG example. #1517
Labels
Comments
|
Regarding the As for your error, could you print out the filenames you are providing? The error seems to suggest that Satpy found 3712 of image data but 7424 of "time" data. |
|
Re. The file list is pretty huge, good luck: |
|
I'm not an MSG expert (far from it), but it looks like you have two separate images in this directory, right? One for 202009060000 and one for 202009060015? If this is the case, Satpy can only really handle one time step at a time. Two images also makes sense with the error message (one number is twice as much as the other). If you pass the files for a single time, does it work? |
|
I was just writing that you might end up with two time slots worth of data with import glob
fnames = glob.glob("/path/to/data/decompressed/2020/20200906/00/H*202009060000*__")
scn = Scene(reader="seviri_l1b_hrit", filenames=fnames) |
|
djhoese — changing the second pnuu — that works! I now have a ginormous view of the Earth and it's beautiful! I think the example script given by satpy at https://nbviewer.jupyter.org/github/pytroll/pytroll-examples/blob/master/satpy/hrit_msg_tutorial.ipynb should be modified for your method. Is there a benefit to specifying a Along with the nice image I do have more errors, which I believe is a different issue relating to |
|
I'll let @pnuu answer the start/end time question. Regarding hidden stuff: It is mostly that Satpy's Scene doesn't have any logic right now for detecting how many scenes/time steps you've given it so the error messages produced are always a little cryptic. Regarding the warnings: The pyproj ones are mostly an issue from newer versions of pyproj deprecating the use of "PROJ Strings". We are slowly updating parts of pyresample to not use PROJ strings but it is slow as they are used everywhere. The last warning about invalid values are expected as numpy/dask will report these anytime a NaN is used. Satpy uses NaNs for invalid pixels including space pixels (like you'll see in a full disk image). |
|
Since the warnings are so huge, can they be suppressed to improve readability with code similar to The information you described above would be great to add to a markdown box in the example MSG |
|
It is more likely to happen if you make an issue on the pytroll-examples github repository 😉 |
|
Ha! I am new to Git but I've found the notebook and will attempt to "commit" my "proposed" changes... 😁 |
Benblob688
pushed a commit
to Benblob688/pytroll-examples
that referenced
this issue
Jan 29, 2021
Based on the thread at pytroll/satpy#1517, this is a better example to avoid importing two timesteps onto a scene, and adds clarification about the warnings which are expected but would throw off a new user.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I am new to SatPy and have an issue using the basic tutorial "example" with MSG data.
The example code is here: https://nbviewer.jupyter.org/github/pytroll/pytroll-examples/blob/master/satpy/hrit_msg_tutorial.ipynb
and I have supplied some data I have from a project archive used by other people, here for your recreation/testing: https://drive.google.com/drive/folders/1fpZBsGy3GnoD-r7RpuA3fbSCKls4Xs34?usp=sharing
I get an error with the
Scene.loadsyntax.Btw, the example says to use 'natural' for the composite, but that does not exist in my scene, I have used
scn.available_composite_ids()to get the list, and have chosen one from there.The code:
The very large error (I feel the start is unrelated, but have included it for completeness)
I setup a new environment today so all packages are the most recent as suggested, doing commands like
pip install satpyand so on. Python 3.8 is being used.The text was updated successfully, but these errors were encountered: