vheimhuber / inlettracker Goto Github PK

If you run the post-process check section #5.3 in InletTracker_master.py, check_inlet_state_detection changes the column 'bin_inlet_state' to 'Inlet_state' (if you run #5.3 in InletTracker_master.py), but plot_inlettracker_resultsV2 is looking for 'bin_inlet_state'.

I think the easiest edit be to change 'Inlet_state' back to 'bin_inlet_state' line 1769 of InletTracker_tools.py.

I would like to add Landsat 9 imagery to InletTracker. I tried to copy code from the updated CoastSat toolbox, but wasn't successful. Could you update to include it? Thanks!

Earth Engine authentication shows access blocked authorization error:400-invalid request while installation.

Hi Valentin,

Working with the InletTracker package I have had some problems related to raster files with incomplete bands, resulting in an interruption of the training process. I think that this could be handled as an exception by analyzing the georef variable. This is not a major issue, but it could avoid interruptions in the training process and having to start again.

Temporarily I solved this issue by removing the text files from the meta directory and loading the data again. But if there are lots of rasters with problem, this maual process could take a long time.

Please find attached a input_locations.shp file with sitename QUILIMARI, which raised this problem. Also, I attach here an example of a raster with no problems and raster which raised the error.

At the end of this message I pasted a screenshot of the error I got and of the rasters in the study area. Please note that the red area in the lower right is the raster with problems.

Regards,

input_locations.zip

sample_rasters.zip

Hi Tino -

I'm having trouble creating the inlettracker environment, since my version of Anaconda has a lot of dependencies that are more recent than those defined in environment.yml (e.g. current numpy is version 1.20.1 and the environment.yml file looks for - numpy=1.16.3=py37h19fb1c0_0, so it doesn't find the package). There seems to be a way to update/prune the environment.yml in anaconda once it is created, but I can't find a way to ignore the version numbers before it is created as an environment.

My error in $ conda env create -f environment.yml -n inlettracker is

Collecting package metadata (repodata.json): done
Solving environment: failed

ResolvePackageNotFound: 
  - zlib==1.2.11=h62dcd97_4
  - psutil==5.8.0=py37h2bbff1b

Is there a version-less way to do this?

Thanks,
Megan