Hi,

Follow the last lines of report.html. I regenerated the client_id and client_secret and it keeps getting the same error.

File “C:\ProgramData\Anaconda3\envs\mm_2\lib\site-packages\sentinelhub\sentinelhub_session.py”, line 34, in init raise ValueError("Configuration parameters ‘sh_client_id’ and ‘sh_client_secret’ have to be set in order " ValueError: Configuration parameters ‘sh_client_id’ and ‘sh_client_secret’ have to be set in order to authenticate with Sentinel Hub service.

Hi,

I received your report.html file. The error inside seems to be:

HTTPSConnectionPool(host='services.sentinel-hub.com', port=443): Max retries exceeded with url: /oauth/token (Caused by SSLError(SSLCertVerificationError(1, '[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self signed certificate in certificate chain (_ssl.c:1129)')))

According to this I believe the issue is with an SSL certificate on your side and has nothing to do with Sentinel Hub credentials. I don’t know much about network certificates but there are a lot of posts on Stack Overflow, such as this one, of people having the same type of error. You might have to configure some permissions on your computer or your local network.

Is there any modification to the script that limits processing to less than 8Gb of RAM?

Training a ML model will take up some computing resources. Since you are in a cell that iterates over patches, you could try reducing the size of the patches when you define the grid, as shown in the example page.

The splitting choice depends on the available resources of your computer. An EOPatch with a size of has around 500 x 500 pixels at 10 meter resolution has a size of about ~1 GB.

I understand your explanation. I’m now trying to run on Google Colab (13 Gb RAM) and it’s showing these difficulties in the attachment:

Hi,

You are getting the same error message as in one of your previous posts. Please refer to the response: Slovenia jupyter notebook - #9 by maxim.lamare

Hi,

How to solve this problem?

ImportErrorTraceback (most recent call last)

in

38 from eolearn.mask import AddValidDataMaskTask

39 from eolearn.geometry import VectorToRaster, PointSamplingTask, ErosionTask

—> 40 from eolearn.features import LinearInterpolation, SimpleFilterTask, NormalizedDifferenceIndexTask

41 from sentinelhub import UtmZoneSplitter, BBox, CRS, DataCollection, MimeType, SentinelHubRequest, SHConfig

~/miniconda3/envs/py38/lib/python3.8/site-packages/eolearn/features/init.py in

9 KrigingInterpolation, LegacyInterpolation

10 from .feature_extractor import FeatureExtractionTask, FeatureExtendedExtractor

—> 11 from .feature_manipulation import SimpleFilterTask, FilterTimeSeries, ValueFilloutTask

12 from .haralick import HaralickTask

13 from .radiometric_normalization import ReferenceScenes, HistogramMatching, BlueCompositing, HOTCompositing, \

~/miniconda3/envs/py38/lib/python3.8/site-packages/eolearn/features/feature_manipulation.py in

17 import numpy as np

18

—> 19 from sentinelhub.time_utils import iso_to_datetime

20 from eolearn.core import EOTask, FeatureType

21

ImportError: cannot import name ‘iso_to_datetime’ from ‘sentinelhub.time_utils’ (/home/idies/miniconda3/envs/py38/lib/python3.8/site-packages/sentinelhub/time_utils.py)

Hi,

This function is depreciated since the last version of sentinelhub-py and has been removed.

You can use the following functions instead, depending on what you are trying to achieve:

• 
  

  sentinelhub.parse_time: time_utils — Sentinel Hub 3.4.0 documentation

  
  


• 
  

  sentinelhub.parse_time_interval: time_utils — Sentinel Hub 3.4.0 documentation