Hello,

I have deployed a Python function as a containerized AWS Lambda function, which uses sentinelhub-py to request Sentinel-2 images via the Process API. However, I am encountering an issue when initializing SHConfig(). It appears that SHConfig attempts to write to a local configuration file (config.toml) in the default directory (~/.config/sentinelhub/config.toml), which is not writable in the Lambda environment due to its read-only file system.

Since AWS Lambda only allows writing to the /tmp directory, I tried redirecting the configuration there, but I still encounter the following error:

Error configuring Sentinel Hub client: Errno 30] Read-only file system: '/home/sbx_user****'
 

My Approach:

I am passing the client_id and client_secret as environment variables and attempting to initialize the configuration as follows:

import os
from sentinelhub import SHConfig, SentinelHubRequest

# Retrieve credentials from environment variables
CLIENT_ID = os.environ.get('SH_CLIENT_ID')
CLIENT_SECRET = os.environ.get('SH_CLIENT_SECRET')

def handler(event, context):
    SHdata = AccessSentinelHub()
    return {"Message": "Fetched Data Successfully"}

def AccessSentinelHub():
    try:
        config = SHConfig(SH_CLIENT_ID=CLIENT_ID, SH_CLIENT_SECRET=CLIENT_SECRET)
        if not config.sh_client_id or not config.sh_client_secret:
            print("Warning! To use the Process API, please provide valid credentials (OAuth client ID and client secret).")
    except Exception as e:
        print("Error configuring Sentinel Hub client:", str(e))

# Rest of the code here using SentinelHubRequest()

Is it possible to initialize SHConfig() with the necessary credentials without it attempting to write to a configuration file or change the default read/write directory for SHConfig to avoid permission issues?

Or is there an alternate way to config to access SentinelHub data in python for the SentinelHubRequest function.

I’m relatively new to using SentinelHub so would appreciate any help/corrections. 

Thank you.