Summary
The user is encountering a "ModuleNotFoundError" for the 'pystac_client' module while using a Docker container in the GCP artifact repository, despite having all necessary libraries included. They have updated the Docker tag, but the command pyflyte run --remote still refers to an old tag. The user believes they may have fixed the tagging issue and is prepared to register workflows, but is still worried about the module import issue. They mention using a prebuilt Docker image and creating an image specification for the task. The user is interested in Flyte's approach to using a separate Flytekit Docker for running workflows while referencing other Docker images with specific libraries. They suggest that keeping imports within the task could be helpful. Additionally, they ask about the base image configuration and the feasibility of running the custom Docker image to verify the availability of 'pystac_client', as well as the workflow registration command. They confirm that the libraries are accessible in their local container and provide a Dockerfile and registration process based on existing FlyteLab repositories.