Add support for local Docker images

[RobertLucian]

Closes #1094.

When validating for the local provider, Cortex first checks if the Docker serving image is available locally and if it isn't, then it checks if it's on DockerHub or on AWS ECR and if it can't find there either, then it will throw the docker client's error (like usual).

Next, before the container is launched, Cortex first tries to pull the image and if it isn't available, then it checks if there's a matching local image and if that too fails, it returns an error message of this sort: "failed to pull image".

---

checklist:

• run make test and make lint
• test manually (i.e. build/push all images, restart operator, and re-deploy APIs)

[deliahu]

Thank you for adding this!

[deliahu]

LGTM!

[RobertLucian]

Tested this the following ways assuming a local provider:

1. With the image present both locally and on the upstream (i.e. DockerHub/ECR) -> worked as expected.
2. With local image present and none available on the upstream (i.e. DockerHub/ECR) -> proceeded to use the local image.
3. With local image not present and none available on the upstream (i.e. DockerHub/ECR) -> failed as expected.
4. With local image not present and one available on the upstream (i.e. DockerHub/ECR) -> proceeed to download the upstream image.
5. With an API already running using an xyz image, build a different image locally and then hit cortex deploy -> proceeded to use the newly updated local image.