Design issues related to autonomy deploy from-token

[jmoreira-valory]

Is your feature request related to a problem? Please describe.
There are 2 related design issues related to the command autonomy deploy from-token .

Issue 1
The flags --docker and --kubernettes achieve different outcomes. This is roughly how the from-tokenworks:

1. Find the hash of the provided service token ID
2. Fetch the service (similar to autonomy fetch)
3. Build images (similar to autonomy build-image)
4. Build the deployment (similar to autonomy deploy build)
5. IF --kubernetes or --no-deploy, quit,
   ELSE run docker deployment.

From the point of view of an user this "asymmetric behaviour" between the flags --docker and --kubernetes can be confusing and not desirable.

Issue 2
The reason why the command is implemented this way is to be able to build deployments given a service ID on-chain. This suggests that the Step 1) above should be integrated in the command fetch service (with a token ID).

Describe the solution you'd like
Some of these should be considered:
a) Make that autonomy from-token behave equal for kubernettes and docker.
b) Integrate Step 1 in autonomy fetch
c) If b) is implemented and not a), remove kubernettes from the from-token command. This way the user will know until what step he has "kubernetes" support form the framework (autonomy fetch - build-image - deploy build, but NOT deploy run or from-token).

Describe alternatives you've considered

Additional context