Caliban

- 'caliban run' and 'caliban shell' now take an --image_id argument; if
provided, these commands will skip their 'docker build' phase and use the
image ID directly.

- AI Platform labels now swap periods for underscores (thanks to vinay!); this
means that floating point numbers will no longer have pre- and post- decimal
components concatenated.

- A new `expansion` script will expand experiment configs into all of the
individual experiments they'd generate. This command can accept `stdin`, just
like the `--experiment_config` argument. Options include `--pprint` and
`--print_flags`. The output of this script can be piped directly into `caliban
cloud --experiment_config stdin`.

- `caliban shell` will now default to bash if you're using a shell that's not
`bash` or `zsh` (fish shell, for example) instead of erroring out.

- `caliban shell` has a new `--shell` argument that you can use to override the
container's default shell.

- consolidated gke tpu/gpu spec parsing with cloud types
- modified all commands to accept as the module argument paths to arbitrary
shell scripts. Any argument of the format "trainer.train" will execute using
"python -m trainer.train", just as before. If instead you pass a python script
as a file, like "trainer/train.py", caliban will execute this file inside the
container using "python trainer/train.py". Any other argument, if it exists in
the local directory, will be executed as a bash script.

This allows users to run commands like "caliban cloud my_script.sh" and have
it all work.

- "caliban run" now supports --experiment_config and --dry_run. These work just
like they do for "caliban cloud"; the experiment config will expand out and
execute N jobs on your local machine.
- moved some methods from cluster/cluster.py to gke/utils.py
- added unit tests for some gke/utils.py methods
- Support for ADC credentials! if application_default_credentials.json is
present on the user's machine, they now get copied into the container.
- if ADC credentials are NOT present but a service account key is we write a
placeholder. this is required to get ctpu working inside containers.

- added tpu driver specification for gke jobs
- added query for getting available tpu drivers for cluster/project

- set host_ipc=True for cluster jobs

- moved cluster constants to separate file
- moved cluster gpu validation to separate file
- added test for gpu limits validation

- TPU and GPU spec now accept validate_count arg to disable count validation.