Add generator subcommand to ansible-dk CLI #65
Conversation
…ke it a dependency of ansible-dk-cli
| @click.option('--generator', '-g', default=get_config()['generator'], help='Path to custom generator repo') | ||
| @click.option('-e', 'extra_vars', multiple=True, callback=parse_extra_vars, help='extra vars to pass to ansible, var=value') | ||
| @click.argument('name') | ||
| def generated_method(debug, generator, name, extra_vars): |
There was a problem hiding this comment.
We should comment this section to describe your intent as to what is going on here. Cleverness can be harmful, remember this code has to be maintained by the whole team.
|
When we add a CLI command, we should add real help. This should be describing what the generate command does, from a user perspective, and give a pointer to more detailed docs. [clinton@and-cheese ansible-dk]$ ansible-dk-cli/ansible-dk generate --help
Usage: ansible-dk generate [OPTIONS] COMMAND [ARGS]...
Options:
--help Show this message and exit.
Commands:
config
filter
inventory
playbook
repo
role |
|
|
||
| ################################################### | ||
| # Defaults | ||
| RUBY_ABI = "2.1.0" | ||
| PYTHON_ABI = "2.7" | ||
| HOME = os.getenv("HOME") | ||
|
|
||
| def get_config(debug=False): | ||
| #TODO: pass debug properly | ||
| config_path=os.path.expanduser('~/.ansible-dk/config') |
There was a problem hiding this comment.
Could we make this config.yaml instead of plain config, to make it obvious?
|
I'm thinking we'd like to be able to leave open the possibility of shipping multiple stock generators. |
| run_ansible(playbook_path, extra_vars) | ||
| generated_method.__name__ = 'generate_%s' % thing | ||
|
|
||
| generate_things = ['repo', 'role', 'filter', 'inventory', 'config', 'playbook'] |
There was a problem hiding this comment.
This is wrong - repo should treat its 'name' parameter differently, and needs to create the directory. The implementation of 'repo' is different than the others.
| variable_manager = VariableManager() | ||
| inventory = Inventory(loader=loader, variable_manager=variable_manager, host_list='localhost') | ||
| variable_manager.set_inventory(inventory) | ||
| if vars: |
There was a problem hiding this comment.
We will ALWAYS have vars to pass into the run. We need to determine what they are, but a good start would be:
adk_generator_command # 'role' in ansible-dk generate role pony
adk_generator_target_path # ./pony if we're in repo mode, . otherwise
adk_generator_target_name # 'pony' in ansible-dk generate role pony
adk_generator_source # /opt/ansible-dk/generators/default or whatever
The ansible-dk CLI will maintain a list of 'context vars' that it passes to the ansible run. We must document these in the README or somewhere else.
not ready - just starting the pull request
Update: I believe this is now ready to merge, have done a test build and all seems well.
We will still need to make sane default playbooks, the ones in place now are dummy placeholders.