Aller au contenu principal

Action execution

An action is a set of instructions that will be executed sequentially. To learn more about them, read the user guide page about actions.

Executing actions from CLI​

This is the simplest way of executing an action. Once silex is installed, run silex action <action-name>.

tip

If you installed silex client as a rez package, don't forget to execute this command in the rez environment: rez env silex_client -- silex action <action-name>

Task ID​

Some actions behave differently according to the context you are executing them from (for example, the publish will set the publish location using the current task defined in the context). You can specify the task using the --task-id argument with the ID if the entity in the CGWire database.

Parameters​

Silex actions sometimes prompt the user with parameters. You can set parameters ahead with the --parameters argument. The expected value is <parameter-path>=<parameter-value> the parameter will be automaticaly cast into the expected parameter type. To set multiple parameters just use this argument multiple times.

Batch mode​

Sometimes you want to execute actions in headless mode (on the renderfarm or on ssh), in this case you don't have access to the Silex UI and when the user will be prompted, the action execution will be stuck waiting for an input in the UI. To prevent that you can use --batch to tell silex to not try to connect to the UI via websocket. All the parameters that require a user input will keep their default values.

Categories​

If your action is not in the default category (the actions category) you can specify a different one by using the --category argument (for more information see the action definition page)

tip

To get more information about the silex client CLI, use silex action --help

Executing actions from Python​

When executing an action from a dcc (with a shelf button) you want to execute the action from python:

from silex_client.action.action_query import ActionQuery

action = ActionQuery("<action-name>")
action.execute()

Categories​

If your action is not in the default category (the actions category) you can specify a different one by using the category parameter when instanciating the action (for more informations see the action definition page)

Methods​

Executing an action from python is much more flexible because the ActionQuery object gives you control on the execution at runtime. Here is a list of usefull methods on the ActionQuery, they can be used while the action is executing.

# Some of the following methods have an async version (prefixed with 'async_')
# use those if you want to use them from the silex event loop.
# (see the event loop page)

# Set a parameter
action.set_parameter("<parameter-path>", <value>)

# Get a command
action.get_command("<command-path>")

# Stop the execution if executing and go back one command
action.undo()

# Pause the execution
action.stop()

# Resume the execution
action.redo()

# Synchronize with the UI, use this method if you modified
# the buffer of the action directly (like get a command and modify its data)
action.update_websocket()

Debugging an action​

When an execution is not going right, you can debug its execution using some action query methods. The following instructions cannot be used for actions executed from the CLI.

  • Make sure to set the log level to INFO, the default is WARNING
from silex_client.utils.log import logger
logger.setLevel("INFO")
caution

DEBUG is very verbose (it logs every diffs of the websocket synchronisation process) so INFO is enough most of the time

  • If you didn't store the ActionQuery instance into a variable (or the action was started from a shelf button) you can get the ActionQuery instance with its ID from the Context:
# To get the ID, click on the button next to the context in
# the panel of the action on the UI. It will store the ID in you clipboard.
from silex_client.core.context import Context
action = Context.get().actions["<action-id>"]

  • You can use the execute method with the step_by_step argument (action.execute(step_by_step=True)). This will execute only one command and pause directly. From that point it will be easier check all the values of the output and inputs. To go to the next command, use action.execute(step_by_step=True) again.

  • Some commands / parameters are hidden, you might want to set them to visible, here is a snippet to set every commands:

for command in action.commands:
    commands.hide = False

# Always use this method after modifying the buffers directly, to keep synced with the UI
action.update_websocket()
  • You can check the output of a command at any time, with the output_result property.
print(action.command[<command-index>].output_result)
  • You can check the input of a parameter at any time with the get_value method. Parameter input can be a link to a command output. This method will resolve all the links if any and give you the real value directly.
print(action.command[<command-index>].parameters["<parameter-name>"].get_value(action))