Note
For the Python package containing pre-built and re-usable tools, see :ref:`gladier_tools_package`
As noted in A simple Gladier application, a Gladier application defines a set of tools that are to be executed by the associated flow.
Gladier Tools are the glue that holds together Globus Flows and Compute Functions. A tool bundles everything that a Compute Function needs to run, so that the Gladier Client can register the function, check its requirements, and run it inside the Globus Flow.
The following code implements a simple tool, MakeDirs, that runs a single Compute Function, makedirs.
.. literalinclude:: example_tool.py :language: python
In the same module, this can be added to a client by simply adding the MakeDirs
class to gladier_tools:
# from gladier import GladierBaseClient @generate_flow_definition class ExampleClient(GladierBaseClient): gladier_tools = [MakeDirs]
A tool is composed of zero or more compute_functions and a flow_definition. A flow can be specified explicitly, or if it involves just a single Compute Function, can be generated automatically by using the @generate_flow_definition decorator. When applied to tools, this decorator will create a simple one-state flow for the function to run.
The main attributes of a Gladier Tool are here:
- compute_functions (list of callables) – A list of functions this tool uses.
- flow_input (dict): Default input that should be used in the flow. This is automatically overridden if the user supplies flow input to the Gladier Client
- required (list of strings): A list of critical flow_input keys that must be present for the function to be run. Gladier will raise an exception if these are not present when the user attempts to run the flow.
- flow_definition (dict): (Optional) A complete Globus Flow for running this tool. Provides a built-in flow new users can use to instantly run your tool.
The :ref:`gladier_tools_package` package provides a set of predefined, generally useful tools for common tasks.