Home - GateNLP/gateplugin-ModularPipelines GitHub Wiki
#GATE ModularPipelines Plugin
The GATE Modular Pipelines Plugin supports two important aspects of developing GATE pipelines, especially more complex pipelines:
- Support modular development, where building blocks can be developed independently and combined to increasingly larger pipelines.
- Support parametrisation of pipelines through config files. Rather than having to update runtime parameters this allows to control how exactly the PRs in a pipeline work from outside the pipeline.
Modularization
This works by representing each contained pipeline as a Processing Resource (Pipline PR) which refers to a a pipeline file. Thus a nested pipeline gets loaded into the containing pipeline via the Pipline PR and the nested pipeline can in turn contain Pipline PRs that may load still further nested pipelines.
Parametrization
- The Parametrized Corpus Controller is a modified version of a Conditional Corpus Controller and it is associated with a configuration file that will be used to change runtime parameters of its PRs, or set document features. The plugin adds a menu entry to the GUI to change an existing Conditional Corpus Controller to a Parametrized Corpus Controller. In addition, the settings from the config file can be overridden by system property settings.
You can find more information in the following Wiki pages:
- Pipeline PR: The "Pipeline" processing resource allows including sub-pipelines getting re-presented as processing resources in a main pipeline and allows setting run-time parameters and document features from a config file.
- Parametrized Corpus Controller: A conditional corpus controller that can be configured from a config file.
- Configuration File documents the format of the config file and the format of the system properties for configuring a Parametrized Corpus Controller
- Downloads: https://github.com/johann-petrak/gateplugin-modularpipelines/releases