Ready Made Pipeline python-stanford-stanza

After loading the Python plugin, the prepared pipeline python-stanza is available from Applications - Ready Made Applications - Python - python-stanford-stanza.

This pipeline allows to annotate a document using the Stanford Stanza NLP software.

In order to use this pipeline, the following requirements must be met:

• Python package stanza must be installed
• The Stanza model for the desired language must be installed

Installing Stanford Stanza

Run the following command:

python -m pip install -U stanza

Install the Model for the Language

See the Stanza documentation for this and Download Models For example to install the default English model run the following command:

python  -c 'import stanza; stanza.download("en")'

Running the Pipeline

The following parameters can be set in the programParams runtime parameter for the ready made pipeline.
This contains a subset of what can be specified directly in Python when a Stanza Pipeline is created, see the Stanza Documentation

• outputAnnotationSet: name of the annotation set where the Stanza annotations should be placed. Default is Stanza
• lang: the language for the model to use, e.g. en. Default is: en
• dir: directory where the downloaded model is stored. If not specified the default download location is used.
• logging_level: one of “DEBUG”, “INFO”, “WARN”, “ERROR”, “CRITICAL”, “FATAL”, default depends on verbose, see below
• verbose: if True, but now logging level is specified, loggin level is set to INFO, otherwise toERROR
• use_gpu: if True, attemt to use the GPU, if available. Set to False to avoid using the GPU if one is available.

NOTE: any parameter which is included in programParams but not set / empty, is ignored.