Skip to content

Latest commit

 

History

History
31 lines (21 loc) · 1.48 KB

registry.rst

File metadata and controls

31 lines (21 loc) · 1.48 KB
Raw

Register your model

In order to extend the library, you need to register your classes and functions; it is done in two steps.

  1. Decorate your ~deeppavlov.core.models.component.Component (or ~deeppavlov.core.data.dataset_reader.DatasetReader, or ~deeppavlov.core.data.data_learning_iterator.DataLearningIterator, or ~deeppavlov.core.data.data_fitting_iterator.DataFittingIterator) using ~deeppavlov.core.common.registry.register and/or metrics function using ~deeppavlov.core.common.metrics_registry.register_metric.
  2. Rebuild the registry running from DeepPavlov root directory:
python -m utils.prepare.registry

This script imports all the modules in deeppavlov package, builds the registry from them and writes it to a file.

However, it is possible to use some classes and functions inside configuration files without registering them explicitly. There are two options available here:

  • instead of {"class_name": "registered_component_name"} in config file use key-value pair similar to {"class_name": "my_package.my_module:MyClass"}
  • if your classes/functions are properly decorated but not included in the registry, use "metadata" section of your config file specifying imports as "metadata": {"imports": ["my_local_package.my_module", "global_package.module"]}; then the second step described above will be unnecessary (local packages are imported from the current working directory).