Interactive curation application

class cord19kg.apps.curation_app.CurationApp

JupyterDash-based interactive entity occurrence curation app.


Get the current curated data table.


table – Current curation data table. The table is indexed by unique entities and contains the following columns: paper/section/paragraph sets of papers/sections/paragraphs where the entity occurs, aggregated_entities set of unique raw entities associated with the entity (by ontology linking, if performed), uid ontology concept id, definition ontology concept definition, paper_frequency number of unique papers where occurs, entity_type type of the entity computed either from the raw NER types or the ontology concept type.

Return type:


run(port, mode='jupyterlab', debug=False, inline_exceptions=False)

Run the curation app.

  • port (int) – Port number to launch the app (localhost:<port>).

  • mode (str, optional) – Mode in which the app should be launched. Possible values: inline inside the current Jupyter notebook, external given an external link that can be opened in a browser, jupyterlab as a new tab of JupyterLab.

  • debug (bool, optional) – Flag indicating whether the app is launched in the debug mode (traceback will be printed).

  • inline_exceptions (bool, optional) – Flag indicating whether app exceptions should be printed in the current active notebook cell.


Set default terms to be fixed in the table.

The application allows to set fixed terms that are included in the table at all times (even when don’t satisfy the conditions of current filters).


terms (list of str) – List of terms to fix in the data table


Set the ontology linking callback.

This function will be called when the Link ontology button is clicked. The curation app will pass the current curation table object to this function.