The pipeline will fail on the first unknown term it encounters, so if there are multiple new terms, then the pipeline has to be re-run multiple times, identifying one new term each time, until it succeeds. Preferably, if there are unknown terms, the pipeline would collect them all and include them in the error message. This would make it much easier to update the categories.
Currently, new clinical significance terms are when the import pipeline fails to map them to a category.
https://github.com/broadinstitute/genetic-prevalence-estimator/blob/121e5cd7cb18a2628fe3fdd0db9806ac6e648554/data-pipelines/import_clinvar.py#L195-L203
The pipeline will fail on the first unknown term it encounters, so if there are multiple new terms, then the pipeline has to be re-run multiple times, identifying one new term each time, until it succeeds. Preferably, if there are unknown terms, the pipeline would collect them all and include them in the error message. This would make it much easier to update the categories.