Drug indication data requires a disease normalizer instance, so the CLI currently tries to identify a disease_concepts table location and update it if necessary. However, there are a few condition branches where various values aren't properly set, raising UnboundLocalErrors.
Right now we can use DISEASE_NORM_DB_URL to set a URL, and I think if you define it as a blank string it will push up to prod (and then use the environment set in DISEASE_NORM_ENV). You can also set THERAPY_NORM_PROD to prevent attempts to check/update the disease database. I think we just need to ensure documentation is updated (if only for my sake).
Oh, also, the CLI _check_disease_normalizer method needs to check against ChEMBL too, not just HemOnc.
Drug indication data requires a disease normalizer instance, so the CLI currently tries to identify a
disease_concepts
table location and update it if necessary. However, there are a few condition branches where various values aren't properly set, raising UnboundLocalErrors.Right now we can use DISEASE_NORM_DB_URL to set a URL, and I think if you define it as a blank string it will push up to prod (and then use the environment set in DISEASE_NORM_ENV). You can also set THERAPY_NORM_PROD to prevent attempts to check/update the disease database. I think we just need to ensure documentation is updated (if only for my sake).
Oh, also, the CLI
_check_disease_normalizer
method needs to check against ChEMBL too, not just HemOnc.