Skip to main content

Notebook Examples

Notebook examples focus on reviewable workflows rather than isolated API calls.

Executable notebooks are available in the repository notebooks/ directory:

  • terminology_lookup.ipynb: lookup, search, source inventory, and term type review
  • patient_friendly_mapping_review.ipynb: patient-friendly names and ICD10CM-to-SNOMED mapping review
  • valueset_optimization.ipynb: optimize diagnosis value sets and inspect include/exclude rules
  • ndc_rxnorm_resolution.ipynb: normalize NDCs, resolve RxCUIs, and preserve original medication identifiers

Notebook examples should keep provenance columns visible. Do not hide match_type, match_depth, matched_via, original source code, resolved code, or review flags.

Run the notebook smoke suite locally:

make notebook-smoke

Base pattern:

import medterm4ds as mt

terms = mt.connect("/mnt/d/medterm4ds/data/umls_current.duckdb", memory_profile="low")

df = terms.search_df("diabetes", sources=["ICD10CM", "SNOMEDCT_US"], limit=25)
df.head()

For mixed terminology work, build CodeRef inputs:

refs = [
    mt.CodeRef("ICD10CM", "E11.9"),
    mt.CodeRef("RXNORM", "1049502"),
    mt.CodeRef("CVX", "208"),
]

terms.patient_friendly_df(refs)