Skip to contents

Maintainer Guide

maintainer-guide.Rmd

This article documents workflows for maintaining and rebuilding package datasets. It is intended for the package maintainer and contributors, not general users.

disease_eponyms — Eponym capitalization vector

disease_eponyms is a named character vector mapping lowercase words to their correctly capitalized forms (e.g. c("waardenburg" = "Waardenburg")). It is used as the default eponyms argument to parse_omim_name().

Source files

File Purpose
data-raw/omim_eponyms.R Mines OMIM for capitalization candidates; updates disease_eponyms_curated.tsv
data-raw/disease_eponyms_curated.tsv Shared hand-curated TSV; single authoritative source for all candidates regardless of provenance
data-raw/build_disease_eponyms.R Reads the curated TSV and saves the .rda dataset

Curated TSV columns

Column Description
word_lower Lowercase word (primary key; unique across all sources)
word_cap Correctly capitalized form; NA for contested words until manually resolved
alt_caps Competing capitalization forms with counts, e.g. "MacLeod (48); Macleod (2)"
examples Up to 3 source names where the word appeared
status "cap" (capitalize), "lower" (leave lowercase), or "pending" (awaiting review)
source Provenance of the candidate (e.g. "OMIM"); used to scope refreshes per source
notes Free-text annotation (optional)

Workflow

  1. Mine candidates — run a source-specific script to update disease_eponyms_curated.tsv:

    source("data-raw/omim_eponyms.R")   # requires OMIM API key; see ?download_omim

    New candidates are appended with status = "pending". Existing OMIM rows have alt_caps and examples refreshed from the latest download.

  2. Review pending rows — open data-raw/disease_eponyms_curated.tsv and for each "pending" row set status to:

    • "cap" — this word should be capitalized in output
    • "lower" — this word should stay lowercase

    For contested words (alt_caps is non-empty), verify word_cap before approving. Roman numerals and alphanumeric suffixes (e.g. IIb, 2a) are handled automatically by fix_disease_caps() and should be marked "lower".

  3. Rebuild the dataset:

    source("data-raw/build_disease_eponyms.R")

Adding a new source

Create a new mining script (e.g. data-raw/do_eponyms.R) that:

  1. Generates a candidate data frame with columns word_lower, word_cap, alt_caps, examples.
  2. Reads disease_eponyms_curated.tsv, appends rows for words not already present (with source = "<SOURCE>" and status = "pending"), and writes the file back.
  3. Does not touch rows belonging to other sources.

After curation, run data-raw/build_disease_eponyms.R as above.

disease_cap_patterns — Phrase-level capitalization pattern vector

disease_cap_patterns is a named character vector of regex substitutions applied after disease_eponyms by parse_omim_name(). Use it for words whose correct capitalization depends on context (e.g. SHORT as an acronym in SHORT syndrome vs short as an adjective elsewhere).

Names are case-insensitive regex patterns; values are replacement strings. Longer patterns take priority over shorter ones and override conflicting disease_eponyms entries.

Source file

data-raw/disease_cap_patterns.R — fully hand-curated; no mining step.

Workflow

  1. Open data-raw/disease_cap_patterns.R and add a named element to the disease_cap_patterns vector, e.g.:

    disease_cap_patterns <- c(
  disease_cap_patterns,
  "\\bshort syndrome\\b" = "SHORT syndrome"
)

  2. Rebuild the dataset:

    source("data-raw/disease_cap_patterns.R")