Changelog

We use Commitizen to automatically increase the versions, release the package on GitHub (with a tag), and auto-generate this NEWS file. The items listed below are generated based on our commit messages that follow the Conventional Commits standard.

0.9.19 (2026-01-12)

Fix

• :bug: URL in docs gave CRAN check error, change to citation (#474)

0.9.18 (2026-01-09)

Fix

• :bug: don’t download from URL inside a function (#470)

0.9.17 (2025-12-04)

Fix

• :bug: split out case data from sim, joining them as one in tests (#456)

0.9.16 (2025-12-01)

Perf

• :zap: classification requires DuckDB/SQL (#410)

0.9.15 (2025-11-27)

Refactor

• :truck: move ICD8, ICD10, and sim definitions into sysdata (#444)

0.9.14 (2025-11-26)

Perf

• :zap: use union_all() for database version of bind_rows() (#439)

0.9.13 (2025-11-21)

Refactor

• ♻ replace stringr functions with base alternatives in algorithm (#419)

0.9.12 (2025-11-21)

Refactor

• :recycle: replace str_detect() with grepl in algorithm (#416)

0.9.11 (2025-11-20)

Refactor

• :truck: use from_*() prefix for logical inclusion helper variables (#415)

0.9.10 (2025-11-20)

Refactor

• ♻ add logical inclusion variables to classification logic (#404)

0.9.9 (2025-11-20)

Perf

• :zap: duckplyr doesn’t support lubridate::as_date(), use as.Date() (#408)

0.9.8 (2025-11-20)

Perf

• :zap: DuckDB can’t use rename_with(), switch to rename() (#409)

0.9.7 (2025-11-19)

Fix

• :bug: Non-cases test filter (#412)

0.9.6 (2025-11-19)

Fix

• 🐛 convert NAs to FALSE at end in prep for classifying (#379)

0.9.5 (2025-11-19)

Fix

• 🐛 changes to drop_pregnancies() to support the new refactored logic (#405)

0.9.4 (2025-11-19)

Fix

• :bug: don’t use collect() or compute() inside code, only outside (#403)
• :bug: convert NAs to FALSE and drop any if TRUE for pregnancy (#398)

0.9.3 (2025-11-19)

Fix

• :bug: don’t need to drop indication code after exclusion (#395)

Refactor

• ♻ add logical helper variables for inclusion events (#400)

0.9.2 (2025-11-17)

Fix

• :bug: wrong value in the c_diag of the cases data (#394)

0.9.1 (2025-11-17)

Refactor

• :recycle: enhance (non-)case data with sim data and convert to duckdb tibble (#396)

0.9.0 (2025-11-14)

Feat

• :sparkles: require using DuckDB in classify_diabetes() (#391)

0.8.8 (2025-10-05)

Fix

• :bug: logic around is.na() in pregnancy event isn’t needed (#367)

0.8.7 (2025-10-03)

Refactor

• :recycle: deselect columns we don’t need later (#377)

0.8.6 (2025-10-03)

Fix

• :bug: use .data$ instead of quotes around is_hba1c (#376)

0.8.5 (2025-10-03)

Fix

• :bug: remove HbA1c rows when calculating insulin columns (#371)

0.8.4 (2025-10-03)

Refactor

• ♻ add separate keep_two_earliest_events() function (#369)

0.8.3 (2025-10-02)

Refactor

• :truck: rename cases() to edge_cases() to match file name (#368)

0.8.2 (2025-10-02)

Fix

• :bug: use full_join() instead of left_join() (#365)

0.8.1 (2025-10-02)

Fix

• :bug: remove rows with NA in both T1 and T2D (#363)

0.8.0 (2025-10-02)

Feat

• :sparkles: add non-case of someone with PCOS (#362)

0.7.1 (2025-09-19)

Refactor

• :truck: drop_potential_pcos() -> drop_pcos() (#359)

0.7.0 (2025-09-19)

Feat

• ✨ create_inclusion_dates() (#316)

Refactor

• :truck: rename inclusion_ and exclusion_ to keep_ and drop_ (#356)

0.6.3 (2025-09-18)

Refactor

• :recycle: convert logic parsing code into its own function (#350)

0.6.2 (2025-09-18)

Refactor

• 🚚 rename to exclude_pregnancies() (#351)

0.6.1 (2025-09-18)

Refactor

• :recycle: “parse, don’t verify”, select correct columns at start (#345)

0.6.0 (2025-09-18)

Feat

• :sparkles: classify_t1d() (#309)

0.5.0 (2025-09-18)

Feat

• ✨ add_t1d_diagnoses_cols() (#288)

0.4.2 (2025-09-18)

Refactor

• :truck: rename to keep_, not get_ for pregnancy dates (#346)

0.4.1 (2025-09-18)

Refactor

• ♻ move add_insulin_purchases_cols() down to after exclusions (#338)

0.4.0 (2025-09-16)

Feat

• :sparkles: add_insulin_purchases_cols() (#296)

0.3.4 (2025-08-27)

Refactor

• :truck: use _diagnosis for clearer language (#332)

0.3.3 (2025-08-25)

Refactor

• :fire: is_non_insulin_gld_code isn’t used, so removing it (#323)

0.3.2 (2025-08-25)

Refactor

• :fire: remove most has_ variables, they aren’t needed anymore (#324)

0.3.1 (2025-08-25)

Refactor

• :fire: no longer need atc after excluding pregnancy cases (#320)

0.3.0 (2025-08-22)

Feat

• :sparkles: check_data_types() (#304)

0.2.0 (2025-08-22)

Feat

• :sparkles: join_inclusions() added as step (#246)
• :sparkles: include_diabetes_diagnoses() (#258)
• ✨ exclude_pregnancy() (#261)
• :sparkles: first exclusion step (#243)
• :sparkles: add exclude_potential_pcos() (#214)
• :sparkles: add simulate_registers() (#208)
• :sparkles: skeleton of the core diabetes classification (#142)
• ✨ add include_podiatrist_services() with tests (#182)
• ✨ add simulated register data to targets pipeline (#167)
• ✨ update algorithm logic to include regex in parentheses (#159)
• :sparkles: add include_gld_purchases() (#138)
• add join_lpr3 with tests (#125)
• function to join the LPR2 registers (#118)
• add join_lpr3 with tests
• :sparkles: add the function to join the LPR2 registers
• create function to include HbA1c criteria
• add function that join inclusion events
• helper functions to insert data into Markdown vignettes
• internal function to extract logic from algorithm data. Untested
• create (draft) function to include HbA1c criteria
• added simulated data
• :construction: code to get list of Danish ICD 8 codes
• :construction: expand on the fixer functions for the simulation data
• add function to convert to lowercase.
• internal function to get the required variables for a register.
• function to get list of the registers abbreviations
• internal function to verify if the dataset has required variables.
• :sparkles: workflow that assigns PR author as assignee

Fix

• :bug: update algorithm() examples with updated logic names (#302)
• :bug: refer to correct workflow in R-CMD-check badge (#299)
• algorithm: :bug: barnmak logic in podiatrist_services (#278)
• :bug: fixes notes about factors and reading CSV file (#248)
• :bug: rename pregnancy event variable in algorithm by removing lpr2_ (#236)
• 🩹 don’t filter to two first observations in include_hba1c() (#202)
• :bug: this function needs to be exported so Arrow can find it (#193)
• :bug: this function needs to be exported so Arrow can find it
• :memo: commit message auto-generated by document workflow (#183)
• :bug: regex works on more than one =~ (#176)
• :bug: misc fixes to successfully run r-cmd-check (#168)
• 🐛 change honuge format to yyww (#165)
• :bug: verify function actually throws error (#128)
• the logic criteria wasn’t being correctly loaded, need to convert to expression
• :art: remove duplicate arrow from function flow puml
• remove return since it’s not needed
• add “indo” to condition since it’s used in the function
• flip the condition so it’s true when both “atc” and “name” is in data
• typos
• :bug: keep earliest two dates, not samples
• diagnosis -> diagnoses in inclusion function
• add missing header in output example table
• minor text edit to make sentence clearer
• specify that we use the primary diagnosis for classification
• update classify diabetes type flow chart based on feedback from @Aastedet
• describe classifying steps as “filters” with “criteria”
• add oxford comma to header
• :fire: remove old figure experiments
• we don’t have tidyverse as a dependency, it isn’t needed here
• add missing bracket
• apply suggestions from code review
• small fix to the comments and proportion of NPU units inserted
• :art: add space before register_abbrev
• NPU should be 8 long in total, not 8 integers
• add “_status” to the name of the “classify_diabetes” function
• minor text edits and change “classification date” to “inclusion date”
• typo
• remove all mentions of “before index date” since these filters are based on all data
• edits from code suggestions
• move set.seed up
• change pnr to only include 001-100 independent of num_samples
• removed leftover man/ file that triggers check warning
• DuckDB requires colnames(), not names()
• Minor text edits
• remove extra starting column
• convert the variable list into an exported object, not internal
• use the created function for getting the abbrev

Refactor

• :recycle: add (non-)insulin drug indicator variables (#298)
• :recycle: rename verify_required_variables() to check_required_variables() (#306)
• ♻ move column name lowercasing to classify_diabetes() (#295)
• :recycle: criteria -> logic (#294)
• :recycle: move lpr2 primary diagnosis logic to algorithm (#289)
• ♻ update join_lpr3() to be prepare_lpr3() (#273)
• :recycle: make yyww_to_yyyymmdd() internal (#283)
• :recycle: update prepare_lpr2() col names to be _dept (#282)
• :recycle: align variable names in algorithm (#277)
• :recycle: update join_lpr2() to be prepare_lpr2() (#272)
• :recycle: remove verifying from internal functions (#259)
• :recycle: rename cpr column to pnr in join_lpr3() (#245)
• :recycle: moved registers into its own R code (#221)
• :recycle: change “!” to “NOT” (#226)
• :recycle: move algorithm into R file (#204)
• :recycle: add contained_doses variable to output (#187)
• :recycle: remove name and vnr (#185)
• :recycle: make joinable generated padded integers (#170)
• take a dataframe rather than a string name of the register
• these functions should only output data
• :recycle: convert the data-raw/ scripts into functions
• kontakter should come first so cpr is first column, plus use inner join
• keep only earliest two dates, might not work with some databases
• move ICD8 processing into own section and store only relevant info
• completely refactor simulation code since simstudy was not working
• :recycle: rewrote and revised the code to add false metformin
• create function to make pnr, plus other small edits