Since we follow Conventional Commits when writing commit messages, we're able to automatically create formal "releases" of the website based on the commit messages. Releases in the context of websites are simply snapshots in time of the website content. We use Commitizen to be able to automatically create these releases, which uses SemVar as the version numbering scheme.
Because releases are created based on commit messages, we release quite often, sometimes several times in a day. This also means that any individual release will not have many changes within it. Below is a list of the releases we've made so far, along with what was changed within each release.
- 🚚 rename
chunk_info$columnscols to be more descriptive (#289)
- ✨
list_parquet_files()andlist_parquet_datasets()helpers (#272)
- ♻️ change
convert()to outputchunk_infotibble (#286)
- 🚚 rename
use-targets-template()touse-template()(#284)
- 🚚 rename
convert_file()->convert()(#281)
- ✨ helper
get_*()for project IDs and directories (#251)
- 🔥 remove
convert_register()(#275)
- ✨ export originally private
read_parquet_*()functions (#259)
- ♻️ use sas paths as input in targets template (#244)
- 🐛 add
gc()toconvert_register()(#240)
- 🐛 set
sas_pathtarget tomode = "always"to detect new SAS files ininput_dir(#233)
- 🐛 always clean up targets template
output_dirbefore converting to Parquet (#224)
- 🐛 use
pipeline_dirasoutput_pathto avoid macOS temp dir permission errors (#219)
- ♻️ change simdata to
befandlmdb(#218)
- 🔥 not using
list_parquet_files()anymore (#213)
- ♻️ change
use_targets_template()to use dir as input (#193)
- ⚰️ remove
is_same_register()(#187)
- ♻️ add input checks in simulate (#184)
- 🐛 abort
use_targets_template()if file name is not_targets.R(#181)
- ♻️ use
assert_string()to check character scalars (#180) - ♻️ add internal
check_parquet_path(#179) - ♻️ use Tidyverse functions instead of base and stats (#178)
- 🔥 remove
split_paths_by_register()(#177)
- ♻️
convert_register()andconvert_file()(#174)
- ♻️ add internal
read_sas_chunk()(#171)
- ♻️ add internal
create_partition_path()(#170)
- ♻️ update targets template (#164)
- ♻️ update success msg in
convert_to_parquet()(#167)
- ♻️ get register name internally in
convert_to_parquetfrompath(#161)
- ♻️ simplify simulation in single function
simulate_register()(#153)
- ✨ simulate register functions; test clean-up (#146)
- 🚚 align
pathparam name in internal functions (#151)
- 🚚
use_targets()->use_targets_template()(#149)
- 🐛 update
convert_*()param name in template;file_paths->path(#150)
- 🚚
use_targets_pipeline()->use_targets()(#145)
- ♻️ simplify parameter naming to use
path(#140)
- ♻️ longer UUIDs for less chance of duplicates (#141)
- ♻️ side effect functions return invisibly (#131)
- 🔥 remove unused functions (#135)
- 🐛 partitioning with missing year (#132)
- ✨ targets pipeline (#124)
- ♻️ rename parameters in
convert_to_parquet()(#129)
- ♻️ give error when no file is found in
list_*()(#122)
- ♻️ add internal
get_register_names()(#123)
- 🩹 ensure
source_fileis character before Arrow schema creation (#120)
- ♻️ ensure same register
pathsinconvert_to_parquet()(#116)
- ✨ read SAS files in chunks; lowercase column names; align Arrow schemas (#112)
- ✨ add
list_*()functions (#96)
- ✨ add
read_register()(#90)
- ✨ implement
convert_to_parquet()(#84)
- 🔥 remove deprecated functions (#70)
- 🐛no visible binding for global variable (#62)
- ♻️ export get helper functions (#51)
- 🔥 remove icd related functions and vignette (#50)
- 🎉 start of project (copied from DST servers)