Releases: rfhb/ctrdata
Releases · rfhb/ctrdata
ctrdata 1.7.1
- fix DBI not needed in Imports anymore (CRAN Note)
- fix potential file name issue in conversion script
- fix dbFindFields() to never return _id (previously depended on database backend)
- changed tests (not on CRAN detection, register availability, additional tests)
ctrdata 1.7.0
- much reduced database backend-specific code, using nodbi 0.4.3 (released 2021-07-23) which also introduces transactions for sqlite using RSQLite >=2.2.4 (released 2021-03-12)
- temporary directory creation only when needed, more automated deletion
- changes in detecting non-functioning register servers
- further streamlined unit testing
ctrdata 1.6.0
- added support for ISRCTN
- refactored checking binaries and caching this info
- updated EUCTR download parameters
- refactored ctrGetQueryUrl and ctrOpenSearchPagesInBrowser
- harmonised error checking
- avoid some errors with external scripts
- refactored url / query mangling, added detailed testing
- refactored storing JSON into database (handle big files, reduce memory)
- improved dbFindIdsUniqueTrials (speed, memory, register coverage)
- factored out conversion to JSON
- accelerated EUCTR results and history download and storage
- external scripts now create multiple chunks of records
- use further identifier fields with dbFindIdsUniqueTrials
ctrdata 1.5.3
- include message how to handle server certificate issues, by propagating user settings for httr to curl operations
- ensure identical return structures when no new trials found
- dfTrials2Long: harmonise identifier level assignment, address cases where field occurs only once in input df
- dfMergeTwoVariablesRelevel: corrected and improved user info
- dfName2Value: remove duplicate rows, e.g. from duplicated criteria
ctrdata 1.5.2
- bugfix of EOL for converting EUCTR files
- remove LazyData from DESCRIPTION
ctrdata 1.5.1
- bugfix for non-matching euctr protocol and result ids: any trials from EUCTR for which results were downloaded with version 1.5.0 should be downloaded again (ctrLoadQueryIntoDb)
- dfTrials2Long refactored and accelerated
- API change: dfTrials2Long return value (identifier replaces main_id and sub_id)
- new option to save EUCTR results PDF files in user-specified directory
ctrdata 1.5.0
- return values of dbGetFieldsIntoDf are now mostly identical whether using src_mongo or src_sqlite, to best ensure portability of analysis code
- permit dots in queries / URLs
- improved handling of queryterm
- renamed ctrGetQueryUrlFromBrowser to ctrGetQueryUrl
- soft deprecated ctrGetQueryUrlFromBrowser
- ensure parallel retrievals from EUCTR
- speed up routines in dbGetFieldsIntoDf
- make dfTrials2Long handle NA better
- improved documentation, clarified examples
- simplified internals for typing fields, start typing results fields
ctrdata 1.4.1
- reset row names on data frames returned by functions
- update curl parameters when accessing EUCTR
ctrdata 1.4.0
- new: easy access to variables with dfTrials2Long() + dfName2Value()
- improved dfMergeTwoVariablesRelevel() to maintain type of data
- revised and simplified vignettes
- deprecated: dfListExtractKey()
- refactored parts of euctr retrieval
- notify user when euctr register server does not permit compression and how long retrieval will take
ctrdata 1.3.2
changes from 1.3.0 onward
- quote system file paths
- fix error in CI tests
- workaround EUCTR certificate issue
- streamline ctrGetQueryUrlFromBrowser()
- better handling of complex fields
- include further tests for query string handling, checking more parameters and return values
- better clean-up after testing
- ctrLoadQueryIntoDb(querytorerun = ...) now looks for the date when the querytorerun was last run, to more often use euctr update options
- switching from travis to github action
- upped coverage of code tested