tb_vigibase() now accepts a new argument
rm_suspdup (default TRUE) to automatically remove cases
identified as suspected duplicates (from
SUSPECTEDDUPLICATES.txt) when creating the main tables. You
can disable this behavior by setting
rm_suspdup = FALSE.
tb_vigibase() now supports a new argument
overwrite_existing_tables (default FALSE). If set to FALSE,
the function will skip the creation of any .parquet tables that already
exist, allowing you to resume the process after a crash or interruption
without rebuilding completed tables. Set to TRUE to force rebuilding all
tables. This makes the table-building process more robust and efficient
for large datasets or limited hardware.
vigi_routine() now supports two new arguments:
suspect_only (default FALSE): controls whether only
suspected drug cases are included in IC analysis.d_code_2 (optional): if provided, a single analysis is
performed on cases exposed to both drugs in d_code and
d_code_2. A CLI message is displayed when dual drug
analysis is performed.vigi_routine() now has left/right justification of the
patient label centered on 3 months (90 days) instead of the observed
median. This resolves an issue where the label could be misplaced on the
time to onset plot when the median did not match the visual center of
the graph.(#145)The ‘smq’ argument in get_llt_smq() no longer
requires length 1.(#126)
The example for get_atc_code() now uses explicit
names for atc_sel.(#126)
The Value section of desc_tto() now clarifies the
default descriptive format and how to change it.(#126)
Typos in routine pharmacovigilance vignette.(#126)
vigi_routine() now plots details on drug liability
and rechallenge when available (#119).
desc_rch(), desc_dch() and
desc_outcome() now handle out of memory arrow Table
(#127).
compute_dispro() now handles factors as
x or y arguments, if levels are strictly equal
to 0 and 1 (#134).
Error and warnings all turned into cli syntax.
Gathered redundant checkers internally (#133).
vigi_routine() now works with out of memory arrow
Tables. Additional checkers to vigi_routine() for input
data type.
compute_or_mod() now has an informative error
message if either args estimate or std_er were
not supplied (#123).
get_drecno() now handles non-WHO names matching a
unique DrecNo that has several WHO names, and throws an error if
matching multiple DrecNos (#139).
Adding dependency to glue.
Working on examples per CRAN query (#128).
create_ex_* family mostly aimed at example set up or
internal use (#128).
Addressing CRAN queries on package DESCRIPTION.
Condensing examples and tb_vigibase() default
output.
tb_* family now has a cli style
progress bar
get_llt_soc() output is reworked with
cli features. The function also now correctly supports
non-data.table data.frame.
vigi_routine() fails if there are no drug or adr
cases found in demo_data.
mp replaces mp_short throughout the
package (#110). Users will have to change existing code to replace calls
to mp_short with mp.
# the old way
mp_short <- dt_parquet(path_who, "mp_short")
d_names <- list(drug1 = "paracetamol")
d_drecno <- get_drecno(d_names, mp_short = mp_short)
# the new way
mp <- dt_parquet(path_who, "mp")
d_names <- list(drug1 = "paracetamol")
d_drecno <- get_drecno(d_names, mp = mp)In get_llt_smq(), smq_list and
smq_content replace smq_list_content(#74).
Users will have to change existing code to replace calls to
smq_list_content with smq_list and
smq_content. Example tables smq_list_ and
smq_content_ are added to the package. This change was made
to streamline the process of querying sub-SMQs.
# the old way
smq_list_content <- dt_parquet(path_meddra, "smq_list_content")
smq_llt <-
get_llt_smq(
list(ihd = "Ischaemic heart disease (SMQ)"),
smq_list_content)
# the new way
smq_list <- dt_parquet(path_meddra, "smq_list")
smq_content <- dt_parquet(path_meddra, "smq_content")
smq_llt <-
get_llt_smq(
list(ihd = "Ischaemic heart disease (SMQ)"),
smq_list = smq_list,
smq_content = smq_content)compute_dispro() is the new name of
compute_or_abcd(), to reflect that other disproportionality
measures are available.
Args inspect and show_all of
get_drecno() are replaced by verbose
(#102).
get_drecno() and get_llt_smq() messages
are cleaner and written with the cli package.
New top_n argument added to
screen_adr() (#86).
screen_drug() let you screen most reported drugs in
drug (#103).
get_llt_smq() now queries sub-SMQs and return all
relevant codes (#74).
Order of output columns in compute_dispro() has
changed, to allow easier access to formatted values. Also new arguments
min_n_obs and export_raw_values
(#101).
tb_subset() handles ID code lists as input, instead
of numeric vectors, to make workflow more consistent with other
processes like add_*. (#112)
screen_adr() now counts effects at the case level
(#86).
arg data_type of add_* functions is now
deprecated. Data_type is internally detected.
add_drug() and add_adr() have
informative error if id code list is incorrect (#88).
#90 Introducing vigi_routine() to streamline the process
for pharmacovigilance practitioners. This function is intended for
non-expert users, to produce an Information Component and time to onset
graph.
#94 The luda table no longer exists. It is replaced by
an augmented version of link - built directly from
tb_vigibase(), which is used downstream in place of
luda.
#92 tb_main() and tb_sub() are fusionned
into tb_vigibase().
#93 It is now possible to export raw data composing the
value column of desc_facvar() and
desc_cont(), with argument
export_raw_values.
#83 vignette("descriptive") was translated in
English.
#89 and #87 the get_* family of functions now correctly
handle out of memory tables (by pulling them into memory). The same is
true for extract_tto().
#85 Useless default values to some arguments of
extract_tto() were removed.
#91 tb_* family correctly handles different ways of providing paths, and check for the existence of the files before working on tables.
#78 compute_or_abcd() now correctly handles large amount of reports.
#80 hot fix on windows 1224 error
#75 clearer example for get_llt_soc()
#76 SizeUnit_Lx file in sub, which is encoded in an old format, is now appropriately treated as such.
Failed ubuntu latest oldrel-1 for licensing issue. Put license in license file.
Fixing Windows 1224, by using different temp sub folder for each test
Changing the package’s name!! All set for a public release.
Improved documentation by adding tags, translating to english the basic_workflow vignette. #62, #67, and #69
Increased code coverage back to 100%.
#60 add table builders to arrow format (tb_* functions).
This is supposed to overwrite the current fst and dt_fst process, which
is being deprecated as of this release.
New getting_started vignette shows the process.
dt_parquet function is to replace dt_fst in
the workflow.
Sounds like minor but… Added tests to hit 100% test coverage for the package!
#56 Had R CMD CHECK pass on 3 different OS! Test coverage also works.
#58 Further anonymized datasets.
get_drecno now correctly inspects DrecNos identified from an mpi_list.
Fix a bug, get_drecno wasn’t correctly returning DrecNos from mpi_list.
#55 add soc option to argument term_level
of get_llt_soc.
#15 compute_or_abcd is now internally vectorized on
drugs and adrs.
#53 get_atc_code now appropriately tolowers and trims
names as it states, even when argument vigilyze is set to
FALSE.
#39 added a d_names argument to
add_drug.
Introducing tb_custom to pharmacocaen.
Improved description of add_drug and
add_adr.
#7 was fixed a long time ago. Internal improvement to
get_llt_smq
#54 introduce the desc_outcome function to collect adr
outcome. Help in the descriptive vignette.
Fix add_adr data_type checker.
Patches:
#47 get_drecno now handles correctly non lower-case
names in d_sel.
#43 improved documentation of add_drug
#50 Experimental screen_adr function to identify main
adverse drug reactions in an adr dataset.
#28 documentation of desc_* functions is now available
in vignettes descriptive, template_main and template_dictionary.
Vanquished the
codetools::checkUsagePackage(pack = "pharmacocaen")
note!
#52 fixed incorrect behavior of get_drecno with arg inspect set to TRUE
#51 desc_facvar now shows appropriate big.marks for large numbers
#46 desc_facvar and desc_cont now use a more reliable regex logic to
display results according to format argument.
#44 columns names of desc_dch, desc_tto and
desc_rch now consistently use drug_s and
adr_s (instead of drug and adr
for desc_rch in the past)
#42 improved get_drecno documentation.
#36 and #40 improved add_drug documentation.
#45 removed argument tto_time_range from
desc_rch (formerly, the function also provided time to
onset descriptives)