mrgsolve 1.0.9

• Fix signatures for compiled.mrgmod() and as_tibble.mrgsims() based on new
  R-devel check findings (#1065).

mrgsolve 1.0.8

• SIGMA() is a new model macro which allows users to access on-diagonal
  elements of SIGMA in the model (e.g. SIGMA(2) in $ERROR) (#1051, #1052).

• mrgsim() and mrgsim_q() gain an etasrc argument, allowing ETAs to
  be either simulated from OMEGA (new default and previously the only
  behavior) or taken from the input data set (new option), similar to the way
  parameters can be scraped from the data set (#1037).

• @etas is a new option for use with the $CAPTURE block to let users name
  ETAs to be captured into the simulated output; for example, use
  @etas 1:last to capture all model ETAs in the simulated output (#1055).

• Drop CXX_STD statement from Makevars file and DESCRIPTION to be consistent
  with current changes in R-devel; mrgsolve continues to require compiler
  capable of implementing C++11 standard, but this should be selected
  automatically by R (#1060).

• mrgsolve now depends on R >= 3.6.2 (#1060).

Bugs Fixed

• Fix bug when the path argument is used in the $NMXML or $NMEXT blocks;
  this bug was introduced through the @cppstem feature in version 1.0.7
  (#1046, #1048).

• Fix bug in mread_cache() where the project directory wasn't getting
  rendered properly when passing the complete path to the model specification
  file (#1056).

mrgsolve 1.0.7

• $NMXML and $NMEXT now accept the run argument set to "@cppstem" (i.e
  run = "@cppstem"); in this case, the stem of the NONMEM run will be assumed
  to be the same as the stem of the mrgsolve model file (#1025).

• Missing values (NA) in input data sets will be replaced with 0 for the
  following columns: AMT, CMT, EVID, II, ADDL, RATE, SS as well as
  their lower case counterparts (#1030).

• Refactored include order when building a model; this is an internal update and
  not expected to be visible to the user (#1038).

mrgsolve 1.0.6

Bugs Fixed

• Fix bug in TOUCH_FUNS where parameter and compartment lists were not
  getting generated properly when mrgsolve was not loaded; bug was
  detected and fixed prior to upload to CRAN or MPN (#1013).

mrgsolve 1.0.5

• Changed behavior for dosing records where EVID = 4 and SS != 0
  to match what is done by NONMEM: the system will be advanced to
  steady-state but will not be reset; behavior prior to 1.0.5
  release was to advance to steady-state and then reset (#1011).

• Any column in an input data set that has a class attribute will
  now be dropped in addition to non-numeric columns; this includes
  columns that are integer64 which can be present in data frames
  derived from data.table::fread() or other .csv readers
  (#1008).

• The $MAIN and $TABLE blocks will no longer get called for
  actual dose administration records with lag times; this could
  change time after dose calculations or other calculations that
  could be happening in those blocks when the actual administration
  is taking place (#992).

• Code to update the parameter list was re-factored to be much more
  efficient (#978).

Bugs Fixed

• Fix bug where apparent dosing events for additional doses with lag
  times were not getting scheduled (only records for the actual
  administration); this doesn't change the simulated output but could
  change time after dose calculation (#992).

• Fix bug where special nm-vars variables were not getting recognized
  as valid capture items during dynamic capture (#987).

• Fix bug when updating the parameter list with a data frame that included
  non-numeric columns that were not parameters (#978).

mrgsolve 1.0.4

• Fix bug where as_data_frame() was not properly working when leading event
  object was evd type (#948, #955).

• Add uctran() to convert nmtran data names to upper case (#955).

• Both lctran() and uctran() are generic and work on data.frame or
  event (ev) objects (#949, #955).

• Fix bug where data records with EVID = 3 were getting shifted in time by
  ALAG (#964, #969).

• Small negative eigenvalues from OMEGA or SIGMA are set to zero in
  multivariate normal simulation of ETA and EPS, following the pattern
  seen in MASS::mvrnorm() (#956, #957).

• Fixed bug where template parameters (T) were getting modified when nm-vars
  plugin was used in conjunction with mrgx plugin (#965, #968).

mrgsolve 1.0.3

• Removed assert() statement in LSODA code found by CRAN check (#943).

• Test class using inherits() not class() from CRAN check (#943).

These changes were made in two versions (1.0.2 and 1.0.3) after iteration with CRAN checks.

mrgsolve 1.0.1

• Add LOG(), EXP(), SQRT() macros when nm-vars plugin is invoked
  #931, #936.

• Use evd() to create an event object which renders nmtran names
  in upper case (e.g. TIME rather than time) #935, #919.

• Fixed bug where rate was not getting set for modeled events #934.

• Fixed bug where self.stop_id() and self.stop_id_cf() had reversed
  behavior as documented #927, #928.

• Refactored EVID=3 behavior to leave NEWIND as-is #934.

mrgsolve 1.0.0

• New model syntax: THETA(n) is interpreted as THETAn in the model code;
  THETA is now a reserved word (#837, #891, #892).

• New functions collapse_omega(), collapse_sigma() and collapse_matrix()
  added which help to reshape OMEGA and SIGMA matrix objects with multiple
  blocks (#897, #900).

• New plugin nm-vars which implements NONMEM-like syntax for referring to
  compartments, differential equations, bioavailability factor, infusion
  duration and rate, and lag time (#904, #891).

• New plugin autodec which will find user-defined variables in the model
  code and automatically declare them as type double (#905, #893).

• Use ii as spacer for setting the between-dose interval when putting event
  objects in a sequence (c.f. wait) (#906, #901).

• Start deprecatingsimeta(n) and simeps(n), calls to simeta() or
  simeps() with an integer argument thus limiting the update to just a single
  ETA(n) or EPS(n); use in model code will generate warning when loading
  the model that contains the n integer value (#908, #909).

release/0.11.2

mrgsolve 0.11.2

• In pk2iv, change scaling volume for CENT from V2 (incorrect) to
  V1 (#831, #832, #833)

• Fix bug collating multiple $OMEGA or $SIGMA matrices when parsing a
  model (#863)

• Refactor how debugging information is processed when using the recover
  argument to mread() (#853)

• Fix typo in documentation for as_cmat(); the off-diagonals are assumed
  to contain correlations (#856)

• Wrote a as.data.frame.matrix() function in Rcpp; all simulation results
  return from the C++ simulation code as a data frame (#857)

• Fix bug where dynamic capture (via mread()) was not allowed for variables
  declared in $GLOBAL (#868)