version 2.2.7 (2024-10-23)

New features and changed behaviour
    o none
    
Other
    o include package anchors in doc links

version 2.2.6 (2023-11-28)

New features and changed behaviour
    o static userdist option; linear mask
    o details may include stepmax argument passed to nlm
    o openCR.fit() acquires details argument 'agebreaks'
    o collate method for 'openCR' objects

Other
    o makeNewData.openCR() rejects covariates with reserved names
    o agebreaks largely replaces agecov for defining age groups
    o openCR.design() acquires argument 'agebreaks', saves agelevels
    o age.matrix() acquires argument 'unborn' (code for age<0)
    o requires secr >= 4.6.1
    o explicit error for integer overflow with many parameter levels

version 2.2.5 (2022-09-25)

New features and changed behaviour
    o ucare.cjs() restored

Bugs fixed
    o modelAverage.openCR() did not pass arguments to modelAverage.openCRlist()

version 2.2.4 (2022-07-02)

New features and changed behaviour
    o ucare.cjs() temporarily removed (requires R2ucare)

Other
    o vignette - new references and tweaked mention of ucare.cjs ()

version 2.2.3 (2022-07-01)

New features and changed behaviour
    o none

Bugs fixed
    o predict() failed with parameters b & h2 when number of sessions differed 
      among strata
    o derived() not working with stratified model
    o bad names in openCRlist output by par.openCR.fit()
    o dummy variable coding failed with R>4.0

Other
    o fix accidental inclusion of GitHub README
    o depends secr >=4.5.0
    o does not expect R2ucare to be on CRAN

version 2.2.2 (2022-01-17)

New features and changed behaviour
    o modelAverage() method for openCR models
    o openCR() new parameter 'settle' for weighted settlement
    o cumMove() allows settlement to be weighted by 0-1 mask covariate

Bugs fixed
    o expected.d() failed with user-defined movement model

Other
    o Git repository synched with GitHub
    o exampledata folder under inst with some previously fitted models 
      (used in Examples e.g. ?make.kernel)

version 2.2.1 (2021-11-09)

New features and changed behaviour
    o none

Bugs fixed
    o openCR.fit() did not correctly fit non-spatial learned response (b) models 
      such as model = p~b, type = 'JSSAfCL' (bug undetected for some years).

Other
    o tidied C++ code, edited vignette

version 2.2.0 (2021-11-03)

New features and changed behaviour
    o openCR.fit() new defaults: kernelradius = 30, sparsekernel = TRUE
    o plot.kernel() accepts meshcol argument
    o make.kernel() acquires argument 'session' for model input
    o expected.d() expected movement from movement parameters, fitted model etc.
    o summary.kernel() expected distances use expected.d() and may differ
    o matchscale() acquires argument 'expected' to match expected distance
    o 'BVN2' movement model (equal mixture two BVN)
    o bivariate Cauchy movement model 'BVC' (BVT with fixed move.b = 0.5)
    o log-sech movement model 'RDLS' (partial implementation)
    o extensions to plot.kernel

Bugs fixed
    o openCR.fit() failed with non-binary detector usage
    o domain errors in inputs to Boost statistical distributions could crash R
    o make.table() could fail with models fitted before 2.1.0

Other
    o new aliases for movementmodel 'RDE', RDG','RDL','RDEzi','IND','INDzi'
    o RDL movementmodel no longer blocked (Boost call bypassed if scale infinite)
    o openCR.fit() more robust to extreme parameter values (spatial models)
    o more documentation of movement models
    o improved weighting of sparse kernels; details argument 'r0' for 
      effective radius of zero cell

version 2.1.0 (2021-08-14)

New features and changed behaviour
    o new movement kernels based on polar distance 
    o zero-inflated option for some movement kernels and 'uncorrelated' movement
    o cumMove() compounds movement kernel and applies edge effect
    o proportionInPolygon() sums probability within arbitrary polygon(s)
    o optional link 'log1' defined as log(x-1) (constrains x>1)
    o plot.kernel() uses par()$cex.main for title font
    o openCR.fit() optional dummy variable coding of factor predictors
    
Bugs fixed
    o openCR.esa() failed and hence derived() failed for spatial models 
    o openCR.make.newdata() failed if model from version <2.0.0 (details$minimumage not specified)
    o m.array() mis-tabulated counts when none caught on some occasions (no flow-on)
    o openCR.fit() version 2.0 halved variance for types 'Pradel' and 'Pradelg' 
    
Other
    o 'BVN','BVE','BVT' aliases for movementmodels 'normal','exponential','t2D'
    o openCR.fit() default move.b for BVT changed to 1.5 (from 0.6)
    o openCR.fit() details argument 'control' passes arguments to nlm()
    o openCR.fit() default iterlim for nlm() raised to 300 
    o makeNewData() S3 method replaces openCR.make.newdata()
    o C++ code uses Boost library for statistical distributions, avoiding R API
    o kernel code simplified (one C++ routine) and tested
    o long operations more responsive to interruption
    o plot.openCR() definition of 'turnover parameters' for intermediate.x
      extended to include move.a and move.b
    o number of threads explicit in calls to RcppParallel parallelFor
    o minor changes to make.kernel() and summary.kernel()
    o c and '[' methods for openCRlist class; separate documentation of openCRlist()
    o "anchored" movement (experimental)

version 2.0.2 (2021-05-14)

New features and changed behaviour
    o make.table() acquires argument 'collapse' for stratified models

Bugs fixed
    o C++ sqrt() ambiguity on r-patched-solaris-x86 (use std::sqrt())
    o clang-UBSAN overflow error in tests (read beyond array in C++ getmoveargs)
    o erroneous likelihood when dimensions differed among strata
    o openCR.make.newdata() omitted session covariates of stratified model
    o sumsims() failed with 'non-numeric argument to binary operator'

Other
    o multithreaded Examples now in dontrun{} to avoid problematic TBB library
    
version 2.0.1 (2021-05-09)

New features and changed behaviour
    o none

Other
    o run tests with Sys.setenv(RCPP_PARALLEL_BACKEND = "tinythread") 
      to avoid CRAN errors
    o added URL to DESCRIPTION
    o onLoad() allows for possibility of single CPU (previous minimum 2)

version 2.0.0 (2021-05-04)

New features and changed behaviour
    o stratified models
    o generic method strata() for retrieving and assigning stratum names
    o stratify() function to arrange data for stratified model
    o openCR.fit() sparse movement kernels (new argument 'sparsekernel')
    o openCR.fit() kernelradius elevated to full argument from 'details'
    o make.kernel() saves empirical cumulative distribution function as attribute
    o make.kernel() accepts fitted model as input
    o pkernel() computes cumulative probability distribution of distance
    o qkernel() computes quantiles of distance distribution
    o dkernel() computes probability density of distance distribution
    o matchscale() kernel scale parameter giving required quantile of distance
    o plot.kernel() acquires argument 'title'
    o new movementmodel "annular" in openCR.fit(), make.kernel() etc.
    o HPX new detection function for binary detection
    
Bugs fixed
    o age miscoded in openCR.make.newdata(), hence bad prediction for age models
    o summary.kernel() proportion truncated misreported for exponential movement
    o LLsurface.openCR() failed

Other
    o vignette drops call for multiple cores
    o Age2 included in openCR.make.newdata()
    o plotKernel() removed (superceded by plot method for kernel objects)
    o 'openCR' object returned by openCR.fit includes kernel coordinates
    o optional logmultinomial term precalculated and not affected by covariates
    o fixed doi format in a citation
    o avoid ':::' (use local versions of unexported 'secr' functions)
    o Imports RcppParallel version >= 5.1.1 to dodge problems in earlier releases

version 1.5.0 (2020-12-13)

New features and changed behaviour
    o number of threads managed via RCPP_PARALLEL_NUM_THREADS as in secr
    o PLB... aliases for JSSA...CL type names
    o openCR.fit() new argument 'edgemethod' ("truncate", "wrap", "none")
    o par.openCR.fit() modified to allow each fit to multithread
    o openCR.fit() and openCR.design() new argument 'agecov' for grouping ages
    o openCR.fit() and openCR.design() new argument 'minimumage'
    o openCR.esa() now returns a list when bysession = TRUE
    o openCR.esa() unsqueezes result (allows for histories with freq>1)

Bugs fixed
    o movementmodel "user" failed (typo in openCR.fit line 80)
    o secr LL component 3 miscalculated when some capthist repeated (biased D-hat)
    o movement models bias in D-hat due to edge effect
    o movement models failed when R = TRUE (dim(PIA) error)
    o derived.openCR() mistakenly report kappa for spatial models
    o some factor covariates mishandled by openCR.make.newdata() and hence by 
      predict, derived

Other
    o more robust user-defined kernel 
    o calcpdotbd details argument of openCR.fit defunct
    o correct description of runsim.spatial() intervals argument
    o par.openCR.fit() ncores>1 is deprecated (slower than ncores=1!)
    o par.openCR.fit() logfile defaults to automatically generated tempfile
    o Suggests testthat, adds some unit tests

version 1.4.4 (2020-09-01)

New features and changed behaviour
    o none

Other
    o suggests rmarkdown
    o tweak url in vignette

version 1.4.2 (2020-09-01)

New features and changed behaviour
    o reported rank takes account of possible negative eigenvalues (eigH)

Other
    o some data.frame() calls specify stringsAsFactors = TRUE for backwards 
      compatibility in R >= 4.0
    o corrected return without ()
    o terminal '/' on URLs in documentation as required
    o updated references

version 1.4.1 (2019-06-29)

New features and changed behaviour
    o sumsims() dropifnoSE condition now also rejects zero SE
    o print.summary.openCR() drops attributes eigH and fit

Bugs fixed
    o multiple problems in naive CJSsecr model (CJSp1 FALSE, the default)
    o summary.openCR table of counts, format error with non-spatial model

Other
    o details switch 'calcpdotbd' defaults to FALSE unless required for movement model; 
      may affect estimates with JSSAsecr models
    o minor improvements to robustness when details$R = TRUE etc.

version 1.4.0 (2019-06-15)

New features and changed behaviour
    o new 2-parameter movement kernel t2D for 2-D t-distribution 
    o new movement kernel uniform (no parameters, depends kernelradius)
    o new function make.kernel() and kernel methods plot and summary
    o plotKernel() deprecated
    o runsims.spatial() more robust with non-standard extractfn
    o see also 1.3.6

Bugs fixed
    o biased estimates from movement models

Other
    o removed redundant C++ code for single-threaded execution
    o default details$autoini changed to NULL, meaning 'use all'
    o classMembership should now work with movement models

version 1.3.6 (2019-04-13; not published)

New features and changed behaviour
    o plot.openCR() invisibly returns the x coordinates of plotted points; dots argument 
      also controls axes, line width and colour of CI
    o new plot method for output from derived.openCR
    o openCR.fit checks whether autoini required (incomplete or no start provided)
    o runsim.nonspatial() and runsim.spatial() return NULL for failed replicates rather than crashing

Bugs fixed
    o derived method for openCR SECR models wrongly adjusted per capita recruitment 
      when intervals varied, with consequences for derived D-hat
    o JSSAsecrXXX models gave erroneous estimates, particularly of lambda0 and sigma, 
      for count data with binomN = 0 (Poisson counts) 
    o predict.openCR failed when one parameter estimated (others fixed)

version 1.3.5 (2019-03-10)

New features and changed behaviour
    o none

Bugs fixed
    o vignette rebuild failed on r-devel-linux-x86_64-debian-gcc (double \vec)

version 1.3.4 (2019-02-05)

Bugs fixed
    o bad keyword in classMembership.Rd

Other
    o removed unnecessary pdf from inst/doc
    
version 1.3.3 (2019-02-04)

New features and changed behaviour
    o enabled detector-level covariates
    o summary method compatible with secr 3.2.0 (moves)
    
Bugs fixed
    o openCR.make.newdata() and dependent methods (print, predict, summary etc.) 
      failed with learned responses (~b etc.) 

version 1.3.2 (2018-12-09)

New features and changed behaviour
    o none
    
Other
    o conditional use of R2ucare in Examples, vignette

version 1.3.1 (2018-12-08)

New features and changed behaviour
    o sumsims accepts simulated fits with method = "none"

Other
    o R2ucare now in Suggests not Imports, with tweak to ucare.cjs() 

version 1.3.0 (2018-11-26)

New features and changed behaviour
    o canned predictors for learned responses (b, bk etc.) redefined and extended
    o AIC method acquires argument n to override the default effective sample size
    o openCR.fit() no longer computes 'posterior' component for mixture models
    o classMembership method for finite mixture models ('posterior' component removed)
    o specifying native R routines for the likelihood forces ncores = 1
    o print and AIC methods for summary.openCR
    o openCR.fit() acquires argument dframe0 for pesky custom models with learned response
    o par.openCR.fit() default trace = FALSE 
    o rev method for reversing primary sessions of capthist objects

Bugs fixed
    o varying usage unreliable with 'multi' detector type in some circumstances
    o summary method for openCR objects: erroneous counts in capthist component (ignored freq)
    o finite mixture model posterior class membership crashed openCR.fit() for spatial models
    o finite mixture models failed for 'multi' detector type

Other
    o help page for openCR.design corrected

version 1.2.1 (2018-10-31)

New features and changed behaviour
    o runsim.spatial() acquires intervals argument
    o summary method for openCR objects 
    o sumsims() accepts output from summary.openCR
    o sumsims() acquires argument 'true' and adds summaries of bias, coverage etc.
    o CJSp1 = TRUE details argument respected by openCR.design etc.

Bugs fixed
    o sumsims() failed with parameters superN and superD
    o derived method for openCR objects: erroneous f and lambda for some models when
      intervals not 1.0
    o print.derivedopenCR did not apply Dscale to session-specific D

Other
    o changed default starting value for superD
    o tweaks to print method for derivedopenCR class
    o not published on CRAN

version 1.2.0 (2018-05-25)

New features and changed behaviour
    o  multithreaded execution of some functions (up to 4x speed gain)
    o  openCR.fit() types 'secrCL' and 'secrD' emulate closed-population analyses of 'secr'
    o  openCR.fit() trace may be a positive integer (frequency of reporting)
    o  ucare.cjs() wrapper for R2ucare test functions
    o  moving.fit() and extractFocal() functions for moving window of sessions
    o  m.array() reports releases on final session (last.session = TRUE)
    o  openCR.fit() details ignoreusage option
    o  runsim.spatial() and runsim.nonspatial() ncores argument defaults to NULL for multithreading
    o  cyclic.fit() exposed as internal function but not exported
    o  derived() method for openCR objects acquires arguments newdata and all.levels

Bugs fixed
    o usage of count detectors not properly accounted for (noted by Richard Glennie)
    o fixedbeta clashed with start argument (some options)
    o session covariates not matched correctly with robust design sessions
    o LLonly reported x -1
    o details$hessian = FALSE failed

Other
    o Links to RcppParallel
    o Additional check for nonidentifiable parameters in print.openCR() 

version 1.1.2 (2018-04-16)

New features and changed behaviour
    o  none
    
Other
    o  tweaks to C++ code for CRAN (fabs -> abs)
    o  check for availability of RMark
    o  better failure criterion in bd.array()
    o  http changed to https where possible

version 1.1.1 (2018-04-11)

New features and changed behaviour
    o  movement kernel exposed
    o  bd.array() probability of (b,d)
    o  improved simulation functions
    o  start may be named numeric vector
    
Other
    o  first CRAN submission
    o  consolidates CJS and JSSA C code (e.g. prwisecrcpp)
    o  timevaryingcov<- disabled until next release secr

version 1.0.2 (2018-03-26)

New features and changed behaviour
    o  movement kernel unsuppressed

version 1.0.1 (2018-03-25)

New features and changed behaviour
    o  movement kernel [temporarily suppressed]
    o  plotKernel()
    o  movementmodel argument replaces staticHR
    o  experimental timevarying individual covariates

version 1.0.0 (2018-03-11)

Known bugs
    o  CI of predicted b[1]

Other
    o  renumbered 1.0.0 (previously unreleased 2.2.3)