Skip to contents

Changelog

Source: NEWS.md

BayLum 0.3.3.9000-13

This version of BayLum requires R >= 4.3!

Miscellaneous

  • Fix various anchors in the manual with broken target (CRAN request).

Internals

  • The vignette manual code takes very long to execute and must be pre-computed. To pre-compute the vignette: make changes in BayLum.Rmd.orig, then run vignettes/precompile.R locally.

BayLum 0.3.2 (2024-04-14)

CRAN release: 2024-06-16

This version of BayLum requires R >= 4.1.x!

New functions

  • create_DataFile(): this function attempts to ease the very pain-stacking process of input data processing (BIN/BINX files) so that it can be used in the BayLum modelling. The function is an entirely new development and will replace the functions Generate_DataFile_MG() and Generate_DataFile() in future versions of BayLum. Key features:
    • Only one function for single-grain and multiple-grain data.
    • Support for BIN/BINX and XSYG files and RLum.Analysis-class objects, i.e. the data can be imported independently of the data file creation.
    • A new single-file configuration based on the YAML format. This makes the old error-prone folder structure obsolete.
    • Various internal consistency checks to avoid input data are erroneousness; which may lead to very hard to-track-down error messages.
    • Carries information about the number of samples and the sample names.
  • plot_RegDosePoints(): the function will replace the deprecated function LT_RegenDose(), which was too error-prone and complicated to maintain.
  • write_YAMLConfigFile(): This function supplements create_DataFile() supporting auto-generated YAML config template files.

Bugfixes and changes

Miscellaneous

  • Drop dependency to deprecated package 'ArchaeoPhases' and make corresponding code an internal function (PR #33, @nfrerebeau).
  • Add dependencies to the packages 'yaml' and 'cli'.
  • Bump version requirement for 'Luminescence' to >= v0.9.22.
  • Add new external files example.yml and yaml_config_reference.yml.
  • Internal datasets DATA1, DATA2, DATA3 were updated to be consistent with the new function create_DataFile().

Deprecated functions

The following functions are deprecated and will be removed in future versions of BayLum:

BayLum 0.3.1 (2023-03-28)

CRAN release: 2023-04-13

This version of BayLum requires R >= 4.1.x!

Bugfixes and changes

  • AgeS_Computation() & Age_OSLC14() had two instances of the MCMC-list. One has been removed. This essentially halves the amount of memory needed to store the output of these functions. (Thanks to @guiguerin for reporting, #27).
  • AgeS_Computation() crashed reproducibly, if the THETA matrix (as one possibility) was provided indeed as a file name instead of a matrix object; fixed (thanks to M. Heydari for reporting).
  • AgeS_Computation() the function handles wrong user input for the Theta matrix more friendly by converting it automatically to a matrix if needed.
  • create_ThetaMatrix() the function did not throw an error in all cases where sigma_s was obviously wrong; fixed (thanks to M. Heydari for reporting).
  • Generate_DataFile_MG() the function now checks that the number of specified BIN/BINX-files matches the actual number of BIN/BINX-files present in the folders. If not, a warning is prompted. This prevents hard-to-track-down error messages and odd modelling results that erroneously encourage searching for problems with JAGS or the BayLum models.
  • plot_Ages() gained a couple of new arguments. First, plot_mode allows to switch between the standard plot output "ages" (the default) and the new option "density". The latter draws probability density curves instead of lines. Further arguments specific to the plot_mode = "density" are d_scale to scale the density plots in height and show_ages to show the Bayes estimator. More generally, legend.cex was added to enable a convenient scaling of the legend fonts; this works for both plot modes.
  • plot_MCMC() gained additional support for BayLum.list returned by Age_OSLC14().
  • plot_Scatterplots() The functions can now handle BayLum.list objects if returned by certain functions. Currently: AgesS_Computation(), Age_OSLC14().

BayLum 0.3.0 (2022-10-28)

This version of BayLum requires R >= 4.1.x!

New functions

  • write_BayLumFiles(): @imgoej contributed an excellent new function that creates all folders and files needed for BayLum, making it a lot easier and less error-prone to work with BayLum.

Bugfixes and changes

  • AgeS_Computation(): the function now supports parallel processing using the package 'rungjags' (thanks to @imgoej)
  • Age_OSLC14():
    • Regression error, the function tried to observe more variables than were actually in the model; fixed. (thanks to @imgoej, #21).
    • The function now supports parallel processing using the package 'rungjags' (thanks to @imgoej).

Internals

  • A few fixes regarding string comparisons in if() conditions.
  • Fix CRAN warning regarding Escaped LaTeX specials: \&.

BayLum 0.2.1 (2022-02-20)

CRAN release: 2022-02-21

Note: This package version needs at least R >= 4.0.

Bugfixes & changes

  • AgeS_Computation():

    • The function crashed if jags_method = "rjparallel" with a runtime error due to missing information about the RNG type; fixed (thanks to @imgoej for spotting).
    • If jags_method = "rjparallel" the inits were preset to a specific value for the random number generator. While this setting led to a crash of the function and probably caused no harm, the setting did not make any sense. Now, the random number generator settings (seed, type etc.) can be provided via the argument inits, which is passed to runjags::autorun.jags. The default is NA to maintain the expected behaviour if jags_method = "rjparallel" is used.
    • The function gained a new argument model to inject custom models to run with the parameters as already defined for the model. This option allows to adjust priors and likelihoods to test different scenarios. Please not that this option works without a safety net, mean the function does not check whether the parameters in the model are actually available.
    • The function now returns the used model as a character string, in combination with the new parameter model it can be used to modify the model on the fly.

  • Age_OSLC14(): add support for "rjparallel" via new argument jags_method = "rjparallel", like in AgeS_Computation() (thanks @imgoej, #20).

  • Generate_DataFile_MG():

    • The function gained a new argument force_run1_at_a_time. If set to TRUE it automatically orders the curves in the BIN/BINX-file following the “Run 1 at a time” option. Regardless of whether this option was used for the measurement or not. Please note: this cannot work if you have assigned a position number more than one time to different aliquots.
    • verbose = FALSE was not respected under all circumstances; fixed.

BayLum 0.2.0 (2020-12-06)

CRAN release: 2020-12-06

Note: This package version needs at least R >= 3.5.0.

Bugfixes & changes

  • AgeS_Computation():
    • Now stops if a mismatch was encountered between the number of samples and the matrix provided via THETA to prevent a JAGS runtime error.
    • The function experimentally gained a new argument jags_method to support full automated processing in parallel using the R package 'runjags'; including a ... argument for a more detailed control. By default the calculation is done simply via 'rjags' as usual.
  • Age_OSLC14():
    • Add additional output to enhance output table by additional information (#11).
    • The CSV-file option now returns the age and the HPD intervals
    • The bugs model for the gaussian distribution did not work due to a wrongly set index (#13); fixed.
    • If a small number of samples was used and the first was a C-14 sample JAGS crashed; now it throws are more comprehensible error message (#14).
  • Generate_DataFile_MG(): now works only on preselected records; the limitation of OSL only was removed.
  • plot_Ages():
    • The sorting of the samples in the graph did not work as expected; along with this a new argument sample_order was added.
    • The function accepts a data.frame as alternative input, if the data.frame is similar the data frames created by the computation functions (e.g., AgeS_Computation()). This option enables the manual creation of age plots from self-compiled data.
  • plot_MCMC():
    • Accidentally the function could not display the chains from more than 9 samples; fixed (thanks to Maryam Heydari).
    • Now the function understand the output of AgeS_Computation() for an automated treatment.
    • Wrong variable names now lead to a warning instead of a stop. This is less aggressive and it makes the function easier to use by users who only use the plot functionality of BayLum.
  • plot_Scatterplots():
    • Line breaks and leading and trailing white space is now removed from the character (vector) provided via sample_names
    • If the number of selected samples via sample_selection was shorter than the number of names provided by sample_names the function returned an subscript out of bounds error; fixed (reported by Maryam Heydari).

Internals

  • All models now come with an example showing how they can be inspected (before the examples not working).
  • Fix a couple of documentation glitches (e.g., part of the documentation was missing).
  • Smooth the creation of BayLum.list objects using a central internal function.
  • Add new C-14 calibration data and unify the names of those datasets. Allowed inputs are "IntCal13","IntCal20","Marine13","Marine20", "SHCal13", "SHCal20" or an own dataset which must be stored in a CSV file.

New functions

BayLum 0.1.3

CRAN release: 2018-09-19

BayLum 0.1.2

CRAN release: 2018-06-22

BayLum 0.1.1

CRAN release: 2017-09-07

BayLum 0.1.0

CRAN release: 2017-09-01

  • Initial CRAN submission.