BayLum 0.2.1 (2022-02-20)

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 @IMHarly 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 @IMHarly, #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)

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

• create_ThetaMatrix() this function is a helper to create the theta matrix that can be provided as input for the functions AgeS_Computation() and Age_OSLC14()
• create_FolderTemplates() creates the folder and file structures required by Generate_DataFile() and Generate_DataFile_MG() on the user’s hard drive

BayLum 0.1.3 (Release date: 2018-09-14)

New functions

• plot_Ages() this function replaces code used internally by the functions AgeC14_Compution() and AgeS_Computation() to create an age overview plot. Before the user had no control over the appearance of age plot output. This is still the case if the functions AgeC14_Compution(), AgeS_Computation() and Age_OSLC14() are used (although internally only the function plot_Ages() is called), however, both functions now return a new list of class BayLum.list which is recognised by the function plot_Ages() and can be used to plot the typical age plot while providing several arguments for adjusting the plot output.

Bugfixes & changes

• The newly introduced function combine_DataFiles() did not work on R < 3.5.0, which caused an error on CRAN; fixed.
• The function ScatterSamples() is replaced by the function plot_Scatterplots()
• All computation functions left text connections open after they have been exited; fixed.
• The output of AgeC14_Compution() and AgeS_Computation() gain a new class BayLum.list
• Redundant code in AgeC14_Compution() and AgeS_Computation() has been removed
• The argument Nb_chaines was replaced by n.chains wherever it occurred in the package (functions and documentation)
• The function plot_Scatterplots() gained a new argument plot_type, allowing to switch between the default hexbin plot and a function developed for ’BayLum(plot_type = “smoothScatter”`)
• The function plot_Scatterplots() gained a new argument plot_mode to allow single xy plots
• The function plot_Scatterplots() now accepts a two column data.frame as inputbuild

Internals

• Add new package dependency KernSmooth used in plot_Scatterplots()

BayLum 0.1.2 (Release date: 2018-06-22)

New functions

• The function MCMC_plot() was replaced by the newly written function plot_MCMC()
• The function Concat_DataFile() is deprecated and is basically replaced by the function combine_DataFiles()

Bugfixes

• The functions AgeC14_Compution() and Age_OSL14() now returns age results with two digits
• The function AgeS_Computation() now returns the MCMC information from all samples of the PDF output is set to TRUE
• The function Age_OSLC14() never returned numerical data; fixed
• Fix Age_OSLC14() age format, now in ka for C-14 instead of years
• The last example in AgeS_Computation() could not be run; fixed
• C-14 calibration datasets now show data in ‘ka’ and not ‘a’
• Various typo corrections.

Enhancements

• Allow different BIN/BINX-file names as input. The functions Generate_DataFile() and Generate_DataFile_MG() do no longer expecting the name sheme bin.BIN
• The functions Generate_DataFile() and Generate_DataFile_MG() now supports the ... argument to pass further arguments to the function Luminescene::read_BIN2R()
• The functions Generate_DataFile() and Generate_DataFile_MG() gained a new argument verbose to control the terminal output
• The function Age_Computation(),AgeS_Computation(),AgeC14_Computation(), Age_OSLC14() now support the argument quiet and pipe it to rjags to suppress the terminal output
• The function AgeC14_Computation() now returns a warning if outliers are detected.
• The function LT_RegenDose() gained two new arguments nrow and ncol became a little bit more user friendly

Internals

• Set patch level to 0 (for 3.3.0) to account for CRAN note
• Vignette was updated and corrected

BayLum 0.1.1 (Release date: 2017-09-01)

• Reduce required R version to 3.3.2

BayLum 0.1.0 (Release date: 2017-08-22)

• Initial version on CRAN