| NEWS | R Documentation |
merge_RLum.Data.Curve(): Function allows merging of RLum.Data.Curve
objects in different ways (#128).
RLum.Analysis-class:
Show warnings messages when empty object are produced by a get_RLum.Analysis()
request (implemented in get method and show method).
analyse_SAR.CWOSL()
rename argument output.plot to plot,
rename argument output.plot.single to plot.single,
rejection criteria added 'De > max. dose point'. These avoids that values are accepted even the De from the fitting should be considered as invalid.
analyse_IRSAR.RF()
Add enhanced support for ..., e.g. log, xlim, ylim,
rename argument output.plot to plot.
analyse_pIRIRSequence()
Fix an error that causes problems during plotting (Error in plot.new(): figure margins too large),
rename argument output.plot to plot,
rename argument output.plot.single to plot.single.
merge_RLum(): Function extended accordingly with the newly implemented function
merge_RLum.Data.Curve() (#128).
plot_AbanicoPlot():
ticks removed from error bars,
line value output removed,
placement of y- and z-labels set to center of plot,
weights-option for KDE integrated,
warning for linear z-scale removed,
fixed an error that occured in Abanico Plot for a large data range (#188),
function output extended by data points in/out of confidence interval,
histogram option added,
dot plot option added,
KDE bandwidth added to plot output,
KDE bandwidth default changed to "SJ",
rugs coloured according to subsets,
examples section extended.
plot_GrowthCurve():
Thanks to a querry from Rachel Smedly (#194) the De error calculation method has been updated. So far only the Lx/Tx values have been re-sampled using a Monte Carlo simlation approach to estimate the error. In contrast to the Analyst the Ln/Tn value was just considered without any error. This has been changed for all fitting methods and the final De error is calculated now also using the Ln/Tn error. Note: Usually this change results in a larger De error, however this approach is much more realistic,
consequently to the changes made in the error calculation the corresponding De resulting from
the MC simulation is now added to the output object ($De$De.MC),
Ln/Tn point including the error is now added on the plot output as red circle,
fix a bug occasionally occurred for fit.method = "EXP+EXP" and the function broke
subsequently (#159).
readBIN2R():
critical bug fixed: GRAIN and GRAINNUMBER were not imported properely
(thanks to Norber Mercier),
import speed has been improved by at least an order of magnitude, due to a different
internal data handling (using of data.table instead of data.frame). This may become
relevant for hugh BIN-files like they are produced by single grain measurements (#191).
Note: The Risoe.BINfileData-class remained untouched.
writeR2BIN():
minor changes due to the changes in the function readBIN2R(),
confusing warnings during writing are suppressed.
Add new dependency for package data.table (used in the function readBIN2R()).
Amongst a few minor changes with this release we continued the homogenisation process to further unify our function arguments. So far meaningful similar function arguments can be used over all functions in the package to produce similar effects (e.g., na.rm = TRUE/FALSE). For you as package user it means that self-written R scripts using package functions are likely to break. Please check the list of changes below.
We apologise for any inconvenience!
RLum.Data.Image: S4-class for handling image data including. Image
data in slot data are stored as RasterBrick
RLum.Results: homogenised structure for containing the following base items in the
slot data: 'summary', 'args' and 'call'. Where applicable the 'data' slot contains the original input data.
Depending on the function various further slots are added containing additional (intermediate) data and objects,
set_RLum.Data.Image(): Set method for class RLum.Data.Image
get_RLum.Data.Image(): get method for class RLum.Data.Image
further: corresponding show and coercing methods added for class RLum.Data.Image.
analyse_pIRIRSequence(): Function to analyse post-IR IRSL sequence
measurements (#133),
calc_MinDose(): the previous calc_MinDose3() and calc_MinDose4()
functions were merged to the new calc_MinDose() function, which was rewritten from
scratch and now contains both the 3- and 4-parametric
Minimum Age Models (introducing the new argument par).
Detailed changes and bugfixes are listed further below,
calc_MaxDose(): a wrapper function that calls calc_MinDose(),
but transforms the data set similar to the method described in Olley et al. (2006).
Rather than estimating the minimum dose of a De distribution, this function estimates the maximum dose.
calc_SourceDoseRate(): calculating the dose rate of the irradiation source at the
date of measurement based on: source calibration date, source dose rate, dose rate error.
get_Layout(): Helper function to return a list with layout definitions for
more homogenous plotting,
get_RLum(): Wrapper function for various get methods for RLum-objects (#148),
merge_RLum.Analysis(): Function allows merging of RLum.Analysis objects and adding
of objects an RLum.Analysis object consists of (#138),
merge_RLum(): Wrapper function for various merge methods for RLum-objects (#139),
plot_RLum.Data.Image(): Plot function for RLum.Data.Image object.
Function also added to wrapper function plot_RLum,
plot_RLum.Results(): Plot function for RLum.Results objects.
RLum.Results objects created, e.g. by the age model functions may now be passed to plot_RLum.Results()
to create specific plots that previously were generated by the functions themselves.
Function also added to wrapper function plot_RLum,
readSPE2R(): Add support Princenton Instruments *.SPE files
(camera and spectrometer measurements). Function output is a RLum.Data.Spectrum
or RLum.Data.Image object.
ExampleData.RLum.Data.Image: Example data for RLum.Data.Image object
(part of #135).
get_RLum.Analysis(): Add support for inverse record.id selection, e.g. record.id = -1 selects
all objects except the first.
get_RLum.Results():
Add support for data objects for objects
produced by the function calc_OSLLxTxRatio(),
match access behavoir of objects produced by the function
calc_TLLxTxRatio(). Now it works similar to the function calc_OSLLxTxRatio().
Risoe.BINfileData-class: Description updated due to established changes
made for a better file format support.
Risoe.BINfileData2RLum.Analysis(): If the pos value
is invalid now the function does not return an error any longer. Instead
NA is returned.
analyse_IRSAR.RF():
Add ... support for mtext,
argument method added to allow selection between the method "FIT" and
"SLIDE". The latter method has been newly implemented (#143).
analyse_SAR.CWOSL():
Allow 'IRSL' curves as input to support rough data analysis of the post-IR IRSL protocol (#132, Alice Versendaal, #140, Georgina King),
support multiple recuperation measurements,
correct wrong TL curve figure scaling,
argument log moved to the ... argument,
correct wrong regeneration point numbering (#144),
arguments cex, main now supported via the ... argument,
argument mtext.outer added,
output object Formula added to return the fromula of the
growth curve,
output object De.values extended for D01, D02 and information
on the applied fittin,
numerical output extended for information on the choosen signal and background integration limits (#122),
enhance argument output.plot.single for numerical values.
This allows to select each single plot output individually,
details from the argument rejection.criteria were removed
from the argument list to allow a more flexible handling (e.g. changing only
one parameter at time) and to clean up the argument list. However, the argument
work still as excpected (#145),
reformat of rejection criteria output to allow easier further treatment,
support graphical output of rejection criteria and IRSL curve (if exists),
minor bug fixes (e.g., not properly caluculated TL curve resolution) and graphical adjustements.
Analyse_SAR.OSLdata():
Error message implemented for non OSL data sets (#130, Georgina King),
add argument dtype to provide further curve limitation options,
add argument keep.SEL to allow a manual curve selection (advanced users),
minor code polishing.
calc_AliquotSize():
new argument plot (logical): option to (dis-)allow plotting of the Monte Carlo estimation,
argument MC.estimate was renamed to MC,
new structure of the returned RLum.Results object, containing the elements
'summary', 'args', 'call' and 'MC'.
plotting now done by plot_RLum.Results().
calc_CentralDose():
removed argument sample.id,
removed argument print.iterations,
renamed argument input.data to data,
new argument '...', which is checked for arguments verbose (logical) and trace
(logical) to (dis-)allow (extended) console output,
new structure of the returned RLum.Results object, containing the elemennts 'summary', 'data', 'args', 'call' and 'profile',
fixed a bug that occured for non-dispersed data sets and resulted in a virtually infinite loop using up all available memory,
plotting now done by
plot_RLum.Results().
calc_CommonDose():
removed argument sample.id,
renamed argument input.data to data,
new argument '...' (currently not used),
new structure of the returned RLum.Results object, containing the elements 'summary', 'data', 'args' and 'call'.
calc_CosmicDoseRate():
new structure of the returned RLum.Results object, containing the elements 'summary', 'args' and 'call'
calc_FiniteMixture():
removed argument sample.id,
removed argument n.iterations,
removed argument main,
renamed argument input.data to data,
new argument plot option to (dis-)allow plotting when a vector of length > 1 is provided for n.components,
new argument '...', which is checked for arguments verbose (logical) and trace
(logical) to (dis-)allow (extended) console output. main may be used to set the title of the plot,
new structure of the returned RLum.Results object, containing the elements
'summary', 'data', 'args', 'call', 'mle', 'BIC', 'llik', 'grain.probability', 'components' and 'single.comp',
plotting now done by plot_RLum.Results().
calc_FuchsLang2001():
removed argument sample.mtext,
removed argument sample.id,
removed argument output.terminal,
renamed argument sample to data,
renamed argument output.plot to plot,
the argument '...' is checked for verbose (logical) to (dis-)allow console output,
new structure of the returned RLum.Results object, containing the elements 'summary', 'data', 'args', 'call' and 'usedDeValues',
plotting now done by plot_RLum.Results().
calc_HomogeneityTest():
removed argument sample.id,
renamed argument input.data to data,
the argument '...' is checked for verbose (logical) to (dis-)allow console output,
new structure of the returned RLum.Results object, containing the elements 'summary', 'data', 'args' and 'call'.
calc_MaxDose3():
Function renamed to calc_MaxDose(),
this is now just a wrapper function that calls the new calc_MinDose() function,
but transforms the data set similar to the method described in Olley et al. (2006).
For further details please see the changes/bugfixes for the calc_MinDose3() / calc_MinDose4() functions,
plotting now done by plot_RLum.Results().
calc_MinDose3() and calc_MinDose4()
the functions calc_MinDose3() and calc_MinDose4() were merged in the new function calc_MinDose(),
renamed argument input.data to data,
renamed argument output.plot to plot,
removed arguments sample.id, gamma.xlb, gamma.xub, sigma.xlb, sigma.xub,
mu.xlb, mu.xub, init.gamma, init.sigma, init.p0, init.mu, ignore.NA,
calc.ProfileLikelihoods, console.ProfileLikelihoods, console.extendedOutput, output.indices,
new argument par that accepts a numeric value (either 3 or 4) to set whether the 3- or 4-parametric Minimum Age Model should be applied,
new argument bootstrap (logical) to apply the recycled bootstrap approach of Cunningham & Wallinga (2012),
new argument boundaries that accepts a named list containing values for the lower and upper boundaries for all parameters,
new argument init.values that accepts a named list containing the starting values for all parameters,
new argument '...' that, amongst others, checks for debug (logical) and verbose (logical) to (dis-)allow (extended) console
output. Further allowed arguments are bs.M, bs.N, bs.h and sigmab.sd to set various bootstrap parameters,
implemented (recycled) bootstrap method as described in Cunningham & Wallinga (2012),
function now iteratively adjusts boundaries for sigma when the estimator reaches the upper boundary,
function now tries to estimate reasonable starting values for all parameters,
calculation and plotting of the maximum log likelihood functions/profiles are now done using the mle2() function of the 'bbmle' package
the optimisation routine was changed from 'nlminb' to 'optim' using the 'L-BFGS-B' method,
confidence intervals for all parameters are now calculated by the 'confint()' method and exclusively on the 95% confidence level
the asymmetric error on gamma (i.e., the minimum dose) is now exclusively determined on the 95% confidence interval of the profile log likelihood function for gamma
a symmetric standard error for gamma is automatically derived from the 95% confidence interval of the profile log likelihood function by dividing the length of this interval by 3.92 (see Galbraith & Roberts, 2012).
new structure of the returned RLum.Results object, containing the elements 'summary', 'data',
'args', 'call', 'mle', 'BIC', 'confint', 'profile', 'bootstrap',
plotting now done by plot_RLum.Results().
calc_OSLLxTxRatio():
Correct an error that causes NA values
within the curve errors for rare combinations of signal and background
integreation limits,
provide a name for the output data frame LxTx.table for the
RLum.Results object,
error fixed for the marginal case that k = 2 and exceed the maximal curve length. In that case accidentially no error was calculated,
argument sigmab was listed but was so far not implement. Fixed (#166)
calc_OSLLxTxRatio():
support of data without given uncertainties.
calc_Statistics():
support of data with NA or zero errors
calc_TLLxTxRatio():
Lx/Tx ratio wrongly calculated as due to a coding error just the first value of the signal integral was taken for summing up the values and not the range of the values,
Output produced by the function now is a list (data slot in the RLum object), to
equal behavoir of the functions.
fit_CWCurve():
Implement ... argument to replace explicit arguments ylab,
xlab and main,
Fix an issue that causes an error when input values were <= 0 (#123 reported by Johanna Lomax),
Correct wrong y-axis labelling (wrong channel resolution value).
plot_AbanicoPlot():
numeric values allowed for parameter centrality,
KDE label value fixed,
layout functionality added,
not supported summary keywords fixed,
dispersion polygon bug fixed,
dispersion type enriched by userdefined confidence interval,
rug option added,
multiple 2-sigma-bars allowed,
rotate-Option added to Abanico Plot,
minor modifications of layout definitions (#124),
grid.col bug fixed (#149),
error message for data with zero or without errors added (#154),
age models as input for lines supported (#107),
support of par(mfcol())l,
bug in lower z-axis limit eliminated,
example list updated (i.e. fast CAM instead of slow MAM)
bug for small values fixed, rounding to 0 decimals replaced by 3 significant digits,
parameter at to allow user-defined z-tick positions,
individual statistical summary order supported (#184).
plot_DRTResults():
Argument na.exclude renamed to na.rm,
by default ylim scaling now is set to the minimum and maximum extend of the data set,
argument given.dose now allows a vector as input, e.g., if more then one data set
is provided (#156),
argument given.dose are not required any longer (became an optional argument).
With this the plot function can be easily re-used, e.g. for preheat plateau tests,
change behaviour of the argument na.rm. If a data set is provided containing
NA values also the corresponding values in provided by the argument preheat are
cleaned up. This allows a proper usage of the function in combination with grouping option
(e.g., boxplot = TRUE,
add argument par.local to allow to combine output plots in broader plot
frameworks,
minor corrections in error handling and graphical appearance.
plot_Histogram():
Change default y-axis labelling from 'Counts' to 'Frequency',
support of par(mfcol()) (#146),
support of data without and NA-values for error values (#152, #162),
implementation of calc_Stats()-call for extended statistic summary,
support of summary keywords "kurtosis" and "skewness",
example y-lim bug removed (#177),
individual statistical summary order supported (#184).
plot_KDE():
wrong x-axis and y-axis limit calculation fixed (#158),
renaming and extension of summary and summary.pos parameter,
plot now longer shows a density error when no error exists (#160),
De.stats error fixed (#163),
support of par(mfcol()),
support of error-free data extended (now also for calculation of kde max) (#178),
individual statistical summary order supported (#184).
plot_GrowthCurve():
Minor graphical and code polishing,
fix wrong regneration point plotting, if sequence contains more then one zero dose point,
argument na.exclude for the sake consistency replaced by na.rm,
proper error message provided if nothing is left after NA value removal,
overshoot of normal distribution curve in MC simulation plot corrected (#111),
argument output.plotExtended.single added allowing a single
window plot output of all plots,
formula of the fitted function is now returend as R expression
instead of a formula object. The return object remains, however,
Formula,
correct wrong error message and graphical output for cases where no De can be calculated for Ln/Tn values much higher than the highest Lx/Tx value,
fix 'EXP+EXP' fit as is was broken, due to a coding error,
correct for wrong par() resetting,
minor code polishing and adjustment.
plot_RadialPlot():
support of user-defined centrality value,
rug option added,
bug for small values fixed, rounding to 0 decimals replaced by 3 significant digits,
individual statistical summary order supported (#184),
minor fixes for the graphical output.
plot_RLum.Analysis():
Add argument abline to allow inserting similar vertical or horizontal lines
in every plot. The decision to implement this argument in this function instead the function
plot_RLum.Data.Curve() was made because in a single plot modus abline is more
sufficiently used as separat function call,
now reset par parameters suffciently,
add argument combine to allow fast combining of curves in single plot window.
plot_RLum.Data.Curve():
Add argument norm to allow for curve normalisation to
the highest value (#129),
Correct for insufficient information on the channel resolution when data was imported from an XSYG file.
plot_RLum.Data.Spectrum():
Correct an error that caused a chrash when xlim was set
in a way that only one wavelength (colour value) was remained,
correct zebra-colour-pattern for rare cases of xlim settings
using plot.type = "persp" or plot.type = "persp3d",
add argument new plot type "transect" for plotting a transect
along the time/temperature axis for a specific wavelength/channel intervall
(#120, requested by Regina DeWitt),
correct function break for a single column matrix (single frame spectrum)
by automatically resetting plot.type to single
for single frame spectra,
colour rug (rug = TRUE) implemented for plot type "multiple.lines"
(#130),
argument xaxis.energy for wavelength energy conversion for
2D plots added (#141).
readBIN2R():
Integrity checks added to avoid unspecific errors for wrong input formats,
complete format support as now a more throughtful distinction is made between
version 03 and 04 (also made for function writeR2BIN()) (#182),
TAG bit is now used to trigger the SEL element during import.
With this, previous curve selections made with other software are kept
(#181, Norbert Mercier).
readXSYG2R():
If the stimulator is infrared light the recordType will be set to IRSL. This
makes the curve types more consistent with the commonly applied nomenclature,
now the function automatically converts time values to temperature values also
for spectra data (#121); behavious is triggered with the argument recalculate.TL.curves.
Second2Gray():
Minor code and example adjustments in along with creating a new function
calc_SourceDoseRate() (#56).
writeBIN2R():
Complete format support as now a more throughtful distinction is made between
version 03 and 04 (also made for function readBIN2R()) (#182),
SEL will now be kept using the TAG bit, releated to changes made
in function readBIN2R (#181, Norbert Mercier).
Add new dependency for package raster and sp, which is used
for the RLum.Data.Image-class,
dependency to package 'Rserve' removed,
dependcy to package 'zoo' added (used for running means),
dependency to package 'bbmle' added (used for maximum likelihood estimations).