base graphics,
ggplot2 or plotly graphics:
plotScoresplotScreeBoth of these functions are actually in ChemoSpecUtils
and called on behalf of the user when needed.
ggplot2 graphics output are generally similar in
layout and spirit to the base graphics output, but
significant improvements have been made in labeling data points using
the ggrepel package. And of course the ggplot2
graphics can be modified after creation through the usual
mechanisms.plotly graphics output are interactive plots which
can be used for better understanding the data. plotly
provides tools such as zoom, interactive labels and custom positioning
which are very useful.options(ChemoSpecGraphics = "option").?GraphicsOptions.leg.loc argument now allow much more
flexibility with regard to how the legend is positioned. See the
documentation.... in sumSpectra
corrected to show how to pass tol to
check4Gaps.Col7 a palette of 7 colorblind-friendly
colors added to files2Spectra2DObject documentation.R 4.1 RCpkgdown site) is now automated via
CI.hats_alignSpectra2D has been
corrected. Only the reporting was foobar’ed. The alignment worked fine.
Defaults for plots also improved.Btotxt added to import2DSpectra,
allowing import of 2D data exported by the Bruker command “totxt”.dx added to import2DSpectra for use
with JCAMP-DX files, via package readJDX which has recently
learned how to import 2D NMR data sets.F1F2RI-F1decF2dec2 added to
import2DSpectra which handles the import of JEOL spectra
exported as “generic ascii”.files2Spectra2DObject gains a new argument
allowSloppy. This will allow one to import data sets that
do not have the same dimensions. The intent here is to deal with data
sets where the number of points in each dimension is similar but not
identical. This is an experimental feature, and additional functions
will be needed to handle this kind of data. See the documentation for
details.files2Spectra2DObject gains a progress bar and will now
accept a path and other arguments to list.files, bringing
it in line with ChemoSpec::files2SpectraObject.normSpectra2D gets a new method to scale
spectra on [-1 … 1].hats_alignSpectra2D gains new arguments
dist_method and maximize which allows the user
to pass their choice of distance measure through to the objective
function used to evaluate the overlap of the spectra. This greatly
improves the quality of the alignment. See the documentation for
additional details.computeVolume added to aid in normalizing
spectra to particular chemical shift regions, which are volumes when the
intensity is taken into account.LofL and
LofC, added to assist with overlaying multiple spectra in
plotSpectra2D. LofL = “List of Levels” and
LofC = “List of Colors.”colorspace package. The color-handling
infrastructure was also changed to allow easy introduction of different
color schemes in the future, though the user cannot yet make changes on
the fly.inspectLvls simplified and arguments changed; one can
now inspect just a single spectrum or a range of spectra.calcLvls rebuilt to be more consistent and
logical. Values will change slightly from previous values.plotSpectra2D when user specifies
xlim and/or ylim is much more robust. Function
.computeTicks was removed as it was no longer needed.plotSlice similarly made more
robust.import2Dspectra cleaned up
(documentation and code).pfacSpectra2D now allows control of the number of cores
in use when using parallel processing. This is to avoid multiple
processes on the same shared machine each trying to use all the cores
for themselves. Per suggestion by Henrik Bengtsson on Twitter.inspectLvls was not playing nice when argument
which was a vector..mapColors in which
NA could be returned as a color, which results in an
unexpected black contour. Now removes NA and gives a
warning that the requested levels are beyond the range of the data.updateGroups which has been in
ChemoSpecUtils for a while but effectively hidden from
users of ChemoSpec2D.roxygen2 warnings.tinytest.robustbase from suggests (not needed)..makeArray gains a unit test..rescale rebuilt to be more flexible.plotScores was fixed. This was a long standing bug that
somehow escaped notice, dating to the early days of
ChemoSpec. Note that plotScores is actually in
ChemoSpecUtils but is called from ChemoSpec2D,
affecting the results here.class(mia) were plotting the
eigenvalues instead of the percent variance explained.class(mia) was incorrect and gave a rotated version of the
correct result.hats_alignSpectra2D was
added, along with a number of supporting functions.conColScheme has moved to ChemoSpecUtils
and can now also handle Spectra2D objects. It is also now
more user friendly.shiftSpectra2D added to permit manual
shifting.centscaleSpectra2D gains the ability to scale by log or
log10.MUD1 was completely rebuilt.MUD2 added, for purposes of testing
alignment algorithms.centscaleSpectra2D defaults have changed.colorSymbol was moved to package
ChemoSpecUtils.tinytest.normSpectra2D now checks the input method as a valid
choice.popSpectra2D computes “plain old PCA” on a
Spectra2D data set, using the IRLBA algorithm. The data is
unstacked before PCA.pfacSpectra2D now has a nfac as an
argument; previously the user was warned to provide it. This is clearer,
more user-friendly and more consistent with other functions.ChemoSpecUtils and ChemoSpec which
introduce more PCA variants.miaLoadings, popLoadings (never publically
released), and pfacLoadings were collapsed into
plotLoadings2D.chkSpectra.files2Spectra2DObject is now careful to remove any
dimnames of the imported matrices, to avoid causing a ruckus with
chkSpectra.toChemoSpec was removed as popSpectra2D
provides a complete workflow corresponding to unstacking, computing PCA
and reassembling the loadings.miaLoadings and pfacLoadings were
incorrectly reordering their matrices.inspectLvls was not excluding any loading matrices when
loading = NULL.miaLoadings and pfacLoadings now check to
see if the requested loading has already been computed, and if so, it is
not computed again. Loadings are named Loading_x in the
Spectra2D object. These functions also gain a plot argument
so that plots can be suppressed if desired. These changes were made for
dealing with large data sets which can occupy a lot of memory and slow
down the computation of the contours. Using plot = FALSE
allows one to compute the loadings and then figure out desirable contour
levels before running with plot = TRUE.pfacSpectra2D gains an argument setup. If
TRUE and parallel = TRUE the parallel
computational environment is automatically configured for the user. If
FALSE the user is responsible for setting up the
environment. This is useful if working on Azure or AWS EC2.plotSlice added.showScale now opens a pdf document showing the
scale used with plotSpectra2D.inspectLvls.MUD1 rebuilt.plotSpectra2D now accepts xlim and
ylim.plotSpectra2D now chosen via
pretty() and are easier to interpret.plotSpectra2D gains an option for a grid aligned with
the ticks.ChemoSpecUtils::.chkArgs.hcaScores added via ChemoSpecUtils.ChemoSpecUtils.ChemoSpecUtils.plotScree, which works with
miaSpectra2D.miaLoadings.groupNcolor moved to non-exported functions, and
re-named to .groupNcolors.import2Dspectra extensively reworked to fix problems
and allow for easy expansion in the future.files2Spectra2DObject tweaked a bit.centscaleSpectra2D and now
centers and optionally scales the data.plotSpectra2D so that it will plot more than
one spectrum..findNA..findNA led to code clean up
(better variable names and code notes).centerSpectra2D.chkSpectra2D.pfacLoadings now returns the modified
Spectra2D object, not the loadings matrix.inspectLvls can now access the loadings matrix, if
present, in a Spectra2D object.pfacSpectra2D was fixed. This task is now handled by helper
function .makeArray since it is also required by
centerSpectra2D. ## Misc.pfacLoadings simplified, and some logical errors
fixed.MUD1 gains some negative peaks for more comprehensive
graphical tests..mapColors overhauled..plotEngine overhauled..normAroundZero overhauled, and renamed to
.symAroundZero.inspectLvls to allow quick
selection of levels.centerSpectra2D function.MUD1 test data set.removePeaks2D.removePeaks2D (part of the problem was in
.findNA).sumSpectra2D now reports the number of data points in
F2 and the number of slices in F1.chkSpectra2D does a better job of checking and
reporting about extra data.ChemoSpec materials.