plot_PCA and plot_LDA did not consistently
work within eg pdf(). Thanks to Bill fof pointing this.
(214)return for return() to
please R CMD checkas_df now uses tibble verbs
everywheredplyr::as_data_frame for
tibble::as_tibbleKMEDOIDS on top of
cluster::pam. Added a plot_silhouette to go
with this friend. Now depends cluster.NMDS on top of
vegan::metaMDS; use plot_NMDS to plot it and
vegan::stressplot for a Shepard plot. Now depends
vegan.MDS on top of
cmdscale; use plot_MDS to plot it.mshapes is now MSHAPES to stick with other
capitalized “multivariate” methods. mshapes now just
announces its future deprecation.MSHAPES now just returns data_frame of PCs or LDs when
used on PCA or LDA objects.plot_MSHAPES is the new method for
plotting it. Works on lists and on the result of
MSHAPES.plot_CV has been refreshed, and is better now at
plotting either small or very big matrices.fac_dispatcher supports NULL which eases a
lot multivariate plots (notably Momecs side)rm_missing to deal with missing
data in $facboxplot methods on Coe have been refreshedhist methods on Coe deprecatedcoo_plot no longer a method and gains a
cex.first.point argumentcoo_scalars that gather all scalar
descriptors of shapeTraCoe class properly data_frameize fac
when build from TraCoe()CLUST methods have been rewrote and now wraps around
dendextend.ape dependency.lists which is
more elegant when working with
chop+combineLDA methods have been partly rewritten and now handles
constant and collinear variables by dropping them and storing them in
the returned listcoo_untiltx now removes (residual) rotational biases
after coo_slidedirection and should be used after it.plot_LDA now on. Pretty much the same as
plot_PCA (which was expected yet nice)..layerize_LDA as an internal to prepare the
previousMomocs_coo and Momocs_FAQ;
others refreshed.morphospace_position and chullfilled in
plot_PCA now properly workingverify replaces validate to avoid conflict
with shiny::validate (for Momecs)subsetize now exported (again)def_ldk gains a close and
points argumentdata.frame rather than data_frameCoo builders gain a .data.frame
method, notably to ease compatibility with Momitas_df now returns more useful data_frame
everywhere and gain a retain argumentcoo_angle_edge1 and friends, now in
coo_angle_edges see 1.2.9.github.com/MomX/Momocsplot.PCA), family pictures
(eg stack will be replace by pile and remove
an annoying conflict with utils::stack, panel)
and single shape plotters (eg ldk_plot,
coo_plot). This strategy is faster, much more generic and
will ease further development and maintenance compared to previous
Momocs graphs.grindr rationale and use.andnow and andnow_method
class tells you what to do with this object, and which classes are
supported by this function/method.coo_* methods: coo_range,
coo_range_enlarge, coo_diffrange,
coo_template_relatively. The latter will prepare ground for
proper size handling, notably for morphospaces.coo functions ported to methods and now supporting
.Coo directly: coo_angle_edges,
coo_angle_tangent, coo_boundingbox,
coo_calliper, coo_chull,
coo_chull_onion, coo_circularity,
coo_circularityharalick, coo_circularity_norm,
coo_convexity, coo_dxy,
coo_eccentricityboundingbox,
coo_eccentricityeigen, coo_elongation,
coo_intersect_angle, coo_intersect_direction,
coo_intersect_segment, coo_perim,
coo_perimcum, coo_perimpts,
coo_rectangularity, coo_rectilinearity,
coo_scalex, coo_scaley,
coo_solidity, coo_truss.pal_manual, pal_qual_solarized and
pal_seq_grey.dispatch_fac is now behind all fac
argumentsfgProcrustes now accepts listsefourier with default norm=TRUE now
messages about how wrong it may bedplyr::tibble() everywhere pertinentis.* aliases for is_* methodsclassify for a whilecalibrate_*(..., method) have been renamed to
calibrate_*_method. See
?calibrate_reconstructions and friends.plot3.PCA (will be replaced in further
versions)Ntable now splitted into of plot_table +
tablecoo_tangle now in coo_angle_tangentcoo_theta3 now in coo_angle_edgestruss now in coo_truss and a method of its
ownplot.Coo is now inspectpos.shapes is now
morphospace_positionsis_closed is deprecated, now
coo_is_closed; same for is_open now
coo_is_open, to comply with all coo_* friends
naming schemeis_clockwise is deprecated, now
coo_likely_clockwise; same for
is_anticlockwise now coo_likely_anticlockwise.
Better reflect the incertainty and gather with coo_*
friendstable (poor shortcut anyway and avoid this
boring startup message)stack2 and panel2 before their
rewritingas_Out, that should have been
efourier_i.OutCoe anyway.panel.OutCoe method that was
additionnaly the only Coe method. May be back in further
versions.., eg .is.error (try
Momocs:::. + <Tab> for a complete list).function_foo renamed to .function_foo):
.coo_angle_edge1, .vecs_param,
.refactorNEWS is now a decent NEWS filereshape2, plyr dependenciesRColorBrewer, progress::. A
nice side effect is to remove annoying messages when attaching Momocs.
Another is the removal of most importFrom.subset is now
subsetize…LDA when retain=1
(#e7704eb)options("verbose")coo_intersect_segment,
coo_intersect_direction, coo_intersect_angle,
def_ldk_direction, def_ldk_angle,
def_ldk, def_ldk_tips,
coo_sample_propcoo_slice.Opn now supports ldk
argument.nsfishes and charring to comply
R CMD CHECK.coo_slice now suports ldk argumentMANOVA_PW now returns p-valuesnsfishestestthatsfourier family implementationapodemus and mouseplot2.PCA deprecated due to ggplot2 2.2.0 breaking
changes0.2.6 (the one from the JSS paper.NEWS)Out, Opn and
Ldk to handle (closed) outlines, open outlines and
configuration of landmarks. Coo becomes a “super class”
encompassing the three others.MANOVA,
MANOVA_PW, LDA, KMEANS,
CLUST.panel,
stack, plot.PCAchaff, flowers,
oak, olea, molars,
shapes, wings,browseVignette("Momocs") is available
Need a high-speed mirror for your open-source project?
Contact our mirror admin team at info@clientvps.com.
This archive is provided as a free public service to the community.
Proudly supported by infrastructure from VPSPulse , RxServers , BuyNumber , UnitVPS , OffshoreName and secure payment technology by ArionPay.