This is a major update to aqp that may create some issues for code depending on specific inputs/outputs in aqp < 1.42, particularly those relying on slice(), slab(), and profile_compare(). slice() and profile_compare() are now deprecated, but will continue to work for the rest of calendar year 2023. There are no plans to maintain these functions beyond aqp 2.0. The new version of slab() is a drop-in replacement for the previous version of the function.

Notable changes include:

  • deprecation of slice() in favor of the new, faster, more robust implementation in dice()
  • complete overhaul of slab(), with new arguments, faster back-end, and weighted aggregation implemented (finally)
  • deprecation of profile_compare() in favor of the NCSP()–a complete overhaul based on Maynard et al., 2020
    • site level attributes now handled by compareSites()
    • variable weights now possible via argument
  • perturb() and estimatePSCS() are now vectorized, and optimized for larger collections
  • mixMunsell() now uses mixingMethod = 'exact' by default for the simulation of subtractive mixtures
  • gower package moved to SUGGESTS
  • plotColorMixture() now using grid graphics functions to determine color swatch geometry and setting overlap detection threshold
  • removal of PMS2Munsell() and support data
  • deprecation of coordinates()<- and proj4string()<- in favor of initSpatial()<-
  • removal of rruff.sample example XRD patterns
  • get.ml.hz() no longer uses the name argument

Major changes to plotSPC():

  • The maximum depth range of the figure is now based on max.depth or max(x). This means that sketches generated with aqp 2.x will generally have less white space at the bottom of the figure. Make more room for additional annotation or visual effect by setting the desired depth range with the max.depth argument.
  • now uses electroStatics_1D() for fixing horizon depth label overlap, solutions are deterministic and almost always better
  • better depth axis interval heuristics (if not specified), varying based on figure depth range
  • depth axis adjustments via new argument depth.axis, logical or list
  • deprecation of arguments:
    • plot.depth.axis: set via depth.axis = TRUE, depth.axis = FALSE, or customize depth.axis = list(...)
    • cex.depth.axis: set via depth.axis = list(cex = 1)
    • axis.line.offset: set via depth.axis = list(line = -2)

New features:

  • example data, wilson2022
  • fast prototyping of SPCs via quickSPC() and list / character templates
  • re-use arguments to plotSPC() via options(.aqp.plotSPC.args = list(...))
  • coarse fragment classification via fragmentSieve() and fragmentClasses()
  • S4 as.data.frame(<SPC>) as shorthand for as(<SPC>, 'data.frame')
  • plotSPC() now marks truncated profiles with a ragged bottom
  • fixOverlap() now has two label-placement solvers, based on 1) electrostatics and 2) simulated annealing
  • new depth axis styles in plotSPC()

Incremental changes, should have no effect on previous code:

  • bug fix in plotSPC() when fixLabelCollisions = TRUE, adjustments suggested to fixOverlap() are now scaled correctly
  • explainPlotSPC() reports label adjustment index when label collision repair is enabled
  • aesthetic cleanup in explainPlotSPC()
  • soilColorSignature() gains arguments and perceptual color distances (dE00) via farver package
  • as(<SPC>, "data.frame"): Replace plyr::join() with merge()
  • correctAWC(): NA handling - return NA when frags are NA
  • mutate_profile(): Faster (data.table-based) evaluation of profile-level expressions (#255)
  • profileApply: Add support for custom lapply()-like function (APPLY.FUN) for processing chunks (#256)
  • Add .interpretHorizonColor() outputs to last_spc_plot in aqp.env for use in custom legend() (#254)
  • Add simplify argument to SoilTextureLevels() and ssc_to_texcl() to optionally convert to an ordered factor with maximum of 12 levels (rather than 21). This smaller list of classes excludes sand grain size variants such as fine sand, loamy coarse sand, and very fine sandy loam.
  • CRAN release
  • getArgillicBounds(), getCambicBounds(), mollic.thickness.requirement(), getSurfaceHorizonDepth() and related functions have been optimized and now work on SoilProfileCollection objects with length > 1
  • Added reduceSPC() function for selecting a subset of minimal plus additional specified columns from a SoilProfileCollection
  • Added new “k-keyword” .NHZ for use with SoilProfileCollection [ method
    • object[,,.NHZ] returns a vector with the number of horizons in each profile
  • proper identification and conversion of Munsell value == 2.5
  • interpolated reference spectra for Munsell value == 2.5
  • neutral color chips in the Munsell look-up table are now based on direct measurement vs. estimation
  • plotSPC() gains new arguments for controlling horizon depth labels
  • plotSPC() gains horizon depth annotation collision fixes (https://github.com/ncss-tech/aqp/issues/240)
  • plotSPC() minor bugfix for cases when n != length(x)
  • plotSPC() important change! y.offset is now automatically re-ordered by plot.order
  • bugfix for R 4.2, related to named arguments passed to base::aggregate (thanks AGB)
  • depthOf() and related functions: better handling of NA results from FUN
  • colorChart() can now show neutral hues in a skinny panel
  • updates to colorContrast(), parseMunsell(), and huePosition() to work properly with neutral hues
  • huePosition() gains new arguments and functionality: origin and direction
  • new function huePositionCircle() for helping visualize differences in Munsell hue on the unit circle
  • plotSPC defaults change:
    • width = 0.25: slightly wider profile sketches
    • name = hzdesgnname(x): use metadata if present, no longer guessing. This may cause your figures to “lose” previously guessed horizon names, specify with name = 'column name' or set metadata with hzdesgnname(x) <- 'hzname'
  • CRAN release
  • Several SoilProfileCollection methods that conflict with {dplyr} 1.0+ have been deprecated:
    • filter, mutate, group_by, summarize
    • New overloaded {base} names:
      • {base}-like syntax: “filter” -> subset, “mutate” -> transform, “combine” -> c
    • New unique {aqp}/SoilProfileCollection aliases:
      • *SPC syntax: filterSPC, mutateSPC, groupSPC, summarizeSPC
    • Be aware dplyr::combine (deprecated in {dplyr} 0.7) still conflicts
  • aqp::union(), previously deprecated, has been removed from namespace
  • Methods that used {rlang} for non-standard evaluation now use {base}
  • plotSPC() gains vectorized y.offset support (demonstrated in: http://ncss-tech.github.io/AQP/aqp/genhz-distance-eval.html)
  • plotSPC() gains argument shrink.thin for shrinking horizon designation labels when horizons are thin
  • new function alignTransect() for simplifying relative positioning of profile sketches
  • plotMultipleSPC() gains ability to automatically merge thematic legends
  • coordinates<- will check formula terms (unique coordinates) in the @horizons slot, if needed.
  • new function spec2Munsell() for converting reflectance spectra into sRGB coordinates or closest Munsell chip
  • mixMunsell:
    • new mixingMethod argument for selecting several mixing strategies
    • suggestions on interpreting spectral distances, message printed when greater than reasonable threshold
  • minor bug-fix in panel.depth_function when plotting grouped step-functions
  • optimization of SoilProfileCollection [,j]-index extraction using {data.table}
  • introduction of .LAST, .FIRST and .HZID SoilProfileCollection “k-keywords”
  • perturb() is the new generalized replacement for sim() and permute_profile()
  • checkHzDepthLogic() now has a byhz argument for checking logic by horizon rather than profile
  • fillHzGaps now has to_top and to_bottom arguments for filling above shallowest top / deepest bottom by profile
  • fixOverlap() more flexible and will usually settle on a solution in fewer iterations:
  • alignTransect helper function for computing relative positions and ordering vector supplied to plotSPC
  • plotSPC() automatically converts logical data supplied to color argument into factor
  • glom() is now vectorized over profiles
  • estimateAWC introduced for testing lookup table estimation of available water capacity (AWC) of fine-earth fraction
  • correctAWC introduced for testing corrections of AWC estimates for rock fragment and salts
  • fillHzGaps: new function for fixing horizon depth topological errors and padding top/bottom of profiles with placeholder (empty) horizons
  • mixMunsell now relies on suggested package {gower} for 5-10x speed bump
  • {aqp} no longer imports from {reshape} (less one dependency), all transformations from wide<->long are done via {data.table}
  • methods from {data.table} are now imported by {aqp} (new dependency)
  • Major overhaul of plotColorQuantiles(), now using {lattice} graphics
  • New dataset equivalent_munsell and method equivalentMunsellChips for “equivalent” Munsell chips lookup list based on all pairwise dE00 contrasts for integer “chips” in {aqp} munsell data set
  • Argillic critical clay contents crit.clay.argillic rounded to whole numbers per NSSH Part 614, subpart B, sections 614.13 and 614.14
  • mutate_profile uses data.table::rbindlist(fill=TRUE) to combine site- and horizon-level transformations
  • updates to horizon boundary encoding functions (hzTopographyCodeToOffset, hzTopographyCodeToLineType, hzDistinctnessCodeToOffset)
  • new function L1_profiles computes multivariate (L1) medians, compare to marginal medians via slab
  • plotSPC updates:
    • argument named changes: hz.boundary.lty is a horizon-level attribute that contains line type codes
    • hz.topography.offset a horizon-level attribute that contains representative offsets that encode horizon boundary topography
    • plotSPC now encodes hz.topography.offset using a vertical “bump” (chevron)
  • addBracket can now accept multiple bracket annotations per profile
  • CRAN release
  • new lookup table pms.munsell.lut for converting Pantone spot color codes to (closest) Munsell chip
  • new function duplicate will makes copies of profiles within a SoilProfileCollection
  • new example data us.state.soils: 50 state soils + PR and VI soils
  • simulate subtractive mixtures of Munsell colors with mixMunsell
    • see companion function plotColorMixture for visualization of spectra / mixture
  • complete overhaul of textureTriangleSummary:
    • uses soiltexture package for visualization (plotrix implementation dropped)
    • argument names changes (! may break old code, sorry)
    • dropped simulation via sim = TRUE argument, see bootstrapSoilTexture for a better approach
  • new function bootstrapSoilTexture for simulating realistic sand/silt/clay compositions
  • combine replaces/expands aqp::union due to conflicts with base::union
  • split receives some upgrades to the S4 definition to increase parity with split.default
  • filter is now an alias for new method subset, which mirrors base::subset
  • estimateSoilDepth loses top and bottom arguments, these are automatically extracted
  • two new SoilProfileCollection wrapper methods: munsell2SPC, spc2mpspline
  • add returnData argument to contrastChart
  • improvements to glom(..., invert=TRUE), glomApply, and better tests
  • new wrapper method around glomApply: aqp::trunc for cases when top and bottom depth interval is the same for all profiles in a SoilProfileCollection
  • fix for routing of NULL through $<- and horizons<- or site<- (https://github.com/ncss-tech/aqp/issues/163)
  • fix handling of missing metadata in (old) serialized SoilProfileCollection objects
  • fix for promotion of data.table with character vector (not formula) interface
  • enhanced SoilProfileCollection object validity checks via S4; new method spc_in_sync (https://github.com/ncss-tech/aqp/pull/152)
  • optimization of [ subset method and optional use of data.table (https://github.com/ncss-tech/aqp/pull/155)
  • depths<- has been optimized and minimally validates input data
  • default horizon ID (hzID) is now a character data type
  • aqp::union uses depths<- internally; explicitly enforcing profile ID + top depth order in horizon data is safer but results in different ordering if union-ing IDs that “intermingle” (need to be re-sorted).
  • new experimental method is permute_profile; similar to sim but for boundaries. The interface to this function is likely to change/be expanded.
  • added segment c/o @smroecker
  • fix for unit-length and zero-length legends in plotSPC
  • fix for plot generic to show aqp::plot in ?plot
  • proof of concept for tidy SoilProfileCollection subsetting
  • define [[ subsetting method; an “ambivalent” accessor for site- or horizon-level properties
  • new subset verbs grepSPC, filter, subApply for use in %>%-lines
  • CRAN release
  • new functions: hzDesgn(), get horizon designations from a SPC
  • new functions: hzdesgnname()/hzdesgnname()<- and hztexclname()/hztexclname()<- get/set column containing horizon designations and texture classes
  • better error/logic handling for glom()
  • new functions colorContrast(), contrastClass(), and huePosition()
  • fixed nasty bug in [-indexing when SPC contains a single profile (https://github.com/ncss-tech/aqp/issues/85)
  • new function validSpatialData() to test for place-holder contents of @sp
  • rebuildSPC() will now correct invalid, place-holder SpatialPoints in @sp
  • fixing #53 and #54
  • bug fix for too many decimal places in legend classes (floating point -> character issue)
  • plotSPC() gets a new argument, n.legend: approximate number of classes
  • enhancement to aggregateColor() for cases when horizon depths are missing or top/bottom logic flipped
  • added a new argument to addDiagnosticBracket() allowing for specification of column containing diagnostic feature kind
  • addBracket() now requires a data.frame with plotting details, see manual page
  • push to CRAN, getting ready for aqp 2.0
  • CIELAB colors are now in the munsell look-up table
  • plotSPC() now able to deal with missing horizon data (thanks Stephen R.)
  • this is likely the last release before aqp 2.0
  • rgb2munsell() now uses CIELAB colorspace for lookup, results are more accurate as compared to using sRGB colorspace
  • texture.triangle.low.rv.high() renamed to textureTriangleSummary(). The old name still works, but a message is issued
  • new argument to textureTriangleSummary() texture.names: for toggling texture class names
  • minor bug fix in textureTriangleSummary(..., sim=TRUE), previous simulated compositional data was not correct because the stats::var() was being used vs. compositions::var.acomp(). the variance / covariance values were 2-5x too small.
  • new function tauW(), added by D.G. Rossiter: see manual page for references
  • munsell2rgb() can now accommodate neutral hues (N), and values of “2.5”, see manual page for details
  • new function: soilColorSignature(), see manual page
  • new function: parseMunsell() for converting strings of ‘10YR 2/2’ into R colors or RGB triplets
    • this is a convenience function that calls munsell2rgb on parsed strings of full Munsell notation
  • documentation for new transition probability functions, see ?hzTransitionProbabilities for details
  • better checking for other places where the bug fixed in rbind.SoilProfileCollection (aqp 1.9.3) could occur
  • new functions: genhzTableToAdjMat, hzTransitionProbabilities
  • fixed a nasty bug in rbind.SoilProfileCollection() related to object re-ordering
    • in aqp < 1.9.3: there was a chance than profile_id(x) and site(x)[[idname(x)]]) would not match
    • bugs like this will no longer be a possibility after aqp 2.0, new SPC object structure planned
  • groupedProfilePlot() gains some new arguments and better documentation (see manual page)
  • plot.SoilProfileCollection gains argument: default.color (see manual page)
  • bug fixes and documentation updates
  • digest::digest is now imported by default
  • better documentation for unique-method of SoilProfileCollection objects
  • updated documentation
  • plotSPC gets a new argument: x.idx.offset: allows for plotting of multiple SPC objects within the same figure, see examples
  • plotSCP gets a new argument: plot.depth.axis: enable / disable depth axis, useful for multi-collection plots
  • new function: profileGroupLabels: simple annotation for groups of profiles within a profile sketch, see manual page for examples
  • new function: plotMultipleSPC: plot several SPC objects on the same axis, see manual page for examples
  • names() method for SoilProfileCollection objects returns a concatenated vector of horizon + site names
  • siteNames() and horizonNames() can now be used to extract from SoilProfileCollection
  • siteNames<- and horizonNames<- can now be used to change names; experimental!
  • note: internal structure of SoilProfileCollection objects will be changing soon!!!
  • texture.triangle.low.rv.high(): new arguments, likely breaking previous usage when method=‘closest’, see manual page for details
  • bug fix c/o Jose Padarian: SPC objects now understand logical indexing rules
  • removed spatial_subset(): this functionality can be accomplished outside of AQP and removes dependency on rgeos package
  • bug fix c/o Jos? Padarian: when promoting coordinates from @site, drop=FALSE is required to prevent a single remaining attribute from being down-graded to a vector– thanks!
  • get.ml.hz() now returns a “pseudo” Brier Score, using the most-likely horizon label as the “correct” value
  • new function rgb2munsell()
  • new function estimateSoilDepth()
  • new function getSoilDepthClass()
  • new documentation / tutorials
  • When applied to a single categorical variable, the results from slab() now contain an attribute ‘original.levels’ that contains the original factor levels. This is important because when casting from long->wide format, illegal column names are scrubbed by make.names(). This process would convert horizon designations like ‘2Bt’ into ‘X2Bt’. You can recover the original horizon names via attr(x, 'original.levels').
  • plotSPC() has a new argument: label, used to set site-level attribute containing profile labels
  • plotSPC() now registers plotting parameters in the environment aqp.env
  • new function addVolumeFraction() for annotating profile plots with volumetric information (e.g. rock fragment volume)
  • plotSPC() attempts to guess which column contains horizon designations
  • plotSPC() can now color horizons based on continuous soil properties (e.g. clay content)
  • rbind.SoilProfileCollection() now re-orders data according to the new set of profile IDs
    • this fixes several problems associated with assumptions in profile_compare()
  • profile_compare() will be re-written soon to work natively with SPC objects
  • new function addBracket() for annotating profile sketches with depth-brackets; useful for plotting diagnostic features
  • new function missingDataGrid() for graphically describing the presence of missing data in an SPC object
  • panel.depth_function() now prints “contributing fraction” values in the same color as corresponding depth-function lines and shaded areas
  • panel.depth_function() gains an argument ‘cf.col’ for manually specifying the CF color
  • panel.depth_function() gains an argument ‘cf.interval’ for manually specifying printed CF depths
  • prepanel.depth_function() now extends x-lim by 5% in either direction for cleaner-looking plots
  • added some details to slab() documentation
  • fixed minor bug in min() and max() methods for SoilProfileCollection objects in the presence of missing data
  • min() and max() methods for SoilProfileCollection objects now have a second argument for checking depth based on the presence of data in that column
  • removed excessive checking in SoilProfileCollection init code, this may speed things up
    • previously identified horizonation errors may no longer result in ERRORS; be sure to test
  • test_hz_logic() now more useful:
    • always checks for NA and overlapping horizons
    • no longer called with initialization of SoilProfileCollection objects
    • note that default behavior has changed since version 1.5
    • moved related documentation to test_hz_logic manual page
  • new mini-vignette on dealing with bad data (dealing-with-bad-data.html)
  • better documentation for slice()
  • new warnings related to missing data in the grouping variable passed to slab()
  • profileApply() can now return arbitrary objects = new subsetting/transformation possibilities
    • see ?profileApply for examples
  • panel.depth_function() can now display contributing fraction values with grouped data
    • see multi-variate example in ?slab
  • HTML version of all documentation (including images) auto-generated via knitr and available on the r-forge website
  • new function: texture.triangle.low.rv.high() plots texture data with low-high range defined by quantiles
  • updated documentation and SoilProfileCollection tutorial
  • plans laid for parallel processing in slab() and profileApply() via ‘parallel’ package
  • plotSPC() gets a density argument for shaded coloring of horizons
  • slab() re-written from scratch, resulting in reduced memory footprint and quicker running time ! older code based on slab() is likely to now broken- sorry, it won’t happen again
    • the new code is much slimmer and based on aggregate() rather than ddply
    • function arguments have changed so be sure to check the manual page
    • user-supplied functions are now much simpler to write and rely on vectors rather than data.frames (= faster)
    • the next major release will contain additional examples of how to use the new features
  • new function sim() for simulation based on a template soil profile, proper documentation pending…
  • new function subsetProfiles() for simpler subsetting of SoilProfileCollection objects via site or horizon-level attributes
  • get.ml.hz() now returns a ‘confidence’ for each horizon (as a percent)
  • added experimental ‘filter’ argument to profile_compare()- see manual page for usage
  • added new argument to plot.SoilProfileCollection()- alt.label: used to add secondary labeling via data in @site
  • plot.SoilProfileCollection() will now do a better job of selecting the location for profile IDs based on label length, font size, and figure size
  • minor tweaks to plot.SoilProfileCollection() for improved output with small/large size collections
  • new function get.ml.hz() attempts to determine the most likely horizonation after slabbing
  • fixed bug in profileApply() where further arguments to FUN weren’t passed on to FUN
  • missing data report from profile_compare() now only includes horizons with some data (e.g. Cr and R horizons are skipped)
  • speed-bump in profileApply()
    • splitProfiles() function is no longer defined/used/needed
  • bug fix in SoilProfileCollection subsetting via []
    • SoilProfiles with a single horizon confused subsetting in the presence of spatial data
  • added rbind.SoilProfileCollection() for concatenating SPC objects
    • this function performs basic sanity checks, but cannot combine unlike objects
  • fixed minor bug in prepanel.depth_function() where an error would occur when input data was all NA
  • new ‘unique’ method for SoilProfileCollection objects, returns indices of unique profiles
  • fixed bug in NAMESPACE that caused an error when ‘sp’ library wasn’t loaded
  • added another sanity check to ‘site<-’ for cases where duplicates occur in the source data set
  • added some new documentation to package help page see ‘help(aqp)’ for details
  • the names() method for SoilProfileCollection objects now returns a named list of horizon and site column names
  • improved examples: see ?sp4 and ?ca630
  • fixed minor bug related to subsetting a SoilProfileCollection object, followed by coercion to SpatialPointsDataFrame
  • plot() method for SoilProfileCollection object no longer sets page margins, this will cause some figures to change
  • added demo ‘slope_effect_hz_thickness’
  • fixed minor bug in panel.depth_function(): grouping + paneling ignored full suite of colors
  • fixed minor bug in NAMESPACE
  • minor bug fix in unroll(), was not previously respecting the max depth argument
  • minor speed-bump and simplification of profile_compare()
  • documentation updates
  • added helper message related to profile_compare() stopping when any profile contains insufficient data
  • profile_compare() now issues a notice when some data are missing, leading to biased results
  • slice() now computes the fraction of variables within a slice that are missing data ! bug introduced into slab() when use.wts=TRUE [should have a fix soon]
  • merged documentation for soil.slot() and slab()
    • slab() should be used in all places where soil.slot() was used
    • soil.slot() will disappear from the package NAMESPACE eventually, so don’t use it directly!
  • minor bug fix in panel.depth_function()
  • cleanup of documentation and examples
  • added horizon boundary distinctness code conversions and plotting functionality
    • see ?sp1 and ?random_profile for new examples
  • More NAMESPACE cleaning
  • Cleaned NAMESPACE and DESCRIPTION
  • removed un-necessary dependencies
  • join added to the plyr imports
  • added scales to the Imports as well, and removed rescaler() in the code, repaced by scales::rescale
  • hack in munsell2rgb to avoid the ‘no visible binding for global variable ’munsell’ at package R CMD check and make the whole thing more robust (using load and system.file rather than data(…)
  • Dylan is now the only maintainer, as per Uwe request
  • when using plyr **ply function, switched from using the .(foo) interface to “foo” - seems more foolproof
  • 1.0 release, still missing condensed vignettes- should be ready soon
  • see http://casoilresource.lawr.ucdavis.edu/drupal/taxonomy/term/56 for samples
  • A small bug in profile_compare() was fixed, where slices were evaluated as ‘soil’ based on the bottom depth of the profile, and NOT on the presence of actual data. See ?profile_compare for details. This change will have a minor affect on profile comparisons in cases where Cr or R horizons (usually missing most characterization data) have been extended down to some arbitrary depth (usually 150 or 200 cm) AND a maximum depth of evaluation (max_d) was set beyond the actual depth of most profiles in the collection.
  • profile_compare() will now use site-level attributes if input is a SoilProfileCollection
    • horizon / site level data currently weighted equally
    • NOTE: this feature is new, test-first!
  • minor bug fixes in profile_compare() to remove spurious warnings
  • better error check in slice()
  • removed old code and documentation: format_slices() and plot_slices()
  • spurious warnings from profile_compare() caused by daisy() now suppressed
    • warnings were generated when calling daisy() on a matrix of all NA which typically occurred when using too-large a lower depth for evaluation of D
  • preliminary support for diagnostic horizons via diagnostic_hz()
  • new examples
  • error and warnings messages cleaned-up
  • converting all messages passed to the user from cat() to message() and warning()
  • profile_compare() and slab() now default to no progress bars, see ‘progress’ argument to adjust
  • profileApply() now provides access to all aspects of the SoilProfileCollection object
    • that means that: profileApply(SPC, function(x) … ) operates on x, which is a soilProfileCollection containing only profile_i note that this may break functionality based on aqp 0.99-9.3
  • site(SCP) <- d now tries to merge data from with ‘d’ via left join
    • this means that all new SPC objects will have a single column of @site data, containing profile IDs
  • plot-SoilProfileCollection now tries to guess the best orientation of IDs
    • override with id.style=‘top’ or id.style=‘side’
  • slice-SoilProfileCollection now accepts ‘.’ to define all columns should be returned at requested slices
    • slice(SPC, 1:50 ~ . ) # this means all vars. ID, top, bottom are automatically removed
  • pedonPC, NASIS, and SDA functions moved out of aqp, and into new package: soilDB
    • removed associated, suggested packages from aqp
    • soilDB will be on CRAN shortly
  • new argument to plot(SoilProfileCollection, …, divide.hz=TRUE|FALSE) that can optionally suppress plotting of dividing lines between horizons (suggested by Ludwig Hilger)
  • new function profileApply() for applying functions by profile
  • almost ready for AQP 1.0, vignettes are the last missing piece :)
  • slice(SPC, …) is now 10-100x faster, scales linearly
  • slice(SPC, …) can simultaneously slice categorical and continuous variables
  • converted ID column in sample data to character class
  • fixed a bug in slice() where NA would be returned when IDs were factors
  • added test to depths()<- such that factor IDs are converted into character IDs and a warning is issued
  • added join conditions for most functions using plyr::join(), resulting in less chatter on the console
  • WARNING: soil.slot.multiple() renamed to slab()
  • fixed minor bug in slice(), when numerical IDs are used: ddply() and cast() implicitly re-order data…
  • added strict argument to many functions, to gracefully account for bad horizon data
    • enforce quality control with strict=TRUE
  • munsell2rgb() is now 100x faster thanks to plyr::join()
  • documentation updates, demo(aqp) updates, more examples
  • bug fix in slice(), now works with numeric OR categorical vars, but NOT both…
  • matrix-style subsetting functional:
    • s[i, j] —> i is the profile index, and j is the horizon/slice index
    • s[i, ] returns a SoilProfileCollection
    • s[, j] returns a SpatialPointsDataFrame when a single slice is used
  • spatial_subset() can be used to extract members of a SoilProfileCollection with a geometry (uses rgeos)
  • smarter $<- methods so that site/horizon data slots are auto-selected
  • added “site(SoilProfileCollection) <-” method for data.frame
  • started VERY basic AQP Introduction vignette… still experimental
  • adding coercion methods for common objects and sp-class objects
  • matrix-style subsetting almost works:
    • s[i, j] —> i is the profile index, and j is the horizon/slice index
  • added a ‘slice’ method for SoilProfileCollection objects
    • returns selected variables along depth slice(s) (i.e. no aggregation)
  • added NAMESPACE file for R >= 2.14 compatibility
  • stabilized S4 class, working on S4-izing all related functions
  • preliminary documentation on the new S4 classes/methods
  • logistic power peak (LPP) can now be used to generate synthetic depth functions
  • added contributing fraction annotation in panel.depth_function()
  • minor bug fixes and enhancements in PedonPC function
  • S4 interface to soil.slot.multiple()
  • added S4 class/methods for SoilProfileCollection
    • this supersedes the (now removed) S3 SoilProfileList classes
    • basic accessors/setters are in place, subject to change
  • profile_plot() now uses the SoilProfileCollection class
    • initSoilProfileList() is no longer supported; see depths() for similar functionality
    • expect some tumultuous times ahead in the API… should be ironed out by 1.0 release
  • soil.slot() will now accept boundaries defining a ‘slab’ over which aggregates are computed
  • soil.slot.multiple() now cleanly wraps soil.slot(), accepting all arguments
    • these two changes make it possible to ask: “what is the wt. mean value of some property within this slab, and among these groups?”
    • soil.slot.multiple() now uses a formula interface: NOTE that this will break existing scripts (sorry)
  • new functions for getting data out of PedonPC (MS Access) databases [windows only for now]

    • get_site_data_from_pedon_db() : site and pedon aggregate data
    • get_hz_data_from_pedon_db() : horizon level data
    • get_colors_from_pedon_db() : formats and mixes multiple colors / horizon
  • implemented in mix_and_clean_colors()

  • test_hz_logic() : basic function for testing horizon logic, returns TRUE/FALSE by ID

  • parallel operations now NON-functional, while we wait for plyr to support doSMP…

  • new ID plotting style for profile_plot() : handy when plotting large collections and/or long IDs

  • code and documentation clean-up
  • Soil Sata Access (SDA) query functions have been added
    • mapunit_geom_by_ll_bbox() : get map unit geometry by bounding box
    • MUKEYS_by_ll_bbox() : get map unit keys by bounding box
    • SDA_query() : retrieve soil tabular data via query written in SQL
  • additional customization added to plotSPC
  • two new sample data sets + examples
  • soil.slot() and profile_compare() now have support for parallel computation thanks to bug-fixes in plyr 1.4
    • plyr >= 1.4, foreach, and doMC packages are required
    • specific examples are not yet documented, but should be soon
    • this feature is still experimental! testing is advised
  • profile_compare() now calculates slice-wise dissimilarity matrices in 1/3 the time (thanks llply!)
  • fixed minor bug in soil.slot() when computing probabilities from profiles that had missing horizons and that had only a single class within the variable to be aggregated
  • added basic demo: demo(aqp)
  • soil.slot() now computes aggregate probabilities over user-defined segments
  • re-write of weighted profile aggregation functions
    • aqp package now requires Hmisc package
    • using Hmisc::wtd.{mean,var,quantile} to compute values
  • better adjustment of weights when computing weighted SD
  • fixes long-standing bugs with wt. mean/SD when NA present in x_i and not in wt_i
  • weighted quantiles now computed
    • addition of new (experimental) S4 classes and methods
    • these are now in the aqp_S4 branch
  • added n.depth.ticks option to profile_plot()
  • profile_compare() algorithm stabilized, new default settings
  • removed code in profile_compare() that is now obsolete
    • note that this will cause changes to numerical classification of soil profiles
  • verified that weighted standard deviations are correct when seg_size > 1
  • removed notices about possible problems with sd calculations
  • soil.slot() streamlined and functionality restored to pre-0.95 condition
  • fixed weighted mean / SD calculation when using 1 unit segments in soil.slot()
    • wt. mean / SD still disabled for user-defined segment sizes
  • temporarily disabled parallel computation, will be re-added in next release
  • bug fixes in panel.depth_function where ‘groups’ was not defined
  • bug fixes and major re-factoring of soil.slot() when called with user-defined segment size, or segmenting vector:
    • SD values are probably too low due to an inflated ‘n’ in the calculation
    • calculation of weighted mean and sd is currently disabled
  • added VERY experimental support for parallel processing based on the latest versions of plyr, doMC, and foreach packages
  • updated vignette with technical details on aggregation and dissimilarity computation
  • new options to profile_plot()
  • fixed bug in soil.slot() when aggregating a categorical variable, with a user-supplied segmenting vector
  • soil.slot() now requires that categorical variables be encoded as factors
  • bug fixes, new documentation + examples
  • AQP now requires R >= 2.9.0
  • including a new column in the results from soil.slot(), denoting number of profiles used in slice-wise aggregation
  • added note to soil.slot() manual page highlighting possible problems with SD calculation
  • better error checking on profile aggregation and classification functions
  • most functions now require that horizon depths are integers
  • added plot_distance_graph() function for visualizing between-profile dissimilarity
  • updated documentation: extended manual is now accessible as a vignette
  • major improvements in profile_compare() when add_soil_flag=TRUE and replace_na=TRUE
    • appears to create much more realistic groupings when there are both shallow and deep soils in the collection
    • new functionality requires further testing
  • plotting under dendrograms generated by ape::plot.phylo() may need some manual adjustments
  • new version of plyr (1.0) should speed up most functions in aqp package
  • added support for user-defined aggregate functions in soil.slot and soil.slot.multiple
  • planning addition of PCA by depth slice
  • planning addition of equal-area spline fitting