OsteoSort 1.3.2

OsteoSort 1.3.2

Changes

OsteoSort 1.3.2 will be the final release of OsteoSort with support for 2D and 3D data. This should be finalized within the next few weeks. The 3D osteometric sorting method will be pulled into a new project and released as an independent tool (OS3D) due to increased complexity. The 2D osteomertic sorting method will be deprecated.

Future iterations of OsteoSort (1.4.x) will include a cloud-based environment (freely accessible with registered accounts) and instructions on deploying a docker container. That release will support the traditional metric-based osteometric sorting methods. This is anticpated to be released sometime next year.

OsteoSort 1.3.2 Installation

require("devtools")
install_github("jjlynch2/OsteoSort", ref="v1.3.2")
library(OsteoSort)
OsteoSort()

R Dependencies

• pixmap
• jpeg
• Morpho
• DT
• shiny
• htmltools
• zip
• rgl
• ClusterR
• JuliaCall
• ggplot2
• shinyalerts
• ggrepel
• networkD3
• network
• sna
• grid
• dplyr
• shinydashboard
• shinyWidgets

Julia Dependencies

• Pkg
• Statistics
• Distributed
• SharedArrays
• Optim
• Rmath

Other Windows Dependencies

• Julia must be in your Windows environrment PATH to run.

Known Issues

• Some systems require the RCall package in Julia to be rebuilt.

using Pkg
Pkg.build("RCall")

OsteoSort Version 1.3.1

OsteoSort 1.3.1

Installation

require("devtools")
install_github("jjlynch2/OsteoSort", ref="v1.3.1")
library(OsteoSort)
OsteoSort()

R Dependencies

• pixmap
• jpeg
• Morpho
• DT
• shiny
• htmltools
• zip
• rgl
• ClusterR
• JuliaCall
• ggplot2
• shinyalerts
• ggrepel
• networkD3
• network
• sna
• grid
• dplyr

Julia Dependencies

• Pkg
• Statistics
• Distributed
• SharedArrays
• Optim
• Rmath

Changes for OsteoSort version 1.3.1

• Bug fix: the independent and dependent mean and standard deviations were switched for probability calculation

Other Windows Dependencies

• Julia must be in your Windows environrment PATH to run.

Known Issues

• Some versions of JuliaCall require RCall to be manually rebuilt in Julia.
• Some newer versions of R may require Sys.setenv(R_REMOTES_NO_ERRORS_FROM_WARNINGS="true") to avoid warnings being converted to errors during installation.

OsteoSort Version 1.3.0

OsteoSort 1.3.0

Installation

require("devtools")
install_github("jjlynch2/OsteoSort", ref="v1.3.0")
library(OsteoSort)
OsteoSort()

R Dependencies

• pixmap
• jpeg
• Morpho
• DT
• shiny
• htmltools
• zip
• rgl
• ClusterR
• JuliaCall
• ggplot2
• shinyalerts
• ggrepel
• networkD3
• network
• sna
• grid
• dplyr

Julia Dependencies

• Pkg
• Statistics
• Distributed
• SharedArrays
• Optim
• Rmath

Changes for OsteoSort version 1.3.0

• The Julia programming language is now integrated as the new analytical platform.
• Reference data can be imported and exported.
• The user interface is now dynamic and generates based on the reference data and user input.
• The ability to check for updates on GitHub has been implemented.
• Z-transform method is now implemeneted.
• Three-dimensional point cloud antimere sorting is now supported.
• Two-dimensional photograph antimere sorting now uses pairwise registration.

Other Windows Dependencies

• Julia must be in your Windows environrment PATH to run.

Known Issues

• Some versions of JuliaCall require RCall to be manually rebuilt in Julia.
• Some newer versions of R may require Sys.setenv(R_REMOTES_NO_ERRORS_FROM_WARNINGS="true") to avoid warnings being converted to errors during installation.

OsteoSort Version 1.2.6

#Current development status for 1.2.6

Installation

Requires Rtools to build on Windows

Requires Xcode to build on OSX

To install directly from GitHub use the devtools R package:

install.packages("devtools")

library(devtools)

install_github("jjlynch2/OsteoSort", ref="v1.2.6")

Depends

• compiler
• Morpho
• parallel
• pixmap
• jpeg
• CCA
• data.table
• Rcpp
• RcppParallel
• car

Changes for OsteoSort version : 1.2.6

Minor changes:

• Updated imports for DESC
• Adusted default scale option for 2D outline extraction

Moderate changes:

• Added Boxcox transformation for t-tests
• Added dependency on car
• Added option to select number of Principal Components for regression analysis in reg.multitest()
• Added option (default) to select number of Principal Components based on cumulative variance

Major changes:

• Added input.3d() and match.3d() functions for importing and pair-matching 3D data respectively
• Added rotation(), translation(), and mean_shape() functions

Bug fixes:

• Fixed bug with prediction interval lines for simple regression

OsteoSort Version 1.2.6rc1

Changes for OsteoSort version : 1.2.6rc1

Minor changes:

Updated imports for DESC

Adusted default scale option for 2D outline extraction

Moderate changes:

Added Boxcox transformation for t-tests

Added dependency on car

Added option to select number of Principal Components for regression analysis in reg.multitest()

Major changes:

Added input.3d() and match.3d() functions for importing and pair-matching 3D data respectively

Added rotation(), translation(), and mean_shape functions

OsteoSort version 1.2.5

Installation

Requires Rtools to build on Windows

Requires Xcode to build on OSX

To install directly from GitHub use the devtools R package:

require("devtools")

install_github("jjlynch2/OsteoSort", ref = "v1.2.5")

Depends

• compiler
• Morpho
• parallel
• pixmap
• jpeg
• CCA
• data.table
• Rcpp
• RcppParallel

Changes for OsteoSort version : 1.2.5

Minor changes:

• Added ByteCompile: yes on install
• Adjusted articulation t-test default options

Moderate changes:

• Removed plyr as a dependency
• Adjusted population specific Trotter corrections for antemortem stature strength association

OsteoSort version 1.2.4

Changes for OsteoSort version : 1.2.4

Minor:

Added example datasets antemortem_data and postmortem_data

Adjusted plots and removed labels

Moderate changes:

Probability values calculated for PCACCA-Regression with options to use alpha level as cut-off

OsteoSort version 1.2.3

OsteoSort:

A statistical osteological sorting R package for commingled human remains using metric, 2D, and 3D data.

Changes for OsteoSort version : 1.2.3

Minor changes:

Adjusted print messages for start/finish of functions

Moderate changes:

parLapply implemented on Windows to replace Lapply

Bug fixes:

Fixed directory creation if not present for sessiontempdir

Fixed testagainstzero argument mislabeled as testagainst

Fixed measurement argument being NULL

Fixed typo with Fem_17 mislabeled as Fem_18

Fixed typo with Tib_11 mislabeled as Tib_10

Fixed Factors being introduced by forcing stringsAsFactors = FALSE within each function excplicity. Value returns to TRUE when complete This solves the memory issue assocaited with articulation and association.

OsteoSort version 1.2.2

OsteoSort:

A statistical osteological sorting R package for commingled human remains using metric, 2D, and 3D data.

Changes for OsteoSort version 1.2.2:

Modules:

Updated 2D to include fragmentary shape registration and pair-matching.

Added match.3d() function (currently broken) for 3D pair-matching.

Minor changes:

Momocs dependency removed: replaced EFA from Momocs with modified version for 2D and 3D EFA.

Pracma dependency removed: replaced distmat from pracma with modified version.

Shapes dependency removed: removed Procrustes distance function call as an option from 2D settings.

Moved code for matching matrices names and alpha color for plots into two internal functions.

Moderate changes:

Renamed segmented_hausdorff_dist() function to hausdorff_dist().

Renamed match.2d.invariant() function to match.2d().

Re-wrote the base code for match.2d() to be more efficient with looping for distance calculations.

Re-wrote the base code for write.TPS to use lists with unequal landmarks and 3D matrices with equal landmarks.

Re-wrote all file output into output_options() including TPS, excel, and plots.

Added pca_align() function to return rotated PCA scores for alignment prior to 3D analysis

Re-wrote the backend for Hausdorff with RcppParallel to use parallel processing for large point clouds. max_directional_hausdorff_rcpp(), mean_directional_hausdorff_rcpp(), dilated_directional_hausdorff_rcpp()

Added new functions, fragment_margins(), remove_fragment_margins(), and minimum_euclidean_deistances_ind() , which are ussed to identifyssss the indices of fragment margins along a orderly 2D outline and remove any non-overlapping landmark from the calculation of Hausdorff distances.

The regression model used for antemortem stature association has been switched to calculate the point estimate of the measurement rather than vice versa.

Added statistical option to change from 1 to 2 tail t-tests.

Switched terminology from cores to threads for functions.

Bug fixes:

OsteoSort version 1.2.1

OsteoSort:

A statistical osteological sorting R package for commingled human remains using metric, 2D, and 3D data.

Changes for OsteoSort version 1.2.1:

Modules:

Added a new module to test the strength of evidence for an antemortem stature to postmortem measurement

Minor changes:

Added ability to switch output metric between inches, millimeters, and centimeters for the statsort function.

Adjusted resolution of graphs to 400x400.

Re-ran roxygenize for man pages.

Adjusted reference to previous code by Julien Claude (Morphometrics with R 2008).

Bug fixes:

Fixed Trotter data in outlier analysis for the over correction by 10 mm

Standard error was estimated slightly wrong for the t-statistic in regression association. This has been fixed.