NEWS.md
In jonesor/Rcompadre: Utilities for using the 'COM(P)ADRE' Matrix Model Database

Rcompadre (development version)

Added a function mpm_elementwise_apply which can apply any function over elements in a matrix. For example, mean, min, max, sd, var etc. to allow summaries to conveniently be made across matrices of the same dimension. Replaces some of the functionality of mpm_sd/mat_sd, mpm_median/mat_median and mpm_mean/mat_mean, which are retained for backwards compatibility.
Removed non-ASCII characters in package (including data) to satisfy CRAN requirements.
Fixed test for cdb_fetch that required connection to COMPADRE Database and thus threw error.

Minor performance enhancements, some of which speed up functions while others improve code readability and stability.
Added functions mpm_sd/mat_sd and mpm_median/mat_median which join the existing mpm_mean/mat_mean functions to calculate element-wise summary statistics across MPMs of the same dimension. Code contributed by Darren Norris with modifications by Owen Jones. A future version will include a generic "summarise matrix" function.
Increased test coverage.

Fixed an error in cdb_build_cdb causing an error in the construction of the A matrix when a C matrix was not provided.
Fixed a technical issue with as_tibble.CompadreDB that caused a build failure on some platforms.

Improvements to cdb_fetch. Can add a comment as an attribute on import. Can download particular versions of the databases by version number. Can force quiet running (no messages).

Added cdb_build_cdb function, which allows users to build compadreDB objects from their own data
Added flag argument to cdb_fetch. If this is set to TRUE then the checks from cdb_flag will be run on the database during the downloaded process thus streamlining coding.
Added additional flags to cdb_fetch: check_zero_F which checks whether the F (sexual reproduction) matrix is all zeros (TRUE if all zero) and check_zero_U_colsum which checks whether there are any columns of the U matrix that are all zero (i.e. whether there are some stages where there is 100% mortality).
Updated Compadre and Comadre data samples with the new names for AnnualPeriodicity (changed to ProjectionInterval) and DOI.ISBN (changed to DOI_ISBN).

Updated DESCRIPTION with new contributors.
Improved documentation to ensure it is more understandable to less experienced users.
Grouped functions by type in the documentation.
Improved vignettes with a range of examples demonstrating the range of basic to complex features.
Added an argument (check_surv_gte_1) to cdb_check to check matU for elements greater than or equal to 1.
Added cdb_metadata function to provide easy access to metadata (without associated matrices).
Matrix accessor functions (matA, matU etc.) now ensure that the matrices are provided with named columns and rows. Names are prefixed by matrix type (e.g. A1, A2, A3 or U1, U2, U3).
Added machine-readable codemeta-data information (codemeta.json)
Unit tests improved to increase coverage.
Modified build checks via continuous integration on Travis, Appveyor and GitHub actions (including weekly checks).

Modifications to improve the class definitions, which are renamed to CompadreDB and CompadreM.
Added generic functionality with accessor methods via ClassUnionMethods. e.g. functions matA, matU, matF, and matC output all A,U,F, or C matrices from a database.
Added cdb_fetch function to obtain the latest version of COMPADRE or COMADRE.
The database object behaves more like a data frame than the complex hierarchical object it is.
Added Tidyverse support so that most dplyr functions can be used with the database (e.g. filter).
Generic subsetting methods including subset, [] now work.
Database can be used with magrittr pipes.
Most functions renamed with cdb_ prefix, followed by verb (e.g. dbCompare becomes cdb_compare and cleanDatabase becomes cdb_flag).
The function cdb_unnest unnests the DB by spreading the matrices into separate list columns.
The function convert2flat is replaced with cdb_flatten, but works on an entire database rather than a single matrix. cdb_unflatten reverses the procedure.
Other functions are also renamed with more intuitive names.
Functions collapseMatrix, identifyReprodStages, rearrangeMatrix, splitMatrix moved to the Rage package.
Added simple vignettes.
Added unit tests for all functions.

First pre-release version!
Established S4 class CompadreData with definition and methods.
Established S4 class CompadreM with definition and methods.
Functions to checkspecies, check for matrix problems (cleanDatabase),
Functions to manipulate the databases: merge databases (mergeDB), compare database versions (dbCompare), subset the database (subsetDB).
Functions to manipulate matrices: to collapse matrix to a smaller number of stages (collapseMatrix), convert matrix to a "flat" format (convert2flat), calculate the mean F matrix (getMeanMatF), segregate reproductive/non-reproductive stages rearrangeMatrix, split the matrix into submatrices (splitMatrix)
Function to produce a matrix from a string representation stringtomatrix.
Added continuous integration.