NEWS (original) (raw)

projections 0.6.1

Fixes and improvements

• fixed tests of dates due failing due to changes inas.Date()
• fixed example in the manpage of summary.projection

projections 0.6.0

Fixes and improvements

• fixed over-dispersion parameter in NegBin model
• improved support for NegBin and Rt models

projections 0.5.4

Fixes and improvements

• add option to use instantaneous R rather than the default case R in project

projections 0.5.3

Fixes and improvements

• harmonise si definition with EpiEstim.

projections 0.5.2

Fixes and improvements

• now using 3rd edition of testthat for checks.

projections 0.5.1

Fixes and improvements

• now using the new default random number generator in tests
• updated tests to pass check with R 4.0.2; this includes re-generation of most reference objects, which were manually inspected and validated against old references

projections 0.5.0

New functions

• new function merge_projections permits to join different sets of simulations, adapting the respective time windows accordingly
• new function merge_add_projections permits to add simulated case incidence from different sets of simulations, adapting the respective time windows accordingly; if objects contain different numbers of simulations, the shortest ones are recycled as needed; also implemented as the operator +
• new function summary will summariseprojections objects by day, using a range of pre-defined statistics (mean, sd, min, max, and user-defined quantiles)

Fixes and improvements

• bug fix for the use of time-varying R inproject
• bug fix for as.data.frame(..., long = TRUE)
• more consistent handling of inputs for the serial interval; if provided as a vector, si now starts at a delay of 1 day, rather than 0 and assuming the first entry is 0; fordistcrete inputs, the mass for si$d(0) is now ignored, and the rest of the distribution is rescaled accorindly to ensure the PMF sums to 1
• (documentation) more realistic serial interval distribution used for the Ebola examples
• (internal) the force of infection is now calculated by the new internal function compute_force_infection
• (testing) revised sets of tests relying less on comparison to references, but testing for meaningful output properties instead
• (testing) 100% coverage

projections 0.4.1

• project() will can now take single dates or single projections as inputs. (@acori, @zkamvar, #18).

projections 0.4.0

• project can now use time-varying R by specifyingtime_change, a vector of dates at which R changes, and providing a vector or a list of values forR instead of a vector, in which case it needs to havelength(time_change) + 1 components, each of which is a vector of R values.

projections 0.3.2

• A bug in project() where R was being resampled recursively was fixed (#11, @jarvisc1; #12, @zkamvar)

projections 0.3.1

• get_dates() now inherits the genericget_dates() from incidence

projections 0.3.0

New features

This is a big release! Plenty of new features have been added, including:

• projections can now be subsetted like matrices usingx[i,j], or using the function subset
• new function cumulate to compute cumulative incidence for projections objects, akin to the similar function in the incidence package
• much improved graphics; plot now call uponsadd_projections, which implements many options for plotting projections, including quantiles lines, ribbon, and boxplots;add_projections can also be used to add such plots to an existing incidence plot
• build_projections can be used to build aprojections object from an input matrix and optional dates

projections 0.1.1

New features

• project can now use a Negative Binomial model as an alternative to the original Poisson model to account for over-dispersion in new cases / super-spreading.

Bug fixes

Fixed the calls to sample, which has a rather irrational behaviour when passing a first argument of length 1 (i.e. not bootstrapping from the vector).

projections 0.0.1

First release of the package!

New features

• project: a function generating projections from an existing incidence object, a serial interval distribution, and a set of plausible reproduction numbers (\(R\)); returns a projectionsobject.
• plot/print: plotting and printing methods for projections objects.
• get_dates: accessors for projectionsobjects.
• as.data.frame: conversion fromprojections objects to data.frame.