Add RStudio project file. Update docs.

.Rbuildignore

@@ -1,2 +1,4 @@
 bench
-notes.md
+notes.md
+^.*\.Rproj$
+^\.Rproj\.user$

.gitignore

@@ -0,0 +1,6 @@
+.Rproj.user
+.Rhistory
+.RData
+src/*.o
+src/*.so
+src/*.dll

NAMESPACE

@@ -1,3 +1,5 @@
+# Generated by roxygen2 (4.1.1): do not edit by hand
+
 S3method("[",dgrid)
 S3method("[<-",ranged)
 S3method(Math,condensed)

bigvis.Rproj

@@ -0,0 +1,20 @@
+Version: 1.0
+
+RestoreWorkspace: Default
+SaveWorkspace: Default
+AlwaysSaveHistory: Default
+
+EnableCodeIndexing: Yes
+UseSpacesForTab: Yes
+NumSpacesForTab: 2
+Encoding: UTF-8
+
+RnwWeave: Sweave
+LaTeX: pdfLaTeX
+
+AutoAppendNewline: Yes
+
+BuildType: Package
+PackageUseDevtools: Yes
+PackageInstallArgs: --no-multiarch --with-keep.source
+PackageRoxygenize: rd,collate,namespace

man/autoplot.condensed.Rd

@@ -1,20 +1,20 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/autoplot.r
 \name{autoplot.condensed}
 \alias{autoplot.condensed}
 \title{Autoplot condensed summaries.}
 \usage{
-  \method{autoplot}{condensed} (x,
-    var = last(summary_vars(x)), ...)
+\method{autoplot}{condensed}(x, var = last(summary_vars(x)), ...)
 }
 \arguments{
-  \item{x}{a condensed summary}
+\item{x}{a condensed summary}
 
-  \item{var}{which summary variable to display}
+\item{var}{which summary variable to display}
 
-  \item{...}{other arguments passed on to individual
-  methods}
+\item{...}{other arguments passed on to individual methods}
 }
 \description{
-  Autoplot condensed summaries.
+Autoplot condensed summaries.
 }
 \examples{
 if (require("ggplot2")) {

man/best_h.Rd

@@ -1,48 +1,44 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/h.r
 \name{best_h}
 \alias{best_h}
 \title{Find "best" smoothing parameter using leave-one-out cross validation.}
 \usage{
-  best_h(x, h_init = NULL, ..., tol = 0.01,
-    control = list())
+best_h(x, h_init = NULL, ..., tol = 0.01, control = list())
 }
 \arguments{
-  \item{x}{condensed summary to smooth}
+\item{x}{condensed summary to smooth}
 
-  \item{h_init}{initial values of bandwidths to start
-  search out. If not specified defaults to 5 times the
-  binwidth of each variable.}
+\item{h_init}{initial values of bandwidths to start search out. If not
+specified defaults to 5 times the binwidth of each variable.}
 
-  \item{...}{other arguments (like \code{var}) passed on to
-  \code{\link{rmse_cv}}}
+\item{...}{other arguments (like \code{var}) passed on to
+\code{\link{rmse_cv}}}
 
-  \item{tol}{numerical tolerance, defaults to 1\%.}
+\item{tol}{numerical tolerance, defaults to 1\%.}
 
-  \item{control}{additional control parameters passed on to
-  \code{\link{optim}} The most useful argument is probably
-  trace, which makes it possible to follow the progress of
-  the optimisation.}
+\item{control}{additional control parameters passed on to \code{\link{optim}}
+The most useful argument is probably trace, which makes it possible to
+follow the progress of the optimisation.}
 }
 \value{
-  a single numeric value representing the bandwidth that
-  minimises the leave-one-out estimate of rmse. Vector has
-  attributes \code{evaluations} giving the number of times
-  the objective function was evaluated. If the optimisation
-  does not converge, or smoothing is not needed (i.e. the
-  estimate is on the lower bounds), a warning is thrown.
+a single numeric value representing the bandwidth that minimises
+  the leave-one-out estimate of rmse. Vector has attributes
+  \code{evaluations} giving the number of times the objective function
+  was evaluated. If the optimisation does not converge, or smoothing is not
+  needed (i.e. the estimate is on the lower bounds), a warning is thrown.
 }
 \description{
-  Minimises the leave-one-out estimate of root mean-squared
-  error to find find the "optimal" bandwidth for smoothing.
+Minimises the leave-one-out estimate of root mean-squared error to find
+find the "optimal" bandwidth for smoothing.
 }
 \details{
-  L-BFGS-B optimisation is used to constrain the bandwidths
-  to be greater than the binwidths: if the bandwidth is
-  smaller than the binwidth it's impossible to compute the
-  rmse because no smoothing occurs. The tolerance is set
-  relatively high for numerical optimisation since the
-  precise choice of bandwidth makes little difference
-  visually, and we're unlikely to have sufficient data to
-  make a statistically significant choice anyway.
+L-BFGS-B optimisation is used to constrain the bandwidths to be greater
+than the binwidths: if the bandwidth is smaller than the binwidth it's
+impossible to compute the rmse because no smoothing occurs. The tolerance
+is set relatively high for numerical optimisation since the precise choice
+of bandwidth makes little difference visually, and we're unlikely to have
+sufficient data to make a statistically significant choice anyway.
 }
 \examples{
 x <- rchallenge(1e4)
@@ -56,8 +52,7 @@ autoplot(smooth(xsum, h))
 }
 }
 \seealso{
-  Other bandwidth estimation functions:
-  \code{\link{h_grid}}, \code{\link{rmse_cv}},
-  \code{\link{rmse_cvs}}
+Other bandwidth estimation functions: \code{\link{h_grid}};
+  \code{\link{rmse_cv}}, \code{\link{rmse_cvs}}
 }
 

man/bigvis.Rd

@@ -1,9 +1,11 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/bigvis.r
 \docType{package}
 \name{bigvis}
 \alias{bigvis}
 \alias{bigvis-package}
 \title{The big vis package.}
 \description{
-  The big vis package.
+The big vis package.
 }
 

man/bin.Rd

@@ -1,33 +1,31 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/bin.r
 \name{bin}
 \alias{bin}
 \title{Create a binned variable.}
 \usage{
-  bin(x, width = find_width(x),
-    origin = find_origin(x, width), name = NULL)
+bin(x, width = find_width(x), origin = find_origin(x, width), name = NULL)
 }
 \arguments{
-  \item{x}{numeric or integer vector}
+\item{x}{numeric or integer vector}
 
-  \item{width}{bin width. If not specified, about 10,000
-  bins will be chosen using the algorithim in
-  \code{\link{find_width}}.}
+\item{width}{bin width. If not specified, about 10,000 bins will be chosen
+using the algorithim in \code{\link{find_width}}.}
 
-  \item{origin}{origin. If not specified, guessed by
-  \code{\link{find_origin}}.}
+\item{origin}{origin. If not specified, guessed by \code{\link{find_origin}}.}
 
-  \item{name}{name of original variable. This will be
-  guessed from the input to \code{group} if not supplied.
-  Used in the output of \code{\link{condense}} etc.}
+\item{name}{name of original variable. This will be guessed from the input to
+\code{group} if not supplied. Used in the output of
+\code{\link{condense}} etc.}
 }
 \description{
-  Create a binned variable.
+Create a binned variable.
 }
 \details{
-  This function produces an R reference class that wraps
-  around a C++ function. Generally, you should just treat
-  this as an opaque object with reference semantics, and
-  you shouldn't call the methods on it - pass it to
-  \code{\link{condense}} and friends.
+This function produces an R reference class that wraps around a C++ function.
+Generally, you should just treat this as an opaque object with reference
+semantics, and you shouldn't call the methods on it - pass it to
+\code{\link{condense}} and friends.
 }
 \examples{
 x <- runif(1e6)

man/breaks.Rd

@@ -1,25 +1,25 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/breaks.r
 \name{breaks}
 \alias{breaks}
 \title{Compute breaks given origin and width.}
 \usage{
-  breaks(x, binwidth, origin = min(x))
+breaks(x, binwidth, origin = min(x))
 }
 \arguments{
-  \item{x}{numeric vector}
+\item{x}{numeric vector}
 
-  \item{origin}{bin origin}
+\item{binwidth}{bin width}
 
-  \item{binwidth}{bin width}
+\item{origin}{bin origin}
 }
 \description{
-  Breaks are right-open, left-closed [x, y), so if
-  \code{max(x)} is an integer multiple of binwidth, then we
-  need one more break. This function only returns the
-  left-side of the breaks.
+Breaks are right-open, left-closed [x, y), so if \code{max(x)} is an integer
+multiple of binwidth, then we need one more break. This function only returns
+the left-side of the breaks.
 }
 \details{
-  The first break is special, because it always contains
-  missing values.
+The first break is special, because it always contains missing values.
 }
 \examples{
 breaks(10, origin = 0, binwidth = 1)

man/condense.Rd

@@ -1,32 +1,31 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/condense.r
 \name{condense}
 \alias{condense}
 \title{Efficient binned summaries.}
 \usage{
-  condense(..., z = NULL, summary = NULL, w = NULL,
-    drop = NULL)
+condense(..., z = NULL, summary = NULL, w = NULL, drop = NULL)
 }
 \arguments{
-  \item{...}{group objects created by \code{\link{bin}}}
+\item{...}{group objects created by \code{\link{bin}}}
 
-  \item{z}{a numeric vector to summary for each group.
-  Optional for some summary statistics.}
+\item{z}{a numeric vector to summary for each group. Optional for some
+summary statistics.}
 
-  \item{summary}{the summary statistic to use. Currently
-  must be one of count, sum, mean, median or sd. If
-  \code{NULL}, defaults to mean if y is present, count if
-  not.}
+\item{summary}{the summary statistic to use. Currently must be one of
+count, sum, mean, median or sd. If \code{NULL}, defaults to mean if
+y is present, count if not.}
 
-  \item{w}{a vector of weights. Not currently supported by
-  all summary functions.}
+\item{w}{a vector of weights. Not currently supported by all summary
+functions.}
 
-  \item{drop}{if \code{TRUE} only locations with data will
-  be returned.  This is more efficient if the data is very
-  sparse (<1\% of cells filled), and is slightly less
-  efficient. Defaults to \code{TRUE} if you are condensing
-  over two or more dimensions, \code{FALSE} for 1d.}
+\item{drop}{if \code{TRUE} only locations with data will be returned.  This
+is more efficient if the data is very sparse (<1\% of cells filled), and
+is slightly less efficient. Defaults to \code{TRUE} if you are condensing
+over two or more dimensions, \code{FALSE} for 1d.}
 }
 \description{
-  Efficient binned summaries.
+Efficient binned summaries.
 }
 \examples{
 x <- runif(1e5)

man/condensed.Rd

@@ -1,35 +1,37 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/condensed.r
 \name{condensed}
 \alias{as.condensed}
 \alias{condensed}
 \alias{is.condensed}
 \title{Condensed: an S3 class for condensed summaries.}
 \usage{
-  condensed(groups, grouped, summary)
+condensed(groups, grouped, summary)
 
-  is.condensed(x)
+is.condensed(x)
 
-  as.condensed(x)
+as.condensed(x)
 }
 \arguments{
-  \item{groups}{list of \code{\link{bin}}ed objects}
+\item{groups}{list of \code{\link{bin}}ed objects}
 
-  \item{grouped,summary}{output from C++ condense function}
+\item{grouped,summary}{output from C++ condense function}
 
-  \item{x}{object to test or coerce}
+\item{x}{object to test or coerce}
 }
 \description{
-  This object managed the properties of condensed
-  (summarised) data frames.
+This object managed the properties of condensed (summarised) data frames.
 }
 \section{S3 methods}{
-  Mathematical functions with methods for \code{binsum}
-  object will modify the x column of the data frame and
-  \code{\link{rebin}} the data, calculating updated summary
-  statistics.
-
-  Currently methods are provided for the \code{Math} group
-  generic, logical comparison and arithmetic operators, and
-  \code{\link[plyr]{round_any}}.
+
+
+Mathematical functions with methods for \code{binsum} object will modify
+the x column of the data frame and \code{\link{rebin}} the data, calculating
+updated summary statistics.
+
+Currently methods are provided for the \code{Math} group generic,
+logical comparison and arithmetic operators, and
+\code{\link[plyr]{round_any}}.
 }
 \examples{
 if (require("ggplot2")) {

man/dchallenge.Rd

@@ -1,26 +1,26 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/challenge.r
 \name{dchallenge}
 \alias{dchallenge}
 \alias{rchallenge}
 \title{Density and random number generation functions for a challenging
 distribution.}
 \usage{
-  dchallenge(x)
+dchallenge(x)
 
-  rchallenge(n)
+rchallenge(n)
 }
 \arguments{
-  \item{x}{values to evaluate pdf at}
+\item{x}{values to evaluate pdf at}
 
-  \item{n}{number of random samples to generate}
+\item{n}{number of random samples to generate}
 }
 \description{
-  This is a 1/3-2/3 mixture of a t-distribution with 2
-  degrees of freedom centered at 15 and scaled by 2, and a
-  gamma distribution with shape 2 and rate 1/3. (The
-  t-distribution is windsorised at 0, but this has
-  negligible effect.) This distribution is challenging
-  because it mixes heavy tailed and asymmetric
-  distributions.
+This is a 1/3-2/3 mixture of a t-distribution with 2 degrees of freedom
+centered at 15 and scaled by 2, and a gamma distribution with shape 2
+and rate 1/3. (The t-distribution is windsorised at 0, but this
+has negligible effect.) This distribution is challenging because it
+mixes heavy tailed and asymmetric distributions.
 }
 \examples{
 plot(dchallenge, xlim = c(-5, 60), n = 500)