[talweg.git] / pkg / R / plot.R

#' plot curves
#'
#' Plot a range of curves in data
#'
#' @param data Object of class Data
#' @param indices Range of indices (integers or dates)
#'
#' @export
plotCurves <- function(data, indices=seq_len(data$getSize()))
{
	yrange = quantile( sapply( indices, function(i) {
		serie = c(data$getCenteredSerie(i))
		if (!all(is.na(serie)))
			range(serie, na.rm=TRUE)
		c()
	}), probs=c(0.05,0.95) )
	par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
	for (i in seq_along(indices))
	{
		plot(data$getSerie(indices[i]), type="l", ylim=yrange,
			xlab=ifelse(i==1,"Temps (en heures)",""), ylab=ifelse(i==1,"PM10",""))
		if (i < length(indices))
			par(new=TRUE)
	}
}

#' plot measured / predicted
#'
#' Plot measured curve (in black) and predicted curve (in red)
#'
#' @param data Object return by \code{getData}
#' @param pred Object as returned by \code{computeForecast}
#' @param index Index in forecasts
#'
#' @export
plotPredReal <- function(data, pred, index)
{
	horizon = length(pred$getSerie(1))
	measure = data$getSerie(pred$getIndexInData(index)+1)[1:horizon]
	yrange = range( pred$getSerie(index), measure )
	par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5, lwd=3)
	plot(measure, type="l", ylim=yrange, xlab="Temps (en heures)", ylab="PM10")
	par(new=TRUE)
	plot(pred$getSerie(index), type="l", col="#0000FF", ylim=yrange, xlab="", ylab="")
}

#' Compute filaments
#'
#' Get similar days in the past + "past tomorrow", as black as distances are small
#'
#' @param data Object as returned by \code{getData}
#' @param index Index in data
#' @param limit Number of neighbors to consider
#' @param plot Should the result be plotted?
#'
#' @export
computeFilaments <- function(data, index, limit=60, plot=TRUE)
{
	index = dateIndexToInteger(index, data)
	ref_serie = data$getCenteredSerie(index)
	if (any(is.na(ref_serie)))
		stop("computeFilaments requires a serie without NAs")
	L = length(ref_serie)

	# Determine indices of no-NAs days followed by no-NAs tomorrows
	fdays = c()
	for (i in 1:(index-1))
	{
		if ( !any(is.na(data$getSerie(i)) | is.na(data$getSerie(i+1))) )
			fdays = c(fdays, i)
	}

	distances = sapply(fdays, function(i) {
		sqrt( sum( (ref_serie - data$getCenteredSerie(i))^2 ) / L )
	})
	indices = sort(distances, index.return=TRUE)$ix[1:min(limit,length(distances))]
	yrange = quantile( c(ref_serie, sapply( indices, function(i) {
		serie = c(data$getCenteredSerie(fdays[i]), data$getCenteredSerie(fdays[i]+1))
		if (!all(is.na(serie)))
			return (range(serie, na.rm=TRUE))
		c()
	}) ), probs=c(0.05,0.95) )
	grays = gray.colors(20, 0.1, 0.9) #TODO: 20 == magic number
	min_dist = min(distances[indices])
	max_dist = max(distances[indices])
	color_values = floor( 19.5 * (distances[indices]-min_dist) / (max_dist-min_dist) ) + 1
	plot_order = sort(color_values, index.return=TRUE, decreasing=TRUE)$ix
	colors = c(grays[ color_values[plot_order] ], "#FF0000")
	if (plot)
	{
		par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5, lwd=2)
		for ( i in seq_len(length(indices)+1) )
		{
			ii = ifelse(i<=length(indices), fdays[ indices[plot_order[i]] ], index)
			plot(c(data$getCenteredSerie(ii),data$getCenteredSerie(ii+1)),
				ylim=yrange, type="l", col=colors[i],
				xlab=ifelse(i==1,"Temps (en heures)",""), ylab=ifelse(i==1,"PM10 centré",""))
			if (i <= length(indices))
				par(new=TRUE)
		}
		abline(v=24, lty=2, col=colors()[56])
	}
	list("indices"=c(fdays[ indices[plot_order] ],index), "colors"=colors)
}

#' Plot similarities
#'
#' Plot histogram of similarities (weights)
#'
#' @param pred Object as returned by \code{computeForecast}
#' @param index Index in forecasts (not in data)
#'
#' @export
plotSimils <- function(pred, index)
{
	weights = pred$getParams(index)$weights
	if (is.null(weights))
		stop("plotSimils only works on 'Neighbors' forecasts")
	par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
	hist(pred$getParams(index)$weights, nclass=20, xlab="Poids", ylab="Effectif")
}

#' Plot error
#'
#' Draw error graphs, potentially from several runs of \code{computeForecast}
#'
#' @param err Error as returned by \code{computeError}
#' @param cols Colors for each error (default: 1,2,3,...)
#'
#' @seealso \code{\link{plotPredReal}},\code{\link{plotFilaments}}
#'   \code{\link{plotSimils}},\code{\link{plotFbox}},\code{\link{plotRelativeVariability}}
#'
#' @export
plotError <- function(err, cols=seq_along(err))
{
	if (!is.null(err$abs))
		err = list(err)
	par(mfrow=c(2,2), mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5, lwd=2)
	L = length(err)
	yrange = range( sapply(1:L, function(index) ( err[[index]]$abs$day ) ), na.rm=TRUE )
	for (i in seq_len(L))
	{
		plot(err[[i]]$abs$day, type="l", xlab=ifelse(i==1,"Temps (heures)",""),
			ylab=ifelse(i==1,"Moyenne |y - y_hat|",""), ylim=yrange, col=cols[i])
		if (i < L)
			par(new=TRUE)
	}
	yrange = range( sapply(1:L, function(index) ( err[[index]]$abs$indices ) ), na.rm=TRUE )
	for (i in seq_len(L))
	{
		plot(err[[i]]$abs$indices, type="l", xlab=ifelse(i==1,"Temps (jours)",""),
			ylab=ifelse(i==1,"Moyenne |y - y_hat|",""), ylim=yrange, col=cols[i])
		if (i < L)
			par(new=TRUE)
	}
	yrange = range( sapply(1:L, function(index) ( err[[index]]$MAPE$day ) ), na.rm=TRUE )
	for (i in seq_len(L))
	{
		plot(err[[i]]$MAPE$day, type="l", xlab=ifelse(i==1,"Temps (heures)",""),
			ylab=ifelse(i==1,"MAPE moyen",""), ylim=yrange, col=cols[i])
		if (i < L)
			par(new=TRUE)
	}
	yrange = range( sapply(1:L, function(index) ( err[[index]]$MAPE$indices ) ), na.rm=TRUE )
	for (i in seq_len(L))
	{
		plot(err[[i]]$MAPE$indices, type="l", xlab=ifelse(i==1,"Temps (jours)",""),
			ylab=ifelse(i==1,"MAPE moyen",""), ylim=yrange, col=cols[i])
		if (i < L)
			par(new=TRUE)
	}
}

#' Functional boxplot
#'
#' Draw the functional boxplot on the left, and bivariate plot on the right
#'
#' @param data Object return by \code{getData}
#' @param fiter Optional filter: return TRUE on indices to process
#' @param plot_bivariate Should the bivariate plot appear?
#'
#' @export
plotFbox <- function(data, filter=function(index) TRUE, plot_bivariate=TRUE)
{
	if (!requireNamespace("rainbow", quietly=TRUE))
		stop("Functional boxplot requires the rainbow package")

	L = length(data$getCenteredSerie(2))
	series_matrix = sapply(1:data$getSize(), function(index) {
		if (filter(index))
			as.matrix(data$getSerie(index))
		else
			rep(NA,L)
	})
	# TODO: merge with previous step: only one pass should be required
	no_NAs_indices = sapply( 1:ncol(series_matrix),
		function(i) all(!is.na(series_matrix[,i])) )
	series_matrix = series_matrix[,no_NAs_indices]

	series_fds = rainbow::fds(seq_len(nrow(series_matrix)), series_matrix)
	if (plot_bivariate)
		par(mfrow=c(1,2))
	par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
	rainbow::fboxplot(series_fds, "functional", "hdr", xlab="Temps (heures)", ylab="PM10",
		plotlegend=FALSE, lwd=2)
	if (plot_bivariate)
		rainbow::fboxplot(series_fds, "bivariate", "hdr", plotlegend=FALSE)
}

#' Functional boxplot on filaments
#'
#' Draw the functional boxplot on filaments obtained by \code{computeFilaments}
#'
#' @param data Object return by \code{getData}
#' @param indices Indices as output by \code{computeFilaments}
#'
#' @export
plotFilamentsBox = function(data, indices, ...)
{
	past_neighbs_indices = head(indices,-1)
	plotFbox(data, function(i) i %in% past_neighbs_indices, plot_bivariate=FALSE)
	par(new=TRUE)
	# "Magic" found at http://stackoverflow.com/questions/13842560/get-xlim-from-a-plot-in-r
	usr <- par("usr")
	yr <- (usr[4] - usr[3]) / 27
	plot(data$getSerie(tail(indices,1)), type="l", lwd=2, lty=2,
		ylim=c(usr[3] + yr, usr[4] - yr), xlab="", ylab="")
}

#' Plot relative conditional variability / absolute variability
#'
#' Draw the relative conditional variability / absolute variability based on filaments
#' obtained by \code{computeFilaments}
#'
#' @param data Object return by \code{getData}
#' @param indices Indices as output by \code{computeFilaments}
#'
#' @export
plotRelativeVariability = function(data, indices, ...)
{
	ref_series = t( sapply(indices, function(i) {
		c( data$getSerie(i), data$getSerie(i+1) )
	}) )
	ref_var = apply(ref_series, 2, sd)

	# Determine indices of no-NAs days followed by no-NAs tomorrows
	fdays = c()
	for (i in 1:(tail(indices,1)-1))
	{
		if ( !any(is.na(data$getSerie(i)) | is.na(data$getSerie(i+1))) )
			fdays = c(fdays, i)
	}
	global_var = c( apply(data$getSerie(fdays),2,sd), apply(data$getSerie(fdays+1),2,sd) )

	yrange = range(ref_var, global_var)
	par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
	plot(ref_var, type="l", col=1, lwd=3, ylim=yrange,
		xlab="Temps (heures)", ylab="Écart-type")
	par(new=TRUE)
	plot(random_var, type="l", col=2, lwd=3, ylim=yrange, xlab="", ylab="")
	abline(v=24, lty=2, col=colors()[56])
}
Commit	Line	Data
	1	#' plot curves
	2	#'
	3	#' Plot a range of curves in data
	4	#'
	5	#' @param data Object of class Data
	6	#' @param indices Range of indices (integers or dates)
	7	#'
	8	#' @export
	9	plotCurves <- function(data, indices=seq_len(data$getSize()))
	10	{
	11	yrange = quantile( sapply( indices, function(i) {
	12	serie = c(data$getCenteredSerie(i))
	13	if (!all(is.na(serie)))
	14	range(serie, na.rm=TRUE)
	15	c()
	16	}), probs=c(0.05,0.95) )
	17	par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
	18	for (i in seq_along(indices))
	19	{
	20	plot(data$getSerie(indices[i]), type="l", ylim=yrange,
	21	xlab=ifelse(i==1,"Temps (en heures)",""), ylab=ifelse(i==1,"PM10",""))
	22	if (i < length(indices))
	23	par(new=TRUE)
	24	}
	25	}
	26
	27	#' plot measured / predicted
	28	#'
	29	#' Plot measured curve (in black) and predicted curve (in red)
	30	#'
	31	#' @param data Object return by \code{getData}
	32	#' @param pred Object as returned by \code{computeForecast}
	33	#' @param index Index in forecasts
	34	#'
	35	#' @export
	36	plotPredReal <- function(data, pred, index)
	37	{
	38	horizon = length(pred$getSerie(1))
	39	measure = data$getSerie(pred$getIndexInData(index)+1)[1:horizon]
	40	yrange = range( pred$getSerie(index), measure )
	41	par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5, lwd=3)
	42	plot(measure, type="l", ylim=yrange, xlab="Temps (en heures)", ylab="PM10")
	43	par(new=TRUE)
	44	plot(pred$getSerie(index), type="l", col="#0000FF", ylim=yrange, xlab="", ylab="")
	45	}
	46
	47	#' Compute filaments
	48	#'
	49	#' Get similar days in the past + "past tomorrow", as black as distances are small
	50	#'
	51	#' @param data Object as returned by \code{getData}
	52	#' @param index Index in data
	53	#' @param limit Number of neighbors to consider
	54	#' @param plot Should the result be plotted?
	55	#'
	56	#' @export
	57	computeFilaments <- function(data, index, limit=60, plot=TRUE)
	58	{
	59	index = dateIndexToInteger(index, data)
	60	ref_serie = data$getCenteredSerie(index)
	61	if (any(is.na(ref_serie)))
	62	stop("computeFilaments requires a serie without NAs")
	63	L = length(ref_serie)
	64
	65	# Determine indices of no-NAs days followed by no-NAs tomorrows
	66	fdays = c()
	67	for (i in 1:(index-1))
	68	{
	69	if ( !any(is.na(data$getSerie(i)) \| is.na(data$getSerie(i+1))) )
	70	fdays = c(fdays, i)
	71	}
	72
	73	distances = sapply(fdays, function(i) {
	74	sqrt( sum( (ref_serie - data$getCenteredSerie(i))^2 ) / L )
	75	})
	76	indices = sort(distances, index.return=TRUE)$ix[1:min(limit,length(distances))]
	77	yrange = quantile( c(ref_serie, sapply( indices, function(i) {
	78	serie = c(data$getCenteredSerie(fdays[i]), data$getCenteredSerie(fdays[i]+1))
	79	if (!all(is.na(serie)))
	80	return (range(serie, na.rm=TRUE))
	81	c()
	82	}) ), probs=c(0.05,0.95) )
	83	grays = gray.colors(20, 0.1, 0.9) #TODO: 20 == magic number
	84	min_dist = min(distances[indices])
	85	max_dist = max(distances[indices])
	86	color_values = floor( 19.5 * (distances[indices]-min_dist) / (max_dist-min_dist) ) + 1
	87	plot_order = sort(color_values, index.return=TRUE, decreasing=TRUE)$ix
	88	colors = c(grays[ color_values[plot_order] ], "#FF0000")
	89	if (plot)
	90	{
	91	par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5, lwd=2)
	92	for ( i in seq_len(length(indices)+1) )
	93	{
	94	ii = ifelse(i<=length(indices), fdays[ indices[plot_order[i]] ], index)
	95	plot(c(data$getCenteredSerie(ii),data$getCenteredSerie(ii+1)),
	96	ylim=yrange, type="l", col=colors[i],
	97	xlab=ifelse(i==1,"Temps (en heures)",""), ylab=ifelse(i==1,"PM10 centré",""))
	98	if (i <= length(indices))
	99	par(new=TRUE)
	100	}
	101	abline(v=24, lty=2, col=colors()[56])
	102	}
	103	list("indices"=c(fdays[ indices[plot_order] ],index), "colors"=colors)
	104	}
	105
	106	#' Plot similarities
	107	#'
	108	#' Plot histogram of similarities (weights)
	109	#'
	110	#' @param pred Object as returned by \code{computeForecast}
	111	#' @param index Index in forecasts (not in data)
	112	#'
	113	#' @export
	114	plotSimils <- function(pred, index)
	115	{
	116	weights = pred$getParams(index)$weights
	117	if (is.null(weights))
	118	stop("plotSimils only works on 'Neighbors' forecasts")
	119	par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
	120	hist(pred$getParams(index)$weights, nclass=20, xlab="Poids", ylab="Effectif")
	121	}
	122
	123	#' Plot error
	124	#'
	125	#' Draw error graphs, potentially from several runs of \code{computeForecast}
	126	#'
	127	#' @param err Error as returned by \code{computeError}
	128	#' @param cols Colors for each error (default: 1,2,3,...)
	129	#'
	130	#' @seealso \code{\link{plotPredReal}},\code{\link{plotFilaments}}
	131	#' \code{\link{plotSimils}},\code{\link{plotFbox}},\code{\link{plotRelativeVariability}}
	132	#'
	133	#' @export
	134	plotError <- function(err, cols=seq_along(err))
	135	{
	136	if (!is.null(err$abs))
	137	err = list(err)
	138	par(mfrow=c(2,2), mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5, lwd=2)
	139	L = length(err)
	140	yrange = range( sapply(1:L, function(index) ( err[[index]]$abs$day ) ), na.rm=TRUE )
	141	for (i in seq_len(L))
	142	{
	143	plot(err[[i]]$abs$day, type="l", xlab=ifelse(i==1,"Temps (heures)",""),
	144	ylab=ifelse(i==1,"Moyenne \|y - y_hat\|",""), ylim=yrange, col=cols[i])
	145	if (i < L)
	146	par(new=TRUE)
	147	}
	148	yrange = range( sapply(1:L, function(index) ( err[[index]]$abs$indices ) ), na.rm=TRUE )
	149	for (i in seq_len(L))
	150	{
	151	plot(err[[i]]$abs$indices, type="l", xlab=ifelse(i==1,"Temps (jours)",""),
	152	ylab=ifelse(i==1,"Moyenne \|y - y_hat\|",""), ylim=yrange, col=cols[i])
	153	if (i < L)
	154	par(new=TRUE)
	155	}
	156	yrange = range( sapply(1:L, function(index) ( err[[index]]$MAPE$day ) ), na.rm=TRUE )
	157	for (i in seq_len(L))
	158	{
	159	plot(err[[i]]$MAPE$day, type="l", xlab=ifelse(i==1,"Temps (heures)",""),
	160	ylab=ifelse(i==1,"MAPE moyen",""), ylim=yrange, col=cols[i])
	161	if (i < L)
	162	par(new=TRUE)
	163	}
	164	yrange = range( sapply(1:L, function(index) ( err[[index]]$MAPE$indices ) ), na.rm=TRUE )
	165	for (i in seq_len(L))
	166	{
	167	plot(err[[i]]$MAPE$indices, type="l", xlab=ifelse(i==1,"Temps (jours)",""),
	168	ylab=ifelse(i==1,"MAPE moyen",""), ylim=yrange, col=cols[i])
	169	if (i < L)
	170	par(new=TRUE)
	171	}
	172	}
	173
	174	#' Functional boxplot
	175	#'
	176	#' Draw the functional boxplot on the left, and bivariate plot on the right
	177	#'
	178	#' @param data Object return by \code{getData}
	179	#' @param fiter Optional filter: return TRUE on indices to process
	180	#' @param plot_bivariate Should the bivariate plot appear?
	181	#'
	182	#' @export
	183	plotFbox <- function(data, filter=function(index) TRUE, plot_bivariate=TRUE)
	184	{
	185	if (!requireNamespace("rainbow", quietly=TRUE))
	186	stop("Functional boxplot requires the rainbow package")
	187
	188	L = length(data$getCenteredSerie(2))
	189	series_matrix = sapply(1:data$getSize(), function(index) {
	190	if (filter(index))
	191	as.matrix(data$getSerie(index))
	192	else
	193	rep(NA,L)
	194	})
	195	# TODO: merge with previous step: only one pass should be required
	196	no_NAs_indices = sapply( 1:ncol(series_matrix),
	197	function(i) all(!is.na(series_matrix[,i])) )
	198	series_matrix = series_matrix[,no_NAs_indices]
	199
	200	series_fds = rainbow::fds(seq_len(nrow(series_matrix)), series_matrix)
	201	if (plot_bivariate)
	202	par(mfrow=c(1,2))
	203	par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
	204	rainbow::fboxplot(series_fds, "functional", "hdr", xlab="Temps (heures)", ylab="PM10",
	205	plotlegend=FALSE, lwd=2)
	206	if (plot_bivariate)
	207	rainbow::fboxplot(series_fds, "bivariate", "hdr", plotlegend=FALSE)
	208	}
	209
	210	#' Functional boxplot on filaments
	211	#'
	212	#' Draw the functional boxplot on filaments obtained by \code{computeFilaments}
	213	#'
	214	#' @param data Object return by \code{getData}
	215	#' @param indices Indices as output by \code{computeFilaments}
	216	#'
	217	#' @export
	218	plotFilamentsBox = function(data, indices, ...)
	219	{
	220	past_neighbs_indices = head(indices,-1)
	221	plotFbox(data, function(i) i %in% past_neighbs_indices, plot_bivariate=FALSE)
	222	par(new=TRUE)
	223	# "Magic" found at http://stackoverflow.com/questions/13842560/get-xlim-from-a-plot-in-r
	224	usr <- par("usr")
	225	yr <- (usr[4] - usr[3]) / 27
	226	plot(data$getSerie(tail(indices,1)), type="l", lwd=2, lty=2,
	227	ylim=c(usr[3] + yr, usr[4] - yr), xlab="", ylab="")
	228	}
	229
	230	#' Plot relative conditional variability / absolute variability
	231	#'
	232	#' Draw the relative conditional variability / absolute variability based on filaments
	233	#' obtained by \code{computeFilaments}
	234	#'
	235	#' @param data Object return by \code{getData}
	236	#' @param indices Indices as output by \code{computeFilaments}
	237	#'
	238	#' @export
	239	plotRelativeVariability = function(data, indices, ...)
	240	{
	241	ref_series = t( sapply(indices, function(i) {
	242	c( data$getSerie(i), data$getSerie(i+1) )
	243	}) )
	244	ref_var = apply(ref_series, 2, sd)
	245
	246	# Determine indices of no-NAs days followed by no-NAs tomorrows
	247	fdays = c()
	248	for (i in 1:(tail(indices,1)-1))
	249	{
	250	if ( !any(is.na(data$getSerie(i)) \| is.na(data$getSerie(i+1))) )
	251	fdays = c(fdays, i)
	252	}
	253	global_var = c( apply(data$getSerie(fdays),2,sd), apply(data$getSerie(fdays+1),2,sd) )
	254
	255	yrange = range(ref_var, global_var)
	256	par(mar=c(4.7,5,1,1), cex.axis=1.5, cex.lab=1.5)
	257	plot(ref_var, type="l", col=1, lwd=3, ylim=yrange,
	258	xlab="Temps (heures)", ylab="Écart-type")
	259	par(new=TRUE)
	260	plot(random_var, type="l", col=2, lwd=3, ylim=yrange, xlab="", ylab="")
	261	abline(v=24, lty=2, col=colors()[56])
	262	}