diff --git a/cran-comments.md b/cran-comments.md index d8c4304..37e9211 100755 --- a/cran-comments.md +++ b/cran-comments.md @@ -3,16 +3,18 @@ * Changed title to include the word 'image'. It now reads as 'Wavelet-Based Enhanced FDR for Signal Detection in Noisy Images' * Changed 'criteria' to 'criterion' in DESCRIPTION * Checked also on R-devel (OK) -* Put long-running vignettes in inst/doc. The minimal vignette then points the user to 'Other documentation' by typing in help(packages="EFDR"). These documents were generated using RMarkdown prior to submission of this package. +* Put long-running tutorials in inst/doc. ## Test environments * local ubuntu 14.04, R 3.1.2 * Windows 8, R 3.1.2 * win-builder (release) * R-devel +* OS X Mavericks, R 3.1.2 ## R CMD check results -There were no ERRORs, WARNINGs or NOTEs. +There were no ERRORs, WARNINGs. +There was one NOTE because this is a new submission. ## Downstream dependencies There are no downstream dependencies diff --git a/inst/doc/EFDR_SST.Rmd b/inst/doc/EFDR_SST.Rmd index cccca00..14f3e01 100755 --- a/inst/doc/EFDR_SST.Rmd +++ b/inst/doc/EFDR_SST.Rmd @@ -67,7 +67,7 @@ The only unusual things here are the setting of `n1` and `n2` which denote the i ```{r,cache=TRUE,results='hide'} ### Carry out EFDR on every frame -cl <- makeCluster(parallel) +cl <- makeCluster(2) registerDoParallel(cl) X <- foreach (t = unique(sst$time),.combine=rbind) %dopar% { @@ -96,7 +96,7 @@ world<-map_data('world')%>% subset(long > 80 & long < 340) base_plot <- ggplot() + - theme(panel.background = element_rect(fill='white', colour='black'), + theme(panel.background = element_rect(fill='white', colour='black'), text = element_text(size=20), panel.grid.major = element_line(colour = "light gray", size = 0.05), panel.border = element_rect(fill=NA, colour='black')) + @@ -164,4 +164,4 @@ plot_time(12*12 + 12) ```{r,fig.height=6,fig.width=6,interval=1} #August 1997 strong El Niño event plot_time(27*12 + 8) -``` +``` \ No newline at end of file diff --git a/inst/doc/EFDR_SST.html b/inst/doc/EFDR_SST.html index 83fe36c..d321af2 100755 --- a/inst/doc/EFDR_SST.html +++ b/inst/doc/EFDR_SST.html @@ -94,7 +94,7 @@

The Code

parallel = detectCores()/2 # use half the number of available cores

The only unusual things here are the setting of n1 and n2 which denote the image size. We will be using a 64 x 32 image as the SST data is at a similar resolution in the area of interest. The registerDoParallel() command allocates half of your machine’s cores to the following foreach operation, which treats each SST time point individually. NB: The following can take a long time and it produces quite a lot of output. On a 64 core machine, this took about 10 minutes.

### Carry out EFDR on every frame
-cl <- makeCluster(parallel)
+cl <- makeCluster(2)
 registerDoParallel(cl)
 
 X <- foreach (t = unique(sst$time),.combine=rbind) %dopar% {
@@ -119,7 +119,7 @@ 
Results

   subset(long > 80 & long < 340)
 
 base_plot <- ggplot() + 
-      theme(panel.background = element_rect(fill='white', colour='black'),
+    theme(panel.background = element_rect(fill='white', colour='black'),
              text = element_text(size=20),
              panel.grid.major =  element_line(colour = "light gray", size = 0.05),
              panel.border  = element_rect(fill=NA, colour='black')) + 
@@ -159,19 +159,19 @@ 
Results

 
Here, instead, we will simply plot some of the interesting time points. Occurrences of El Niño and La Niña, and the respective strength, are given in http://ggweather.com/enso/oni.htm. The video (not shown) gives a clear indication when these have occurred. The time points which stand-out correspond to those given in the cited website, although on one occasion (1999–2000) the EFDR-smoothed plot did not show the La Niña event in the original SST data. We plot some of the detected anomalies below:

 
#August 1972 strong El Niño event
 plot_time(2*12 + 8)

+
plot of chunk unnamed-chunk-7

 
#January 1974 strong La Niña event
 plot_time(4*12 + 1)

+
plot of chunk unnamed-chunk-8

 
#November 1975 strong La Niña event
 plot_time(5*12 + 11)

+
plot of chunk unnamed-chunk-9

 
#December 1982 strong El Niño event
 plot_time(12*12 + 12)

+
plot of chunk unnamed-chunk-10

 
#August 1997 strong El Niño event
 plot_time(27*12 + 8)

-
-
-
+
plot of chunk unnamed-chunk-11

 
 
 
diff --git a/vignettes/EFDR_documents.Rmd b/vignettes/EFDR_documents.Rmd
index d421f82..433559e 100755
--- a/vignettes/EFDR_documents.Rmd
+++ b/vignettes/EFDR_documents.Rmd
@@ -7,10 +7,9 @@ vignette: >
 ---
   
 Enhanced False Discovery Rate (EFDR) tutorials
-
 ===================
   **Andrew Zammit Mangion**
   
-  **04-Dec-2014**
+  **14-Jan-2015**
   
-  Since the main function of this package `test.efdr` is computationally intensive, we do not distribute vignettes so as not to bog down CRAN servers during vignette building. Instead we provide two tutorials as documentation with the package. These can be accessed through the link *User guides, package vignettes and other documentation* in the HTML help page which can be loaded by typing `help(package = "EFDR")`. 
\ No newline at end of file
+  Since the main function of this package `test.efdr` is computationally intensive, we do not distribute vignettes so as not to bog down CRAN servers during vignette building. Instead we provide two tutorials as documentation with the package. These can be accessed through the link *User guides, package vignettes and other documentation* in the HTML help page which can be loaded by typing `help(package="EFDR",help_type="html")` in the console. 
\ No newline at end of file