Merge pull request #5 from snoweye/master

snoweye · snoweye · commit c4884a32df70 · 2014-02-02T14:26:19.000-08:00
0.1-6 to CRAN
diff --git a/ChangeLog b/ChangeLog
@@ -1,3 +1,6 @@
+2014-02-02: Ver. 0.1-6
+  * Add I/O demo to read data and dump clustering results.
+
 2013-07-03: Ver. 0.1-6
   * Move inst/doc/vignettes/ to vignettes/.
 
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,10 +1,10 @@
 Package: pmclust
 Version: 0.1-6
-Date: 2014-01-09
+Date: 2014-02-02
 Title: Parallel Model-Based Clustering
 Authors@R: c(person("Wei-Chen", "Chen", role = c("aut", "cre"), email =
         "wccsnow@gmail.com"), person("George", "Ostrouchov", role = "aut"))
-Depends: R (>= 3.0.0), methods, rlecuyer, pbdMPI (>= 0.1-7)
+Depends: R (>= 3.0.0), methods, rlecuyer, pbdMPI (>= 0.2-2), MASS
 Enhances: MixSim, pbdSLAP (>= 0.1-7), pbdBASE (>= 0.3-0), pbdDMAT (>=
         0.2-4)
 LazyLoad: yes
@@ -25,7 +25,7 @@ URL: http://r-pbd.org/
 BugReports: http://group.r-pbd.org/
 MailingList: Please send questions and comments regarding pbdR to
         RBigData@gmail.com
-Packaged: 2014-01-10 05:28:26 UTC; snoweye
+Packaged: 2014-02-02 22:05:31 UTC; snoweye
 Author: Wei-Chen Chen [aut, cre],
   George Ostrouchov [aut]
 Maintainer: Wei-Chen Chen <wccsnow@gmail.com>
diff --git a/NAMESPACE b/NAMESPACE
@@ -1,6 +1,6 @@
 useDynLib(pmclust)
 
-import(methods, rlecuyer, pbdMPI)
+import(methods, rlecuyer, pbdMPI, MASS)
 
 # exportPattern("^[[:alpha:]]+")
 # exportPattern("^[\\.]")
diff --git a/demo/00Index b/demo/00Index
@@ -17,3 +17,4 @@ gbd_kmeans      pmclust         An GBD example of K-means algorithm.
 iris_dmat       pmclust         iris example in ddmatrix.
 iris_gbdr       pmclust         iris example in GBD row-major format.
 
+io_csv          pmclust         An example of simple data input and output.
diff --git a/demo/io_csv.r b/demo/io_csv.r
@@ -0,0 +1,35 @@
+### Setup environment.
+library(pmclust, quietly = TRUE)
+comm.set.seed(123, diff = TRUE)
+if(comm.size() != 4){
+  comm.stop("4 processors are needed.")
+}
+
+### Generate an example data.
+N <- 400
+p <- 2
+K <- 2
+data.spmd <- generate.MixSim(N, p, K, MaxOmega = 0.001)
+
+### Dump fake data to a csv file.
+X.df.org <- as.data.frame(cbind(data.spmd$X.spmd, data.spmd$CLASS.spmd))
+colnames(X.df.org) <- c(paste("V", 1:p, sep = ""), "ID")
+comm.write.csv(X.df.org, file = "toys_org.csv", row.names = FALSE)
+
+### Read data from the csv file.
+X.df.new <- comm.read.csv("toys_org.csv")
+X.spmd <- as.matrix(X.df.new[, 1:p])
+
+### Run clustering.
+PARAM.org <- set.global(K = K)          # Set global storages.
+PARAM.org <- initial.em(PARAM.org)      # One initial.
+PARAM.new <- apecma.step(PARAM.org)     # Run APECMa.
+em.update.class()                       # Get classification.
+
+### Dump data and clustering results to a new csv file.
+X.df.new$ID.est <- .pmclustEnv$CLASS.spmd
+comm.write.csv(X.df.new, file = "toys_new.csv", row.names = FALSE)
+
+### Quit.
+finalize()
+
diff --git a/inst/CITATION b/inst/CITATION
@@ -16,7 +16,7 @@ citEntry(
   author = personList(as.person("W.-C. Chen"), as.person("G. Ostrouchov")),
   year = "2012",
   note = "R Vignette, URL http://cran.r-project.org/package=pmclust",
-  textVersion = "W.-C. Chen, G. Ostrouchov (2012). A Quick Guide for the pmclust Package. R Vignette, URL http://cran.r-project.org/package=pmclust",
+  textVersion = "W.-C. Chen, G. Ostrouchov (2014). A Quick Guide for the pmclust Package. R Vignette, URL http://cran.r-project.org/package=pmclust",
   key = "Chen2012pmclustvignette"
 )
 
diff --git a/inst/doc/pmclust-guide.R b/inst/doc/pmclust-guide.R
diff --git a/inst/doc/pmclust-guide.Rnw b/inst/doc/pmclust-guide.Rnw
@@ -2,7 +2,6 @@
 
 \documentclass[article,shortnames,nojss]{pmclust-include/jss}
 
-
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 %% declarations for jss.cls %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
@@ -15,9 +14,9 @@ Wei-Chen Chen$^1$ and George Ostrouchov$^{1,2}$
 \end{minipage}
 \\
 \\
-$^1$Computer Science and Mathematics Division, \\
-Oak Ridge National Laboratory, \\
-Oak Ridge, TN, USA \\
+$^1$Department of Ecology and Evolutionary Biology, \\
+University of Tennessee, \\
+Knoxville, TN, USA \\
 \\
 $^2$Remote Data Analysis and Visualization Center,\\
 University of Tennessee, \\
diff --git a/inst/doc/pmclust-guide.pdf b/inst/doc/pmclust-guide.pdf
diff --git a/vignettes/pmclust-guide.Rnw b/vignettes/pmclust-guide.Rnw
@@ -2,7 +2,6 @@
 
 \documentclass[article,shortnames,nojss]{pmclust-include/jss}
 
-
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 %% declarations for jss.cls %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
@@ -15,9 +14,9 @@ Wei-Chen Chen$^1$ and George Ostrouchov$^{1,2}$
 \end{minipage}
 \\
 \\
-$^1$Computer Science and Mathematics Division, \\
-Oak Ridge National Laboratory, \\
-Oak Ridge, TN, USA \\
+$^1$Department of Ecology and Evolutionary Biology, \\
+University of Tennessee, \\
+Knoxville, TN, USA \\
 \\
 $^2$Remote Data Analysis and Visualization Center,\\
 University of Tennessee, \\
diff --git a/vignettes/pmclust-include/01-introduction.tex b/vignettes/pmclust-include/01-introduction.tex
@@ -50,7 +50,7 @@
 Other operating systems such as Mac or Windows are also possible to run the
 \pkg{pmclust} if MPI systems and \pkg{pbdMPI}~\citep{Chen2012pbdMPIpackage}
 are correctly installed, and the instruction can be found in
-the vignette of \pkg{pbdMPI}~\citep{Chen2012pbdMPIvignette}.
+the vignette of \pkg{pbdMPI}~\citep{Chen2014pbdMPIvignette}.
 For master and workers models,
 it is also possible to run the \pkg{pmclust} within \pkg{Rmpi}~\citep{Yu2010}
 via \pkg{pbdMPI}.
diff --git a/vignettes/pmclust-include/02-example.tex b/vignettes/pmclust-include/02-example.tex
@@ -1,6 +1,6 @@
-\section[Examples]{Example}
+\section[Examples]{Examples}
 \label{sec:example}
-\addcontentsline{toc}{section}{\thesection. Example}
+\addcontentsline{toc}{section}{\thesection. Examples}
 
 The \pkg{pmclust} is a package of \proglang{R}~\citep{Rcore} and
 designed in the single program multiple data (SPMD)
@@ -218,3 +218,29 @@
 \end{tabular}
 \end{table}
 
+
+
+
+\subsection[Input and Output via CSV Files]{Input and Output via CSV Files}
+\label{sec:data_io}
+\addcontentsline{toc}{subsection}{\thesubsection. Input and Output via CSV Files}
+
+Since \pkg{pbdMPI} version 0.2-2, there are serveral simple data input
+and output methods implemented either for text or csv files in distributed
+manner (spmd/gbd). Therefore, through csv files, we may simply read data from
+disk and dump clustering results for further analysis. A demo example is
+in \pkg{pmclust} as next.
+\begin{Command}
+mpiexec -np 4 Rscript -e "demo(io_csv,'pmclust',ask=F,echo=F)"
+\end{Command}
+
+We use four processors to generate 400 samples (100 per processor)
+in two dimension and two clusters by
+\pkg{MixSim}~\citep{Melnykov2012}. The fake data with true classification IDs
+are dump to a csv file named \code{toy_org.csv}.
+Then, we read data from the file back to memory, use \pkg{pmclust} to
+cluster them, and dump the data with new classification IDs to a new csv
+file named \code{toy_new.csv}.
+Similar input and output process can be replaced by other file formats,
+data streams, or databases.
+
diff --git a/vignettes/pmclust-include/pmclust.bib b/vignettes/pmclust-include/pmclust.bib
@@ -109,11 +109,11 @@ @MISC{Chen2012pbdMPIpackage
   note = {R Package},
   url = {http://cran.r-project.org/package=pbdMPI}
 }
-@MANUAL{Chen2012pbdMPIvignette,
+@MANUAL{Chen2014pbdMPIvignette,
   author = {Chen, W.-C. and Ostrouchov, G. and Schmidt, D. and Patel, P. and
             Yu, H.},
-  title = {A Quick Guide for the pbdMPI package},
-  year = {2012},
+  title = {A Quick Guide for the pbdMPI package (Ver. 0.2-2)},
+  year = {2014},
   note = {R Vignette},
   url = {http://cran.r-project.org/package=pbdMPI}
 }

Original file line number	Diff line number	Diff line change
`@@ -17,3 +17,4 @@ gbd_kmeans pmclust An GBD example of K-means algorithm.`
`17`	`17`	`iris_dmat pmclust iris example in ddmatrix.`
`18`	`18`	`iris_gbdr pmclust iris example in GBD row-major format.`
`19`	`19`
	`20`	`+io_csv pmclust An example of simple data input and output.`
Original file line number	Diff line number	Diff line change
`@@ -16,7 +16,7 @@ citEntry(`
`16`	`16`	`author = personList(as.person("W.-C. Chen"), as.person("G. Ostrouchov")),`
`17`	`17`	`year = "2012",`
`18`	`18`	`note = "R Vignette, URL http://cran.r-project.org/package=pmclust",`
`19`		`- textVersion = "W.-C. Chen, G. Ostrouchov (2012). A Quick Guide for the pmclust Package. R Vignette, URL http://cran.r-project.org/package=pmclust",`
	`19`	`+ textVersion = "W.-C. Chen, G. Ostrouchov (2014). A Quick Guide for the pmclust Package. R Vignette, URL http://cran.r-project.org/package=pmclust",`
`20`	`20`	`key = "Chen2012pmclustvignette"`
`21`	`21`	`)`
`22`	`22`