@@ -4,5 +4,22 @@

 ^\.astylerc$

 ^.*\.Rproj$

 ^\.Rproj\.user$

+

 ^inst/benchmarks/

-^inst/profiling/

\ No newline at end of file

+^inst/profiling/

+^inst/vignettes/

+

+^src/Algorithms.o

+^src/AtomicSupport.o

+^src/Cogaps.o

+^src/GibbsSampler.o

+^src/Matrix.o

+^src/Random.o

+^src/RcppExports.o

+^src/test-runner.o

+^src/cpp_tests/testAlgorithms.o

+^src/cpp_tests/testAtomicSupport.o

+^src/cpp_tests/testGibbsSampler.o

+^src/cpp_tests/testMatrix.o

+^src/cpp_tests/testRandom.o

+^src/cpp_tests/testSerialization.o

\ No newline at end of file

@@ -1,26 +1,10 @@

 # options file for running astyle to "beautify" c++ code

 # run using `astyle --options=.astylerc src/*.{cc,cpp,h}`

-# don't create suffixes

-suffix=none

-

-# use the google style of braces

-style=google

-

 # use 4 spaces instead of tabs

 indent=spaces=4

 convert-tabs

-# indent 'class' and 'switch' blocks

-indent-classes

-indent-switches

-

-# empty lines around blocks, labels, class, closing headers, ...

-break-blocks=all

-

-# spaces around operator

-pad-oper

-

 # remove spaces in and around parenthesis except in front of headers

 # Example: foo ( bar, baz ); -> foo(bar, baz);

 unpad-paren

@@ -28,9 +12,6 @@ unpad-paren

 # insert a space after if, while, for, etc (exception to previous rule)

 pad-header

-# delete empty lines not created by `break-blocks=all`

-delete-empty-lines

-

 # pointer and reference operators to name

 align-pointer=name

@@ -41,9 +22,5 @@ add-brackets

 # use linux line-endings

 lineend=linux

-# why not keep the date the same?

-# although the date is modified by micro seconds to force compile

-preserve-date

-

 # restrict lines to 80 characters

 max-code-length=120

@@ -4,15 +4,7 @@

 # ignore compiled files

 *.so

-*.o*

-*.sw?

-*.dll

-

-# ignore R data

-*.RDS

-

-# ignore PDFs

-*.pdf

+*.o

 # ignore OS X files

 *.DS_Store

@@ -2,9 +2,8 @@ Package: CoGAPS

 Version: 2.99.0

 Date: 2014-08-23

 Title: Coordinated Gene Activity in Pattern Sets

-Author: Wai-shing Lee, Conor Kelton, Ondrej Maxian, Jacob Carey, Genevieve

-    Stein-O'Brien, Michael Considine, John Stansfield, Shawn Sivy, Carlo Colantuoni,

-    Alexander Favorov, Mike Ochs, Elana Fertig

+Author: Thomas Sherman, Wai-shing Lee, Conor Kelton, Ondrej Maxian,

+    Jacob Carey, Genevieve Stein-O'Brien, Michael Considine, John Stansfield, Shawn Sivy, Carlo Colantuoni, Alexander Favorov, Mike Ochs, Elana Fertig

 Description: Coordinated Gene Activity in Pattern Sets (CoGAPS)

     implements a Bayesian MCMC matrix factorization algorithm,

     GAPS, and links it to gene set statistic methods to infer biological

@@ -27,8 +26,6 @@ Imports:

     utils,

     doParallel,

     foreach,

-    iterators,

-    parallel,

     ggplot2,

     reshape2

 Suggests:

@@ -51,7 +51,6 @@

 #'@param fixedPatterns matrix of fixed values in either A or P matrix

 #'@param whichMatrixFixed character to indicate whether A or P matrix

 #'  contains the fixed patterns

-#'@param checkpoint_file_name name of file to store checkpoint

 #'@param checkpoint_interval time (in seconds) between cogaps checkpoints

 #'@export

@@ -142,7 +141,14 @@ gapsRun <- function(D, S, ABins = data.frame(), PBins = data.frame(),

     return(cogapResult);

 }

+# TODO: remove dependency on D,S matrices

+#' Restart Run from Checkpoint File

 #' @export

+#'

+#' @param D data matrix

+#' @param S uncertainty matrix

+#' @param path path to checkpoint file

+#' @return list containing information about A and P matrix

 gapsRunFromCheckpoint <- function(D, S, path)

 {

     # call to C++ Rcpp code

deleted file mode 100644

@@ -1,28 +0,0 @@

-# get directory of script

-initial.options <- commandArgs(trailingOnly = FALSE)

-file.arg.name <- "--file="

-script.name <- sub(file.arg.name, "", initial.options[grep(file.arg.name, initial.options)])

-script.basename <- dirname(script.name)

-

-# load packages

-library(microbenchmark)

-library(CoGAPS)

-

-# load benchmarks

-source(paste(sep="/", script.basename, "cogaps.R"))

-

-# display package version

-print(packageVersion('CoGAPS'))

-

-# run benchmarks

-data(GIST_TS_20084)

-nIter <- 3000

-print(runCogapsBenchmark(

-    D=GIST.D[1:75,],

-    S=GIST.S[1:75,],

-    nEquil=nIter,

-    nSample=nIter,

-    nFactor=7,

-    seed=12345,

-    reps=10

-))

\ No newline at end of file

deleted file mode 100644

@@ -1,13 +0,0 @@

-library(CoGAPS)

-library(microbenchmark)

-

-runCogapsBenchmark <- function(D, S, nEquil, nSample, nFactor, seed, reps)

-{

-    return(microbenchmark(gapsRun(D, S,

-        nEquil=nEquil,

-        nSample=nSample,

-        nFactor=nFactor,

-        seed=seed,

-        messages=FALSE),

-    times=reps))

-}

\ No newline at end of file

@@ -71,8 +71,6 @@ domain for relative probabilities}

 contains the fixed patterns}

 \item{checkpoint_interval}{time (in seconds) between cogaps checkpoints}

-

-\item{checkpoint_file_name}{name of file to store checkpoint}

 }

 \description{

 \code{gapsRun} calls the C++ MCMC code and performs Bayesian

new file mode 100644

@@ -0,0 +1,22 @@

+% Generated by roxygen2: do not edit by hand

+% Please edit documentation in R/gapsRun.R

+\name{gapsRunFromCheckpoint}

+\alias{gapsRunFromCheckpoint}

+\title{Restart Run from Checkpoint File}

+\usage{

+gapsRunFromCheckpoint(D, S, path)

+}

+\arguments{

+\item{D}{data matrix}

+

+\item{S}{uncertainty matrix}

+

+\item{path}{path to checkpoint file}

+}

+\value{

+list containing information about A and P matrix

+}

+\description{

+Restart Run from Checkpoint File

+}

+

@@ -144,7 +144,7 @@ AtomicProposal AtomicSupport::makeProposal() const

     }

     float unif = gaps::random::uniform();

-    if ((mNumAtoms < 2 && unif <= 0.6667) || unif <= 0.5) // birth/death

+    if ((mNumAtoms < 2 && unif <= 0.6667f) || unif <= 0.5f) // birth/death

     {

         if (mNumAtoms >= mMaxNumAtoms)

         {

@@ -8,7 +8,6 @@

 #include <fstream>

 #include <vector>

 #include <stdint.h>

-//#include <boost/serialization/map.hpp>

 struct AtomicProposal

 {

@@ -31,6 +30,18 @@ struct AtomicProposal

     {}

 };

+struct Atom

+{

+    uint64_t pos;

+    float mass;

+};

+

+struct AtomNeighbors

+{

+    Atom left;

+    Atom right;

+};

+

 class AtomicSupport 

 {

 private:

@@ -43,6 +54,8 @@ public:

     char mLabel;

     // storage of the atomic domain

+    //std::vector<Atom> mAtoms;

+    //std::map<uint64_t, uint64_t> mAtomPositions;

     std::map<uint64_t, float> mAtomicDomain;

     uint64_t mNumAtoms;

     uint64_t mMaxNumAtoms;

@@ -62,6 +75,11 @@ public:

     // expected magnitude of each atom (must be > 0)

     float mLambda;

+    // functions for dealing with atomic data structure

+    //void addAtom(Atom atom); // O(logN)

+    //void removeAtom(Atom atom); // O(logN)

+    //AtomNeighbors getNeighbors(uint64_t pos); // O(logN)

+

     // convert atomic position to row/col of the matrix

     uint64_t getRow(uint64_t pos) const;

     uint64_t getCol(uint64_t pos) const;

@@ -24,7 +24,7 @@ static bpt::ptime lastCheckpoint;

 static void createCheckpoint(GapsInternalState &state)

 {

     state.numCheckpoints++;

-    std::cout << "creating gaps checkpoint...";

+    Rcpp::Rcout << "creating gaps checkpoint...";

     bpt::ptime start = bpt::microsec_clock::local_time();

     Archive ar("gaps_checkpoint_" + SSTR(state.numCheckpoints) + ".out",

         ARCHIVE_WRITE);

@@ -39,7 +39,7 @@ static void createCheckpoint(GapsInternalState &state)

     ar.close();

     bpt::time_duration diff = bpt::microsec_clock::local_time() - start;

     double elapsed = diff.total_milliseconds() / 1000.;

-    std::cout << " finished in " << elapsed << " seconds\n";

+    Rcpp::Rcout << " finished in " << elapsed << " seconds\n";

 }

 static void runGibbsSampler(GapsInternalState &state, unsigned nIterTotal,

@@ -92,7 +92,7 @@ Vector &chi2Vec, Vector &aAtomVec, Vector &pAtomVec)

             if ((state.iter + 1) % state.nOutputs == 0 && state.messages)

             {

                 std::string temp = state.phase == GAPS_CALIBRATION ? "Equil: " : "Samp: ";

-                std::cout << temp << state.iter + 1 << " of " << nIterTotal

+                Rcpp::Rcout << temp << state.iter + 1 << " of " << nIterTotal

                     << ", Atoms:" << aAtomVec[state.iter] << "(" << pAtomVec[state.iter]

                     << ") Chi2 = " << state.sampler.chi2() << '\n';

             }

@@ -187,7 +187,7 @@ Rcpp::List cogapsFromCheckpoint(const std::string &fileName)

     ar >> magicNum;

     if (magicNum != ARCHIVE_MAGIC_NUM)

     {

-        std::cout << "invalid checkpoint file" << std::endl;

+        Rcpp::Rcout << "invalid checkpoint file" << std::endl;

         return Rcpp::List::create();

     }

@@ -1,4 +1,4 @@

-PKG_CPPFLAGS = -Wall -Wextra -O2

+PKG_CPPFLAGS = -DBOOST_MATH_PROMOTE_DOUBLE_POLICY=0

 OBJECTS =   Algorithms.o \

             AtomicSupport.o \

             Cogaps.o \

@@ -6,8 +6,6 @@

 #include <Rcpp.h>

 #include <vector>

-// use CRTP for different matrix types?

-

 // temporary: used for testing performance of float vs double

 typedef float matrix_data_t;

@@ -7,20 +7,15 @@

 #include <boost/random/normal_distribution.hpp>

 #include <boost/random/poisson_distribution.hpp>

 #include <boost/random/exponential_distribution.hpp>

-

 #include <boost/random/mersenne_twister.hpp>

-// need -O0 to run in valgrind, -O2 for performance

-#pragma GCC push_options

-#pragma GCC optimize ("O2")

 #include <boost/math/distributions/normal.hpp>

 #include <boost/math/distributions/exponential.hpp>

 #include <boost/math/distributions/gamma.hpp>

-#pragma GCC pop_options

 #include <stdint.h>

-#define Q_GAMMA_THRESHOLD 1E-6

+#define Q_GAMMA_THRESHOLD 0.000001f

 #define Q_GAMMA_MIN_VALUE 0.0

 typedef boost::random::mt19937 RNGType;

@@ -8,6 +8,8 @@

  *  Distributed under the Boost Software License, Version 1.0. (See accompanying

  *  file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)

  */

+#include <Rcpp.h>

+

 #ifndef TWOBLUECUBES_SINGLE_INCLUDE_CATCH_HPP_INCLUDED

 #define TWOBLUECUBES_SINGLE_INCLUDE_CATCH_HPP_INCLUDED

@@ -7705,10 +7707,10 @@ namespace Catch {

 #ifndef CATCH_CONFIG_NOSTDOUT // If you #define this you must implement these functions

     std::ostream& cout() {

-        return std::cout;

+        return Rcpp::Rcout;

     }

     std::ostream& cerr() {

-        return std::cerr;

+        return Rcpp::Rcerr;

     }

     std::ostream& clog() {

         return std::clog;