Dev

R/RcppExports.R

@@ -1,14 +1,6 @@
 # Generated by using Rcpp::compileAttributes() -> do not edit by hand
 # Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
 
-WritePhylipFile <- function(xPosition, yPosition, data, cutoff, countTable, saveLocation) {
-    invisible(.Call('_clustur_WritePhylipFile', PACKAGE = 'clustur', xPosition, yPosition, data, cutoff, countTable, saveLocation))
-}
-
-WriteColumnFile <- function(xPosition, yPosition, data, cutoff, countTable, saveLocation) {
-    invisible(.Call('_clustur_WriteColumnFile', PACKAGE = 'clustur', xPosition, yPosition, data, cutoff, countTable, saveLocation))
-}
-
 DetermineIfPhylipOrColumnFile <- function(filePath) {
     .Call('_clustur_DetermineIfPhylipOrColumnFile', PACKAGE = 'clustur', filePath)
 }
@@ -37,7 +29,7 @@ OptiCluster <- function(DistanceData, featureColumnName, binColumnName, cutoff)
     .Call('_clustur_OptiCluster', PACKAGE = 'clustur', DistanceData, featureColumnName, binColumnName, cutoff)
 }
 
-CreateDataFrameFromSparse <- function(countTable) {
-    .Call('_clustur_CreateDataFrameFromSparse', PACKAGE = 'clustur', countTable)
+CreateDataFrameFromSparseCountTable <- function(countTable) {
+    .Call('_clustur_CreateDataFrameFromSparseCountTable', PACKAGE = 'clustur', countTable)
 }
 

R/cluster.R

@@ -193,7 +193,8 @@ read_count <- function(count_table_path) {
   if (grepl("#", test_read[1, 1], fixed = TRUE)) {
     count_table_sparse <- read.delim(count_table_path, sep = "\t", skip = 2)
     count_table_sparse <- lapply(count_table_sparse, as.character)
-    return(validate_count_table(CreateDataFrameFromSparse(count_table_sparse)))
+    ct <- CreateDataFrameFromSparseCountTable(count_table_sparse)
+    return(validate_count_table(ct))
   }
   return(validate_count_table(read.delim(count_table_path, sep = "\t")))
 }

README.Rmd

@@ -63,4 +63,3 @@ To get started, look at the ["Getting started"](www.schlosslab.org/clustur/) pag
 
 For feature request please visit the issue page in github and use the feature
 tag. If you would like to contribute pull request are welcomed!
-

README.md

@@ -44,7 +44,7 @@ library(clustur)
 ## Getting Started
 
 To get started, look at the [“Getting
-started”](https:://www.schlosslab.org/clustur/) page.
+started”](www.schlosslab.org/clustur/) page.
 
 ## Contributions
 

src/Adapters/CountTableAdapter.h

@@ -12,7 +12,7 @@
 #include <queue>
 #include <set>
 
-#include "../RowData.h"
+
 
 class CountTableAdapter {
 public:
@@ -29,7 +29,6 @@ class CountTableAdapter {
     std::vector<std::string> GetGroups() const;
     Rcpp::DataFrame GetCountTable() const {return countTable;}
     Rcpp::DataFrame ReCreateDataFrame() const;
-    std::set<std::string> CheckDistanceFileOnlyHasNamesInCount(const std::vector<RowData>&) const;
 private:
     void CreateNameToIndex();
     struct IndexAbundancePair {

src/Adapters/DistanceFileReader.h

@@ -9,7 +9,6 @@
 
 #include "CountTableAdapter.h"
 #include "../MothurDependencies/ListVector.h"
-#include "../RowData.h"
 #include "../MothurDependencies/SparseDistanceMatrix.h"
 
 class DistanceFileReader {
@@ -18,12 +17,11 @@ class DistanceFileReader {
     // We need to deduce type, the easy way to do that is to see if there is a number for the first item read.
     // Phylip files have a number of sequences located at the top. We can use that to our advantage.
     virtual bool Read(const std::string& filePath) {return false;}
-    DistanceFileReader(SparseDistanceMatrix*, ListVector*, bool);
+    DistanceFileReader(SparseDistanceMatrix*, ListVector*, double, bool);
     DistanceFileReader() = default;
     // Phylip files do not need a count table
     SparseDistanceMatrix* GetSparseMatrix() const {return new SparseDistanceMatrix(*sparseMatrix);}
     ListVector* GetListVector() const {return new ListVector(*list);}
-    void SetRowDataMatrix(const std::vector<RowData>& data);
     void SetCountTable(CountTableAdapter data);
     Rcpp::DataFrame SparseMatrixToDataFrame() const;
     Rcpp::DataFrame GetCountTable() const;

src/Adapters/MatrixAdapter.h

@@ -15,8 +15,6 @@ class MatrixAdapter {
     MatrixAdapter(const std::vector<int> &iIndexes, const std::vector<int> &jIndexes,
         const std::vector<double> &dataValues, double cutOff, bool isSimularity, CountTableAdapter table);
     ~MatrixAdapter() = default;
-    bool CreatePhylipFile(const std::string &saveFileLocation);
-    bool CreateColumnDataFile(const std::string &saveFileLocation);
     SparseDistanceMatrix CreateSparseMatrix();
     ListVector CreateListVector() const;
 

src/Adapters/OptimatrixAdapter.h

@@ -15,9 +15,6 @@
 #include <unordered_map>
 #include <sstream>
 
-#include "../RowData.h"
-
-
 class OptimatrixAdapter {
     ///Closeness is the data represntation, it is a vector of sets that tells us which values are closes to it ->
     /// So, close[0], if Namemap[0] = A, then close[0] is the values that are within the cutoff to A.
@@ -28,14 +25,8 @@ class OptimatrixAdapter {
     explicit OptimatrixAdapter(const double cutoff) {
         this->cutoff = cutoff;
     }
-    OptiMatrix* ConvertToOptimatrix(const std::vector<int>
-        &xPosition, const std::vector<int>& yPosition, const std::vector<double>& data);
-    OptiMatrix* ConvertToOptimatrix(const std::vector<RowData>&, bool);
     OptiMatrix* ConvertToOptimatrix(const SparseDistanceMatrix* matrixData,
     const ListVector* listVector, bool sim);
-    std::vector<std::unordered_set<long long>> GetCloseness() {return {};}
-    std::vector<std::string> GetNameList() {return {};}
-    std::vector<std::string> GetSingletons() {return {};}
 private:
     double cutoff;
     //TODO Get rid of values inside of the adapter, it should not care about them

src/ClusterCommand.cpp

@@ -84,10 +84,7 @@ ClusterExport* ClusterCommand::runOptiCluster(OptiMatrix *optiMatrix, const doub
         }
         util.AddRowToDataFrameMap(dataframeMapClusterMetrics, clusterMetrics, clusterMetricsHeaders);
         std::chrono::time_point<std::chrono::system_clock> start, end;
-        //m->mothurOutEndLine();
-        // Stable Metric -> Keep the data stable, to prevent errors (rounding errors)
-        // The difference between what the current and last metric (delta)
-        // MaxIters -> is an exit condition
+
         while ((delta > stableMetric) && (iters < maxIters)) {
             //long start = std::time(nullptr);
             double oldMetric = listVectorMetric;
@@ -126,7 +123,6 @@ ClusterExport* ClusterCommand::runOptiCluster(OptiMatrix *optiMatrix, const doub
         clusterInformation.clusterBins = list->print(listFile);
         data->AddToData(clusterInformation);
         data->SetListVector(*list, std::to_string(cutoff));
-        // clusterMatrixOutput = list->print(listFile);
         stats = cluster.getStats(tp, tn, fp, fn);
 
         sensFile += std::to_string(cutoff) + ',' + std::to_string(cutoff) + ',' + std::to_string(tp) + ',' +
@@ -157,7 +153,7 @@ ClusterExport* ClusterCommand::runMothurCluster(const std::string &clusterMethod
     float previousDist = 0.00000;
     float rndPreviousDist = 0.00000;
     oldList = *list;
-    bool printHeaders = false;
+    constexpr bool printHeaders = false;
     std::string clusterResult;
     double highestDistLabel =  -1;
     std::string binResults;

src/CountTableAdapter.cpp

@@ -147,16 +147,6 @@ Rcpp::DataFrame CountTableAdapter::ReCreateDataFrame() const {
     return countTable;
 }
 
-std::set<std::string> CountTableAdapter::CheckDistanceFileOnlyHasNamesInCount(const std::vector<RowData>& data) const {
-    std::set<std::string> names;
-    for(const auto& value : data) {
-        if(nameToRowIndex.find(value.name) == nameToRowIndex.end()) {
-            names.insert(value.name);
-        }
-    }
-    return names;
-}
-
 
 void CountTableAdapter::CreateNameToIndex() {
     for(size_t i = 0; i < sampleNames.size(); i++) {

src/DistanceFileReader.cpp

@@ -6,7 +6,8 @@
 
 
 DistanceFileReader::DistanceFileReader(SparseDistanceMatrix *sparseDistanceMatrix,
-    ListVector *listVector, const bool isSim):sparseMatrix(sparseDistanceMatrix), list(listVector), sim(isSim){}
+    ListVector *listVector, const double cutoff, const bool isSim):sparseMatrix(sparseDistanceMatrix),
+list(listVector), cutoff(cutoff), sim(isSim){}
 
 Rcpp::DataFrame DistanceFileReader::SparseMatrixToDataFrame() const {
     const size_t size = sparseMatrix->seqVec.size();

src/MatrixAdapter.cpp

@@ -11,52 +11,6 @@ MatrixAdapter::MatrixAdapter(const std::vector<int> &iIndexes, const std::vector
                             data(dataValues) {
 }
 
-
-bool MatrixAdapter::CreatePhylipFile(const std::string &saveFileLocation) {
-    // if(saveFileLocation.empty())
-    //     return false;
-    // const auto matrix = DistanceMatrixToSquareMatrix();
-    // int count = 0;
-    // const size_t size = matrixNames.size();
-    // std::string distanceString = "\t" + std::to_string(size) + "\n";
-    // for (const auto &cells: matrix) {
-    //     distanceString += matrixNames[count];
-    //     for (int i = 0; i < count; i++) {
-    //         distanceString += "\t" + std::to_string(cells.rowValues[i]);
-    //     }
-    //     distanceString += "\n";
-    //     count++;
-    // }
-    // std::ofstream writeOut(saveFileLocation);
-    // if (!writeOut.is_open()) {
-    //     TestHelper::Print("Failed to open: \n");
-    // }
-    // writeOut << distanceString;
-    // writeOut.close();
-    return true;
-}
-
-bool MatrixAdapter::CreateColumnDataFile(const std::string &saveFileLocation) {
-    if(saveFileLocation.empty())
-        return false;
-    const auto matrix = CreateSparseMatrix();
-    std::string data;
-    for (size_t i = 0; i < matrix.seqVec.size(); i++) {
-        std::string firstCellName = matrixNames[i];
-        for(size_t j = 0; j < matrixNames[i].size(); j++) {
-            std::string otherCell = matrixNames[matrix.seqVec[i][j].index];
-            data += firstCellName + "\t" + otherCell + "\t" + std::to_string(matrix.seqVec[i][j].dist) + "\n";
-        }
-    }
-    std::ofstream writeOut(saveFileLocation);
-    if (!writeOut.is_open()) {
-        Rcpp::Rcout << "Failed to open: \n";
-    }
-    writeOut << data;
-    writeOut.close();
-    return true;
-}
-
 SparseDistanceMatrix MatrixAdapter::CreateSparseMatrix() {
     // The indexes are +1, i need to push them back so that 1 -> 0, 2-> 1, etc (name map maybe?)
     std::set<std::string> names;

src/MatrixAdapterTestFixture.cpp

@@ -11,13 +11,6 @@ bool MatrixAdapterTestFixture::TestCreateSparseMatrix(const size_t expectedResul
     return result == expectedResult;
 }
 
-bool MatrixAdapterTestFixture::TestCreatePhylipFile(const bool expectedResult) {
-    Setup();
-    const auto result = matrixAdapter->CreatePhylipFile("");
-    TearDown();
-    return result == expectedResult;
-}
-
 bool MatrixAdapterTestFixture::TestGetListVector(const bool createSparseFirst, const bool expectedResult) {
     Setup();
     if(createSparseFirst) matrixAdapter->CreateSparseMatrix();

src/MothurDependencies/ColumnDistanceMatrixReader.h

@@ -10,7 +10,6 @@
 #include <vector>
 #include "ListVector.h"
 #include "SparseDistanceMatrix.h"
-#include "../RowData.h"
 #include "../Adapters/CountTableAdapter.h"
 #include "../Adapters/DistanceFileReader.h"
 

src/MothurDependencies/OptiMatrix.h

@@ -15,13 +15,13 @@ class OptiMatrix final : public OptiData{
 
 public:
     OptiMatrix(const std::vector<std::unordered_set<long long>>& close,  const std::vector<std::string>& name,
-     const std::vector<std::string>& single, const double c): OptiData(c)
+     const std::vector<std::string>& singleton, const double c): OptiData(c)
 {
     closeness = close;
     nameMap = name;
-    singletons = single;
+    singletons = singleton;
 }//closeness, namemap, singleton, cutoff
-    ~OptiMatrix(){}
+    ~OptiMatrix() override = default;
     std::vector<std::unordered_set<long long>> GetCloseness() {return closeness;}
     std::vector<std::string> GetNameList() {return nameMap;}
     std::vector<std::string> GetSingletons() {return singletons;}

src/MothurDependencies/ReadPhylipMatrix.h

@@ -8,7 +8,6 @@
 #include <fstream>
 #include <string>
 #include <vector>
-#include "../RowData.h"
 #include <unordered_set>
 #include "ListVector.h"
 #include "SparseDistanceMatrix.h"

src/MothurDependencies/SharedFileBuilder.h

@@ -15,7 +15,6 @@
 
 #include "ClusterExport.h"
 #include "ListVector.h"
-#include "RSparseMatrix.h"
 #include "SharedFile.h"
 #include "OtuAbundancePair.h"
 #include "../Adapters/CountTableAdapter.h"

src/MothurDependencies/SparseDistanceMatrix.h

@@ -8,7 +8,7 @@
 #include "PDistCell.h"
 
 #include <vector>
-
+#include <Rcpp.h>
 #include "Utils.h"
 
 
@@ -44,6 +44,7 @@ class SparseDistanceMatrix {
     int addCellSorted(unsigned long, PDistCell);
     std::vector<std::vector<PDistCell> > seqVec;
     bool print() const;
+    void FilterSparseMatrix(float cutoff);
 
 private:
     PDistCell smallCell;				//The cell with the smallest distance

src/OptiClusterTestFixture.cpp

@@ -66,12 +66,26 @@ bool OptiClusterTestFixture::OptiClusterGetsTheCorrectNumberOfBins(const long lo
 
 
 void OptiClusterTestFixture::Setup() {
-    const auto xVals = std::vector<int>{0,0,0,1,1,2,3};
-    const auto yVals = std::vector<int>{1,2,4,2,4,4,4};
-    const auto data = std::vector<double>{0.02,0.04,0.025,0.01,0.028,0.045,0.05};
+    const auto xVals = std::vector<int>{0,0,0,1,1};
+    const auto yVals = std::vector<int>{1,2,4,2,4};
+    const auto data = std::vector<float>{0.02,0.04,0.025,0.01,0.028};
+    SparseDistanceMatrix sparse;
+    ListVector vec;
+    vec.push_back("0");
+    vec.push_back("1");
+    vec.push_back("2");
+    vec.push_back("3");
+    vec.push_back("4");
+    const size_t size = xVals.size();
+    sparse.resize(size);
+    for(size_t i = 0; i < size; i++) {
+        sparse.addCell(xVals[i], PDistCell(yVals[i], data[i]));
+    }
     OptimatrixAdapter adapter(0.03);
-    OptiData *matrix = adapter.ConvertToOptimatrix(xVals, yVals, data);
+    OptiData *matrix = adapter.ConvertToOptimatrix(&sparse, &vec, false);
     optiCluster = new OptiCluster(matrix, new MCC(), 0);;
+    double value = 0;
+    optiCluster->initialize(value, false, "");
 }
 
 void OptiClusterTestFixture::TearDown() {