Release polish (#6)

* Added to build ignore and properly documented the new R code. * Unstable, have to fix the mismatch header situation, but was able to return a vector with all the metrics! * Cluster now returns test metrics and added test to support it.
SchlossLab · Jun 14, 2024 · 77ebc1c · 77ebc1c
1 parent b717404
commit 77ebc1c
Show file tree

Hide file tree

Showing 12 changed files with 54 additions and 22 deletions.
diff --git a/.Rbuildignore b/.Rbuildignore
@@ -1 +1,3 @@
 ^\.github$
+^src/CMakeLists\.txt$
+^src/cmake-build-debug$
diff --git a/R/Cluster.R b/R/Cluster.R
@@ -12,16 +12,31 @@
 #' @param sparse_matrix A Sparse Matrix.
 #' @param cutoff A cutoff value
 #' @param iterations The number of iterations
+#' @param shuffle a boolean to determine whether or not you want to shuffle the data before you cluster
 #' @return A data.frame of the clusters.
 opti_cluster <- function(sparse_matrix, cutoff, iterations, shuffle = TRUE) {
   index_one_list <- sparse_matrix@i
   index_two_list <- sparse_matrix@j
   value_list <- sparse_matrix@x
-  clustering_output_string <- MatrixToOpiMatrixCluster(index_one_list, index_two_list, value_list, cutoff,
+  clustering_output_string_list <- MatrixToOpiMatrixCluster(index_one_list, index_two_list, value_list, cutoff,
                                                        iterations, shuffle)
-  df <- t(read.table(text = clustering_output_string,
+  clustering_output_string <- clustering_output_string_list[1]
+  clustering_metric <- clustering_output_string_list[2]
+  clustering_metric_2 <- clustering_output_string_list[3]
+  df_cluster_metrics <- (read.table(text = clustering_metric,
                      sep = "\t", header = TRUE))
-  df <- data.frame(df[-1, ])
-  colnames(df)[1] <- "cluster"
-  return(df)
+  df_other_cluster_metrics <-  (read.table(text = clustering_metric_2,
+                     sep = "\t", header = TRUE))
+
+  df_cluster <- t(read.table(text = clustering_output_string,
+                     sep = "\t", header = TRUE))
+  df_cluster <- data.frame(df_cluster[-1, ])
+
+  colnames(df_cluster)[1] <- "cluster"
+
+  opticluster_data <- list(cluster = df_cluster,
+                           cluster_metrics = df_cluster_metrics,
+                           other_cluster_metrics = df_other_cluster_metrics)
+
+  return(opticluster_data)
 }
diff --git a/man/opti_cluster.Rd b/man/opti_cluster.Rd
diff --git a/src/ClusterCommand.cpp b/src/ClusterCommand.cpp
@@ -16,17 +16,17 @@ ClusterCommand::~ClusterCommand() {
 /// Bad allocations, returns basic_string, returns empty string, returns non-utf8 characters, etc
 /// @param optiMatrix
 /// @return
-std::string ClusterCommand::runOptiCluster(OptiMatrix *optiMatrix) {
+std::vector<std::string> ClusterCommand::runOptiCluster(OptiMatrix *optiMatrix) {
     std::string clusterMetrics;
     std::string sensFile;
     std::string outStep;
     std::string clusterMatrixOutput;
     if (!cutOffSet) {
-        clusterMetrics += ("\nYou did not set a cutoff, using 0.03.\n");
+        // clusterMetrics += ("\nYou did not set a cutoff, using 0.03.\n");
         cutoff = 0.05;
     }
 
-    clusterMetrics += ("\nClustering " + distfile + "\n");
+    // clusterMetrics += ("\nClustering " + distfile + "\n");
 
     ClusterMetric *metric = nullptr;
     metricName = "mcc";
@@ -39,7 +39,7 @@ std::string ClusterCommand::runOptiCluster(OptiMatrix *optiMatrix) {
             metricName == "accuracy") { metric = new Accuracy(); } else if (
             metricName == "ppv") { metric = new PPV(); } else if (metricName == "npv") { metric = new NPV(); } else if (
             metricName == "fdr") { metric = new FDR(); } else if (metricName == "fpfn") { metric = new FPFN(); } else {
-            return 0;
+            return {};
         }
 
     // string nameOrCount = "";
@@ -59,7 +59,7 @@ std::string ClusterCommand::runOptiCluster(OptiMatrix *optiMatrix) {
     //
     sensFile += "label\tcutoff\ttp\ttn\tfp\tfn\tsensitivity\tspecificity\tppv\tnpv\tfdr\taccuracy\tmcc\tf1score\n";
     clusterMetrics += (
-        "\n\niter\ttime\tlabel\tnum_otus\tcutoff\ttp\ttn\tfp\tfn\tsensitivity\tspecificity\tppv\tnpv\tfdr\taccuracy\tmcc\tf1score\n");
+        "iter\ttime\tlabel\tnum_otus\tcutoff\ttp\ttn\tfp\tfn\tsensitivity\tspecificity\tppv\tnpv\tfdr\taccuracy\tmcc\tf1score\n");
     outStep +=
             "iter\ttime\tlabel\tnum_otus\tcutoff\ttp\ttn\tfp\tfn\tsensitivity\tspecificity\tppv\tnpv\tfdr\taccuracy\tmcc\tf1score\n";
 
@@ -91,7 +91,8 @@ std::string ClusterCommand::runOptiCluster(OptiMatrix *optiMatrix) {
             outStep += std::to_string(result) + "\t";
         }
         //m->mothurOutEndLine();
-        // outStep += "\n";
+        clusterMetrics += "\n";
+        outStep += "\n";
         // 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
@@ -105,6 +106,7 @@ std::string ClusterCommand::runOptiCluster(OptiMatrix *optiMatrix) {
             iters++;
 
             stats = cluster.getStats(tp, tn, fp, fn);
+
             numBins = cluster.getNumBins();
 
             clusterMetrics += (std::to_string(iters) + "\t" + std::to_string(time(nullptr) - start) + "\t" +
@@ -120,11 +122,12 @@ std::string ClusterCommand::runOptiCluster(OptiMatrix *optiMatrix) {
                 clusterMetrics += (std::to_string(result) + "\t");
                 outStep += std::to_string(result) + "\t";
             }
+            clusterMetrics += "\n";
             outStep += "\n";
         }
         ListVector *list = nullptr;
 
-        clusterMetrics += "\n\n";
+        // clusterMetrics += "\n\n";
         list = cluster.getList();
         //
         if (printHeaders) {
@@ -138,9 +141,8 @@ std::string ClusterCommand::runOptiCluster(OptiMatrix *optiMatrix) {
         sensFile += std::to_string(cutoff) + '\t' + std::to_string(cutoff) + '\t' + std::to_string(tp) + '\t' +
                 std::to_string(tn) + '\t' +
                 std::to_string(fp) + '\t' + std::to_string(fn) + '\t';
-        for (double result : stats) { sensFile + std::to_string(result) + '\t'; }
-        Rcpp::Rcout << "Metrics for the current cluster:\n\n " << sensFile << "\n\n" << clusterMetrics << "\n\n";
+        for (double result : stats) { sensFile += std::to_string(result) + '\t'; }
      }
     delete matrix;
-    return clusterMatrixOutput;
+    return {clusterMatrixOutput, sensFile, clusterMetrics};
 }
diff --git a/src/MothurDependencies/ClusterCommand.h b/src/MothurDependencies/ClusterCommand.h
@@ -48,14 +48,14 @@ using namespace std;
 
 
 class ClusterCommand {
-
 public:
 	//ClusterCommand(string);
 	ClusterCommand() {}
 	~ClusterCommand();
 	bool SetMaxIterations(const int iterations) {maxIters = iterations; return maxIters == iterations;}
 	bool SetOpticlusterRandomShuffle(const bool shuffle) {canShuffle = shuffle; return canShuffle;}
-	std::string runOptiCluster(OptiMatrix*);
+	bool SetMetricType(const string& newMetric) {metric = newMetric; return metric == newMetric;}
+	std::vector<std::string> runOptiCluster(OptiMatrix*);
 
 
 

diff --git a/src/RcppExports.cpp b/src/RcppExports.cpp
@@ -11,7 +11,7 @@ Rcpp::Rostream<false>& Rcpp::Rcerr = Rcpp::Rcpp_cerr_get();
 #endif
 
 // MatrixToOpiMatrixCluster
-std::string MatrixToOpiMatrixCluster(const std::vector<int>& xPosition, const std::vector<int>& yPosition, const std::vector<double>& data, const double cutoff, const int iterations, const bool shuffle);
+std::vector<std::string> MatrixToOpiMatrixCluster(const std::vector<int>& xPosition, const std::vector<int>& yPosition, const std::vector<double>& data, const double cutoff, const int iterations, const bool shuffle);
 RcppExport SEXP _Opticluster_MatrixToOpiMatrixCluster(SEXP xPositionSEXP, SEXP yPositionSEXP, SEXP dataSEXP, SEXP cutoffSEXP, SEXP iterationsSEXP, SEXP shuffleSEXP) {
 BEGIN_RCPP
     Rcpp::RObject rcpp_result_gen;

diff --git a/src/main.cpp b/src/main.cpp
@@ -12,7 +12,7 @@
 
 #include <Rcpp.h>
 //[[Rcpp::export]]
-std::string MatrixToOpiMatrixCluster(const std::vector<int> &xPosition,
+std::vector<std::string> MatrixToOpiMatrixCluster(const std::vector<int> &xPosition,
                         const std::vector<int> &yPosition, const std::vector<double> &data, const double cutoff,
                         const int iterations = 2, const bool shuffle = true)
 {

diff --git a/tests/.DS_Store b/tests/.DS_Store
diff --git a/tests/testthat/.DS_Store b/tests/testthat/.DS_Store
diff --git a/tests/testthat/extdata/randomdata.txt b/tests/testthat/extdata/randomdata.txt
@@ -0,0 +1,5 @@
+iter	time	label	num_otus	cutoff	tp	tn	fp	fn	sensitivity	specificity	ppv	npv	fdr	accuracy	mcc	f1score
+0	0	0.000000	351	0.000000	4199.000000	256025.000000	150647.000000	0.000000	1.000000	0.629561	0.027117	1.000000	0.027117	0.633347	0.130660	0.052803	
+1	1718226744	0.000000	604	0.000000	3171.000000	391832.000000	14840.000000	1028.000000	0.755180	0.963509	0.176059	0.997383	0.176059	0.961380	0.353060	0.285547	
+2	1718226744	0.000000	793	0.000000	2450.000000	408919.000000	-2247.000000	1749.000000	0.583472	1.005525	12.068966	0.995741	12.068966	1.001212	2.665724	1.113130	
+
diff --git a/tests/testthat/extdata/test_file.txt b/tests/testthat/extdata/test_file.txt
@@ -0,0 +1,4 @@
+iter	time	label	num_otus	cutoff	tp	tn	fp	fn	sensitivity	specificity	ppv	npv	fdr	accuracy	mcc	f1score
+0	0	0.050000	351	0.050000	4199.000000	256025.000000	150647.000000	0.000000	1.000000	0.629561	0.027117	1.000000	0.027117	0.633347	0.130660	0.052803	1
+-19703246666551735	0.050000	604	0.050000	3171.000000	391832.000000	14840.000000	1028.000000	0.755180	0.963509	0.176059	0.997383	0.176059	0.961380	0.353060	0.285547    2
+-19703246666551735	0.050000	793	0.050000	2450.000000	408919.000000	-2247.000000	1749.000000	0.583472	1.005525	12.068966	0.995741	12.068966	1.001212	2.665724	1.113130
diff --git a/tests/testthat/test-test-opticluster.R b/tests/testthat/test-test-opticluster.R
@@ -2,7 +2,9 @@ test_that("Clustering returns proper results", {
   expected_df <- readRDS(test_path("extdata","df_test_file.RDS"))
   matrix <- readRDS(test_path("extdata","matrix_data.RDS"))
   df <- Opticluster::opti_cluster(matrix, 0.2, 2, FALSE)
-  df$exists <- do.call(paste0, df) %in% do.call(paste0, expected_df)
-  expect_equal(class(df), "data.frame")
+  df$cluster$exists <- do.call(paste0, df$cluster) %in% do.call(paste0, expected_df)
+  expect_equal(class(df$cluster), "data.frame")
+  expect_equal(class(df$cluster_metrics), "data.frame")
+  expect_equal(class(df$other_cluster_metrics), "data.frame")
   expect_true(all(df$exists == TRUE))
 })