- fixes when reading file with empty rows (now values within a row)

- fixes #3
rexyai · Jan 13, 2020 · 5cf9c19 · 5cf9c19
1 parent a2f5f46
commit 5cf9c19
Show file tree

Hide file tree

Showing 10 changed files with 53 additions and 36 deletions.
diff --git a/.Rbuildignore b/.Rbuildignore
@@ -1,3 +1,4 @@
 ^.*\.Rproj$
 ^\.Rproj\.user$
 ^cran-comments\.md$
+^CRAN-RELEASE$
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,8 +1,8 @@
 Package: sparsio
 Type: Package
 Title: I/O Operations with Sparse Matrices
-Version: 1.0.0
-Date: 2017-06-28
+Version: 1.0.1
+Date: 2020-01-13
 Authors@R: c(person("Dmitriy", "Selivanov", role = c("aut", "cre"),
     email = "selivanov.dmitriy@gmail.com"), 
     person("Felix", "Riedel", role = c("aut"),
@@ -25,4 +25,4 @@ Suggests:
     testthat
 URL: https://github.com/dselivanov/sparsio
 BugReports: https://github.com/dselivanov/sparsio/issues
-RoxygenNote: 6.0.1
+RoxygenNote: 6.1.1
diff --git a/NAMESPACE b/NAMESPACE
@@ -5,4 +5,4 @@ export(write_svmlight)
 import(Matrix)
 import(Rcpp)
 import(methods)
-useDynLib(sparsio)
+useDynLib(sparsio, .registration = TRUE)
diff --git a/R/RcppExports.R b/R/RcppExports.R
@@ -2,10 +2,10 @@
 # Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
 
 read_svmlight_cpp <- function(filename, zero_based = 1L) {
-    .Call('sparsio_read_svmlight_cpp', PACKAGE = 'sparsio', filename, zero_based)
+    .Call(`_sparsio_read_svmlight_cpp`, filename, zero_based)
 }
 
 write_svmlight_cpp <- function(x, y, filename, zero_based = 1L) {
-    invisible(.Call('sparsio_write_svmlight_cpp', PACKAGE = 'sparsio', x, y, filename, zero_based))
+    invisible(.Call(`_sparsio_write_svmlight_cpp`, x, y, filename, zero_based))
 }
 
diff --git a/R/sparsio.R b/R/sparsio.R
@@ -1,4 +1,4 @@
-#' @useDynLib sparsio
+#' @useDynLib sparsio, .registration = TRUE
 #' @import Matrix
 #' @import Rcpp
 #' @import methods

diff --git a/cran-comments.md b/cran-comments.md
diff --git a/src/RcppExports.cpp b/src/RcppExports.cpp
@@ -7,7 +7,7 @@ using namespace Rcpp;
 
 // read_svmlight_cpp
 List read_svmlight_cpp(Rcpp::String filename, int zero_based);
-RcppExport SEXP sparsio_read_svmlight_cpp(SEXP filenameSEXP, SEXP zero_basedSEXP) {
+RcppExport SEXP _sparsio_read_svmlight_cpp(SEXP filenameSEXP, SEXP zero_basedSEXP) {
 BEGIN_RCPP
     Rcpp::RObject rcpp_result_gen;
     Rcpp::RNGScope rcpp_rngScope_gen;
@@ -19,7 +19,7 @@ END_RCPP
 }
 // write_svmlight_cpp
 void write_svmlight_cpp(S4 x, NumericVector y, Rcpp::String filename, int zero_based);
-RcppExport SEXP sparsio_write_svmlight_cpp(SEXP xSEXP, SEXP ySEXP, SEXP filenameSEXP, SEXP zero_basedSEXP) {
+RcppExport SEXP _sparsio_write_svmlight_cpp(SEXP xSEXP, SEXP ySEXP, SEXP filenameSEXP, SEXP zero_basedSEXP) {
 BEGIN_RCPP
     Rcpp::RNGScope rcpp_rngScope_gen;
     Rcpp::traits::input_parameter< S4 >::type x(xSEXP);
@@ -32,8 +32,8 @@ END_RCPP
 }
 
 static const R_CallMethodDef CallEntries[] = {
-    {"sparsio_read_svmlight_cpp", (DL_FUNC) &sparsio_read_svmlight_cpp, 2},
-    {"sparsio_write_svmlight_cpp", (DL_FUNC) &sparsio_write_svmlight_cpp, 4},
+    {"_sparsio_read_svmlight_cpp", (DL_FUNC) &_sparsio_read_svmlight_cpp, 2},
+    {"_sparsio_write_svmlight_cpp", (DL_FUNC) &_sparsio_write_svmlight_cpp, 4},
     {NULL, NULL, 0}
 };
 

diff --git a/src/read_svmlight.cpp b/src/read_svmlight.cpp
@@ -39,19 +39,20 @@ List read_svmlight_cpp(Rcpp::String filename, int zero_based = 1) {
       nrow++;
     }
 
+    int id;
+    double value;
     // foreach token (id:value) in line 
     while (!tokenStream.eof()) {
-        std::getline(tokenStream, token, ' ');
-        int id;
-        double value;
+      std::getline(tokenStream, token, ' ');
+      if (!token.empty()) {
         sscanf(token.c_str(), "%d:%lf", &id, &value);
         // case when indices start from 1
         id = id - index_start_with;
         ncol = std::max(id, ncol); 
         x_values.push_back(value);
-
         col_indices.push_back(id);
         pointer_counter++;
+      }
     }
   }
   ncol++;

diff --git a/tests/testthat/test-empty-rows.svmlight b/tests/testthat/test-empty-rows.svmlight
@@ -0,0 +1,17 @@
+0 
+0 
+0 
+0 
+0 
+0 
+0 
+0 
+0 
+0 
+0 
+0 
+0 
+0 
+0 
+0 
+0 70628:1
diff --git a/tests/testthat/test-read.R b/tests/testthat/test-read.R
@@ -9,8 +9,25 @@ test_that("read basic", {
 
 test_that("read different formats", {
   formats = c("CsparseMatrix", "RsparseMatrix", "TsparseMatrix")
-  for( f in formats) {
+  for (f in formats) {
     m = read_svmlight("test.svmlight", type = f)
     expect_true(inherits(m$x, f))
   }
-})
+})
+
+test_that("read file with empty rows", {
+  ln = readLines("test-empty-rows.svmlight")
+  # find max col
+  all_vals = strsplit(ln, " ", fixed = T)
+  all_vals = all_vals[grepl(":", all_vals)]
+  all_vals = unlist(all_vals, recursive = F, use.names = F)
+  all_vals = strsplit(all_vals, ":", fixed = T)
+  all_vals = unlist(all_vals, recursive = F, use.names = F)
+  # +1 because of 0-based indexing
+  n_col_true = max(as.numeric(all_vals)) + 1
+
+  n_row_true = length(ln)
+  m = read_svmlight("test-empty-rows.svmlight", type = "RsparseMatrix")
+  expect_equal(dim(m$x)[1], n_row_true)
+  expect_equal(dim(m$x)[2], n_col_true)
+})