rnabioco · jayhesselberth · Jan 20, 2025 · Jan 20, 2025 · Jan 20, 2025 · Jan 21, 2025
diff --git a/.gitignore b/.gitignore
@@ -12,3 +12,4 @@ index.html
 .env
 docs
 .DS_Store
+.vscode
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -51,7 +51,8 @@ Suggests:
     tidyr
 biocViews:
 LinkingTo:
-    Rcpp (>= 1.0.0)
+    Rcpp (>= 1.0.0),
+    cpp11
 VignetteBuilder:
     knitr
 Encoding: UTF-8

diff --git a/R/RcppExports.R b/R/RcppExports.R
@@ -1,10 +1,6 @@
 # Generated by using Rcpp::compileAttributes() -> do not edit by hand
 # Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
 
-bed12toexons_impl <- function(x) {
-    .Call(`_valr_bed12toexons_impl`, x)
-}
-
 closest_impl <- function(x, y, grp_idx_x, grp_idx_y, suffix_x, suffix_y) {
     .Call(`_valr_closest_impl`, x, y, grp_idx_x, grp_idx_y, suffix_x, suffix_y)
 }
@@ -21,10 +17,6 @@ dist_impl <- function(x, y, x_grp_indexes, y_grp_indexes, distcalc) {
     .Call(`_valr_dist_impl`, x, y, x_grp_indexes, y_grp_indexes, distcalc)
 }
 
-flank_impl <- function(df, genome, both = 0, left = 0, right = 0, fraction = FALSE, stranded = FALSE, trim = FALSE) {
-    .Call(`_valr_flank_impl`, df, genome, both, left, right, fraction, stranded, trim)
-}
-
 gcoverage_impl <- function(gdf, max_coords) {
     .Call(`_valr_gcoverage_impl`, gdf, max_coords)
 }
@@ -33,10 +25,6 @@ intersect_impl <- function(x, y, x_grp_indexes, y_grp_indexes, invert = FALSE, s
     .Call(`_valr_intersect_impl`, x, y, x_grp_indexes, y_grp_indexes, invert, suffix_x, suffix_y)
 }
 
-makewindows_impl <- function(df, win_size = 0L, num_win = 0L, step_size = 0L, reverse = FALSE) {
-    .Call(`_valr_makewindows_impl`, df, win_size, num_win, step_size, reverse)
-}
-
 merge_impl <- function(gdf, max_dist = 0L, collapse = TRUE) {
     .Call(`_valr_merge_impl`, gdf, max_dist, collapse)
 }
@@ -45,10 +33,6 @@ partition_impl <- function(gdf, max_dist = -1L) {
     .Call(`_valr_partition_impl`, gdf, max_dist)
 }
 
-random_impl <- function(genome, length, n, seed = 0L) {
-    .Call(`_valr_random_impl`, genome, length, n, seed)
-}
-
 shuffle_impl <- function(df, incl, within = FALSE, max_tries = 1000L, seed = 0L) {
     .Call(`_valr_shuffle_impl`, df, incl, within, max_tries, seed)
 }

diff --git a/R/cpp11.R b/R/cpp11.R
@@ -0,0 +1,17 @@
+# Generated by cpp11: do not edit by hand
+
+bed12toexons_impl <- function(x) {
+  .Call(`_valr_bed12toexons_impl`, x)
+}
+
+flank_impl <- function(df, genome, both, left, right, fraction, stranded, trim) {
+  .Call(`_valr_flank_impl`, df, genome, both, left, right, fraction, stranded, trim)
+}
+
+makewindows_impl <- function(df, win_size, num_win, step_size, reverse) {
+  .Call(`_valr_makewindows_impl`, df, win_size, num_win, step_size, reverse)
+}
+
+random_impl <- function(genome, length, n, seed) {
+  .Call(`_valr_random_impl`, genome, length, n, seed)
+}
diff --git a/R/read_bed.r b/R/read_bed.r
@@ -160,8 +160,8 @@ peak_coltypes <- list(
 
 bed12_coltypes <- list(
   chrom = readr::col_character(),
-  start = readr::col_integer(),
-  end = readr::col_integer(),
+  start = readr::col_double(),
+  end = readr::col_double(),
   name = readr::col_character(),
   score = readr::col_character(),
   strand = readr::col_character(),

diff --git a/inst/include/genome.h b/inst/include/genome.h
@@ -24,7 +24,7 @@
   IntegerVector sizes = genome[col_size] ;
 
   if (unique(refs).length() != refs.length())
-    stop("duplicate reference names in genome file.") ;
+    Rcpp::stop("duplicate reference names in genome file.") ;
 
   int nchrom = genome.nrows() ;
   for (int i = 0; i < nchrom; ++i) {

diff --git a/inst/include/group_apply.h b/inst/include/group_apply.h
@@ -28,7 +28,7 @@
   int ng_y = shared_grps_y.size() ;
 
   if (ng_x != ng_y) {
-    stop("incompatible groups found between x and y dataframes") ;
+    Rcpp::stop("incompatible groups found between x and y dataframes") ;
   }
 
   // access the group .rows list

diff --git a/inst/include/utils.h b/inst/include/utils.h
@@ -18,6 +18,9 @@ DataFrame subset_dataframe(const DataFrame& df,
 DataFrame subset_dataframe(const DataFrame& df,
                            IntegerVector indices) ;
 
+writable::data_frame subset_dataframe(const data_frame& df,
+                           std::vector<int> indices) ;
+
 inline DataFrame check_is_grouped(const DataFrame& x) {
   bool is_grouped(Rf_inherits(x, "grouped_df")) ;
 

diff --git a/inst/include/valr.h b/inst/include/valr.h
@@ -10,11 +10,12 @@
 #ifndef valr__valr_H
 #define valr__valr_H
 
-// [[Rcpp::plugins(cpp11)]]
-
 #include <Rcpp.h>
 using namespace Rcpp ;
 
+#include <cpp11.hpp>
+using namespace cpp11;
+
 #include "utils.h"
 #include "grouped_dataframe.h"
 #include "IntervalTree.h"

diff --git a/src/RcppExports.cpp b/src/RcppExports.cpp
@@ -11,17 +11,6 @@ Rcpp::Rostream<true>&  Rcpp::Rcout = Rcpp::Rcpp_cout_get();
 Rcpp::Rostream<false>& Rcpp::Rcerr = Rcpp::Rcpp_cerr_get();
 #endif
 
-// bed12toexons_impl
-DataFrame bed12toexons_impl(DataFrame x);
-RcppExport SEXP _valr_bed12toexons_impl(SEXP xSEXP) {
-BEGIN_RCPP
-    Rcpp::RObject rcpp_result_gen;
-    Rcpp::RNGScope rcpp_rngScope_gen;
-    Rcpp::traits::input_parameter< DataFrame >::type x(xSEXP);
-    rcpp_result_gen = Rcpp::wrap(bed12toexons_impl(x));
-    return rcpp_result_gen;
-END_RCPP
-}
 // closest_impl
 DataFrame closest_impl(ValrGroupedDataFrame x, ValrGroupedDataFrame y, IntegerVector grp_idx_x, IntegerVector grp_idx_y, const std::string& suffix_x, const std::string& suffix_y);
 RcppExport SEXP _valr_closest_impl(SEXP xSEXP, SEXP ySEXP, SEXP grp_idx_xSEXP, SEXP grp_idx_ySEXP, SEXP suffix_xSEXP, SEXP suffix_ySEXP) {
@@ -79,24 +68,6 @@ BEGIN_RCPP
     return rcpp_result_gen;
 END_RCPP
 }
-// flank_impl
-DataFrame flank_impl(DataFrame df, DataFrame genome, double both, double left, double right, bool fraction, bool stranded, bool trim);
-RcppExport SEXP _valr_flank_impl(SEXP dfSEXP, SEXP genomeSEXP, SEXP bothSEXP, SEXP leftSEXP, SEXP rightSEXP, SEXP fractionSEXP, SEXP strandedSEXP, SEXP trimSEXP) {
-BEGIN_RCPP
-    Rcpp::RObject rcpp_result_gen;
-    Rcpp::RNGScope rcpp_rngScope_gen;
-    Rcpp::traits::input_parameter< DataFrame >::type df(dfSEXP);
-    Rcpp::traits::input_parameter< DataFrame >::type genome(genomeSEXP);
-    Rcpp::traits::input_parameter< double >::type both(bothSEXP);
-    Rcpp::traits::input_parameter< double >::type left(leftSEXP);
-    Rcpp::traits::input_parameter< double >::type right(rightSEXP);
-    Rcpp::traits::input_parameter< bool >::type fraction(fractionSEXP);
-    Rcpp::traits::input_parameter< bool >::type stranded(strandedSEXP);
-    Rcpp::traits::input_parameter< bool >::type trim(trimSEXP);
-    rcpp_result_gen = Rcpp::wrap(flank_impl(df, genome, both, left, right, fraction, stranded, trim));
-    return rcpp_result_gen;
-END_RCPP
-}
 // gcoverage_impl
 DataFrame gcoverage_impl(const ValrGroupedDataFrame& gdf, const IntegerVector& max_coords);
 RcppExport SEXP _valr_gcoverage_impl(SEXP gdfSEXP, SEXP max_coordsSEXP) {
@@ -126,21 +97,6 @@ BEGIN_RCPP
     return rcpp_result_gen;
 END_RCPP
 }
-// makewindows_impl
-DataFrame makewindows_impl(DataFrame df, int win_size, int num_win, int step_size, bool reverse);
-RcppExport SEXP _valr_makewindows_impl(SEXP dfSEXP, SEXP win_sizeSEXP, SEXP num_winSEXP, SEXP step_sizeSEXP, SEXP reverseSEXP) {
-BEGIN_RCPP
-    Rcpp::RObject rcpp_result_gen;
-    Rcpp::RNGScope rcpp_rngScope_gen;
-    Rcpp::traits::input_parameter< DataFrame >::type df(dfSEXP);
-    Rcpp::traits::input_parameter< int >::type win_size(win_sizeSEXP);
-    Rcpp::traits::input_parameter< int >::type num_win(num_winSEXP);
-    Rcpp::traits::input_parameter< int >::type step_size(step_sizeSEXP);
-    Rcpp::traits::input_parameter< bool >::type reverse(reverseSEXP);
-    rcpp_result_gen = Rcpp::wrap(makewindows_impl(df, win_size, num_win, step_size, reverse));
-    return rcpp_result_gen;
-END_RCPP
-}
 // merge_impl
 DataFrame merge_impl(ValrGroupedDataFrame gdf, int max_dist, bool collapse);
 RcppExport SEXP _valr_merge_impl(SEXP gdfSEXP, SEXP max_distSEXP, SEXP collapseSEXP) {
@@ -166,20 +122,6 @@ BEGIN_RCPP
     return rcpp_result_gen;
 END_RCPP
 }
-// random_impl
-DataFrame random_impl(DataFrame genome, int length, int n, int seed);
-RcppExport SEXP _valr_random_impl(SEXP genomeSEXP, SEXP lengthSEXP, SEXP nSEXP, SEXP seedSEXP) {
-BEGIN_RCPP
-    Rcpp::RObject rcpp_result_gen;
-    Rcpp::RNGScope rcpp_rngScope_gen;
-    Rcpp::traits::input_parameter< DataFrame >::type genome(genomeSEXP);
-    Rcpp::traits::input_parameter< int >::type length(lengthSEXP);
-    Rcpp::traits::input_parameter< int >::type n(nSEXP);
-    Rcpp::traits::input_parameter< int >::type seed(seedSEXP);
-    rcpp_result_gen = Rcpp::wrap(random_impl(genome, length, n, seed));
-    return rcpp_result_gen;
-END_RCPP
-}
 // shuffle_impl
 DataFrame shuffle_impl(DataFrame df, DataFrame incl, bool within, int max_tries, int seed);
 RcppExport SEXP _valr_shuffle_impl(SEXP dfSEXP, SEXP inclSEXP, SEXP withinSEXP, SEXP max_triesSEXP, SEXP seedSEXP) {

diff --git a/src/bed12toexons.cpp b/src/bed12toexons.cpp
@@ -27,14 +27,14 @@ std::vector<int> csv_values(std::string csv) {
   return values ;
 }
 
-// [[Rcpp::export]]
-DataFrame bed12toexons_impl(DataFrame x) {
+[[cpp11::register]]
+writable::data_frame bed12toexons_impl(data_frame x) {
 
   // input
-  IntegerVector starts = x["start"] ;
-  std::vector<std::string> exon_sizes = x["exon_sizes"] ;
-  std::vector<std::string> exon_starts = x["exon_starts"] ;
-  std::vector<std::string> strands = x["strand"] ;
+  doubles starts = x["start"] ;
+  strings exon_sizes = x["exon_sizes"] ;
+  strings exon_starts = x["exon_starts"] ;
+  strings strands = x["strand"] ;
 
   // storage
   std::vector<int> starts_out ;
@@ -66,18 +66,23 @@ DataFrame bed12toexons_impl(DataFrame x) {
     }
   }
 
-  DataFrame out = subset_dataframe(x, df_idx) ;
-
-  out["start"] = starts_out ;
-  out["end"] = ends_out ;
-  out["score"] = nums_out ;
-
-  return out ;
+  writable::data_frame out = subset_dataframe(x, df_idx) ;
+
+  return writable::data_frame({
+    "chrom"_nm = out["chrom"],
+    "start"_nm = starts_out,
+    "end"_nm = ends_out,
+    "score"_nm = nums_out,
+    "strand"_nm = out["strand"],
+    "name"_nm = out["name"],
+    "cdsStart"_nm = out["cdsStart"],
+    "cdsEnd"_nm = out["cdsEnd"],
+    "exonCount"_nm = out["exonCount"],
+    "exonSizes"_nm = out["exonSizes"],
+    "exonStarts"_nm = out["exonStarts"],
+    "itemRgb"_nm = out["itemRgb"],
+    "blockCount"_nm = out["blockCount"],
+    "blockSizes"_nm = out["blockSizes"],
+    "blockStarts"_nm = out["blockStarts"]
+  }) ;
 }
-
-/***R
-library(valr)
-library(dplyr)
-x <- read_bed12(valr_example('mm9.refGene.bed.gz'))
-bed12_to_exons(x)
-*/
diff --git a/src/closest.cpp b/src/closest.cpp
@@ -233,7 +233,7 @@
   int ng_y = grp_idx_y.size() ;
 
   if (ng_x != ng_y) {
-    stop("incompatible groups found between x and y dataframes") ;
+    Rcpp::stop("incompatible groups found between x and y dataframes") ;
   }
 
   // access the group .rows list
@@ -297,4 +297,3 @@
   return res ;
 
 }
-
diff --git a/src/cpp11.cpp b/src/cpp11.cpp
@@ -0,0 +1,75 @@
+// Generated by cpp11: do not edit by hand
+// clang-format off
+
+#include <cpp11/R.hpp>
+#include <Rcpp.h>
+using namespace Rcpp;
+#include "cpp11/declarations.hpp"
+#include <R_ext/Visibility.h>
+
+// bed12toexons.cpp
+writable::data_frame bed12toexons_impl(data_frame x);
+extern "C" SEXP _valr_bed12toexons_impl(SEXP x) {
+  BEGIN_CPP11
+    return cpp11::as_sexp(bed12toexons_impl(cpp11::as_cpp<cpp11::decay_t<data_frame>>(x)));
+  END_CPP11
+}
+// flank.cpp
+writable::data_frame flank_impl(data_frame df, data_frame genome, double both, double left, double right, bool fraction, bool stranded, bool trim);
+extern "C" SEXP _valr_flank_impl(SEXP df, SEXP genome, SEXP both, SEXP left, SEXP right, SEXP fraction, SEXP stranded, SEXP trim) {
+  BEGIN_CPP11
+    return cpp11::as_sexp(flank_impl(cpp11::as_cpp<cpp11::decay_t<data_frame>>(df), cpp11::as_cpp<cpp11::decay_t<data_frame>>(genome), cpp11::as_cpp<cpp11::decay_t<double>>(both), cpp11::as_cpp<cpp11::decay_t<double>>(left), cpp11::as_cpp<cpp11::decay_t<double>>(right), cpp11::as_cpp<cpp11::decay_t<bool>>(fraction), cpp11::as_cpp<cpp11::decay_t<bool>>(stranded), cpp11::as_cpp<cpp11::decay_t<bool>>(trim)));
+  END_CPP11
+}
+// makewindows.cpp
+writable::data_frame makewindows_impl(data_frame df, int win_size, int num_win, int step_size, bool reverse);
+extern "C" SEXP _valr_makewindows_impl(SEXP df, SEXP win_size, SEXP num_win, SEXP step_size, SEXP reverse) {
+  BEGIN_CPP11
+    return cpp11::as_sexp(makewindows_impl(cpp11::as_cpp<cpp11::decay_t<data_frame>>(df), cpp11::as_cpp<cpp11::decay_t<int>>(win_size), cpp11::as_cpp<cpp11::decay_t<int>>(num_win), cpp11::as_cpp<cpp11::decay_t<int>>(step_size), cpp11::as_cpp<cpp11::decay_t<bool>>(reverse)));
+  END_CPP11
+}
+// random.cpp
+writable::data_frame random_impl(data_frame genome, double length, int n, int seed);
+extern "C" SEXP _valr_random_impl(SEXP genome, SEXP length, SEXP n, SEXP seed) {
+  BEGIN_CPP11
+    return cpp11::as_sexp(random_impl(cpp11::as_cpp<cpp11::decay_t<data_frame>>(genome), cpp11::as_cpp<cpp11::decay_t<double>>(length), cpp11::as_cpp<cpp11::decay_t<int>>(n), cpp11::as_cpp<cpp11::decay_t<int>>(seed)));
+  END_CPP11
+}
+
+extern "C" {
+/* .Call calls */
+extern SEXP _valr_closest_impl(SEXP, SEXP, SEXP, SEXP, SEXP, SEXP);
+extern SEXP _valr_complement_impl(SEXP, SEXP);
+extern SEXP _valr_coverage_impl(SEXP, SEXP, SEXP, SEXP);
+extern SEXP _valr_dist_impl(SEXP, SEXP, SEXP, SEXP, SEXP);
+extern SEXP _valr_gcoverage_impl(SEXP, SEXP);
+extern SEXP _valr_intersect_impl(SEXP, SEXP, SEXP, SEXP, SEXP, SEXP, SEXP);
+extern SEXP _valr_merge_impl(SEXP, SEXP, SEXP);
+extern SEXP _valr_partition_impl(SEXP, SEXP);
+extern SEXP _valr_shuffle_impl(SEXP, SEXP, SEXP, SEXP, SEXP);
+extern SEXP _valr_subtract_impl(SEXP, SEXP, SEXP, SEXP);
+
+static const R_CallMethodDef CallEntries[] = {
+    {"_valr_bed12toexons_impl", (DL_FUNC) &_valr_bed12toexons_impl, 1},
+    {"_valr_closest_impl",      (DL_FUNC) &_valr_closest_impl,      6},
+    {"_valr_complement_impl",   (DL_FUNC) &_valr_complement_impl,   2},
+    {"_valr_coverage_impl",     (DL_FUNC) &_valr_coverage_impl,     4},
+    {"_valr_dist_impl",         (DL_FUNC) &_valr_dist_impl,         5},
+    {"_valr_flank_impl",        (DL_FUNC) &_valr_flank_impl,        8},
+    {"_valr_gcoverage_impl",    (DL_FUNC) &_valr_gcoverage_impl,    2},
+    {"_valr_intersect_impl",    (DL_FUNC) &_valr_intersect_impl,    7},
+    {"_valr_makewindows_impl",  (DL_FUNC) &_valr_makewindows_impl,  5},
+    {"_valr_merge_impl",        (DL_FUNC) &_valr_merge_impl,        3},
+    {"_valr_partition_impl",    (DL_FUNC) &_valr_partition_impl,    2},
+    {"_valr_random_impl",       (DL_FUNC) &_valr_random_impl,       4},
+    {"_valr_shuffle_impl",      (DL_FUNC) &_valr_shuffle_impl,      5},
+    {"_valr_subtract_impl",     (DL_FUNC) &_valr_subtract_impl,     4},
+    {NULL, NULL, 0}
+};
+}
+
+extern "C" attribute_visible void R_init_valr(DllInfo* dll){
+  R_registerRoutines(dll, NULL, CallEntries, NULL, NULL);
+  R_useDynamicSymbols(dll, FALSE);
+  R_forceSymbols(dll, TRUE);
+}
-Original file line number
+Diff line change
@@ Expand Up / @@ -12,3 +12,4 @@ index.html @@
     .env
     docs
     .DS_Store
+    .vscode