Reduce overhead of extent and res calls

isciences · Jun 27, 2022 · 5b88ecf · 5b88ecf
1 parent 6523ad5
commit 5b88ecf
Show file tree

Hide file tree

Showing 6 changed files with 45 additions and 16 deletions.
diff --git a/R/RcppExports.R b/R/RcppExports.R
@@ -5,12 +5,12 @@ CPP_coverage_fraction <- function(rast, wkb, crop) {
     .Call('_exactextractr_CPP_coverage_fraction', PACKAGE = 'exactextractr', rast, wkb, crop)
 }
 
-CPP_exact_extract <- function(rast, rast_uncropped, weights, wkb, default_value, default_weight, include_xy, include_cell_number, include_area, area_weights, coverage_areas, p_area_method, include_cols, src_names, p_weights_names, warn_on_disaggregate, grid_compat_tol) {
-    .Call('_exactextractr_CPP_exact_extract', PACKAGE = 'exactextractr', rast, rast_uncropped, weights, wkb, default_value, default_weight, include_xy, include_cell_number, include_area, area_weights, coverage_areas, p_area_method, include_cols, src_names, p_weights_names, warn_on_disaggregate, grid_compat_tol)
+CPP_exact_extract <- function(rast, rast_ext, rast_res, rast_uncropped, weights, wkb, default_value, default_weight, include_xy, include_cell_number, include_area, area_weights, coverage_areas, p_area_method, include_cols, src_names, p_weights_names, warn_on_disaggregate, grid_compat_tol) {
+    .Call('_exactextractr_CPP_exact_extract', PACKAGE = 'exactextractr', rast, rast_ext, rast_res, rast_uncropped, weights, wkb, default_value, default_weight, include_xy, include_cell_number, include_area, area_weights, coverage_areas, p_area_method, include_cols, src_names, p_weights_names, warn_on_disaggregate, grid_compat_tol)
 }
 
-CPP_stats <- function(rast, weights, wkb, default_value, default_weight, coverage_areas, p_area_method, stats, max_cells_in_memory, grid_compat_tol, quantiles) {
-    .Call('_exactextractr_CPP_stats', PACKAGE = 'exactextractr', rast, weights, wkb, default_value, default_weight, coverage_areas, p_area_method, stats, max_cells_in_memory, grid_compat_tol, quantiles)
+CPP_stats <- function(rast, rast_ext, rast_res, weights, wkb, default_value, default_weight, coverage_areas, p_area_method, stats, max_cells_in_memory, grid_compat_tol, quantiles) {
+    .Call('_exactextractr_CPP_stats', PACKAGE = 'exactextractr', rast, rast_ext, rast_res, weights, wkb, default_value, default_weight, coverage_areas, p_area_method, stats, max_cells_in_memory, grid_compat_tol, quantiles)
 }
 
 CPP_update_max_coverage <- function(extent, res, max_coverage, max_coverage_index, tot_coverage, wkb, index) {

diff --git a/R/exact_extract.R b/R/exact_extract.R
@@ -447,6 +447,10 @@ NULL
     weights <- .eagerLoad(weights, geoms, max_cells_in_memory, message_on_fail = progress)
   }
 
+  # Calculate extent and resolution outside of loop for performance (yes, really)
+  x_ext <- .extent(x)
+  x_res <- .res(x)
+
   # At this point, if the data have not been preloaded into memory, either:
   # - we don't have the terra package available
   # - we can't fit the whole processing area in memory
@@ -497,7 +501,7 @@ NULL
       }
 
       results <- lapply(sf::st_as_binary(geoms, EWKB=TRUE), function(wkb) {
-        ret <- CPP_stats(x, weights, wkb, default_value, default_weight, coverage_area, area_method, fun, max_cells_in_memory, grid_compat_tol, quantiles)
+        ret <- CPP_stats(x, x_ext, x_res, weights, wkb, default_value, default_weight, coverage_area, area_method, fun, max_cells_in_memory, grid_compat_tol, quantiles)
         update_progress()
         return(ret)
       })
@@ -590,6 +594,8 @@ NULL
         warn_on_disaggregate <- feature_num == 1
 
         col_list <- CPP_exact_extract(x,
+                                      x_ext,
+                                      x_res,
                                       x_orig,
                                       weights,
                                       wkb,

diff --git a/R/exact_extract_helpers.R b/R/exact_extract_helpers.R
@@ -379,7 +379,7 @@
   if (inherits(r, 'BasicRaster')) {
     return(raster::inMemory(r))
   } else if (inherits(r, 'SpatRaster')) {
-    return(terra::inMemory(r[[1]]))
+    return(terra::inMemory(r)[1])
   } else {
     stop('Unknown type: ', class(r))
   }

diff --git a/src/RcppExports.cpp b/src/RcppExports.cpp
@@ -24,12 +24,14 @@ BEGIN_RCPP
 END_RCPP
 }
 // CPP_exact_extract
-Rcpp::List CPP_exact_extract(Rcpp::S4& rast, Rcpp::Nullable<Rcpp::S4>& rast_uncropped, Rcpp::Nullable<Rcpp::S4>& weights, const Rcpp::RawVector& wkb, double default_value, double default_weight, bool include_xy, bool include_cell_number, bool include_area, bool area_weights, bool coverage_areas, Rcpp::Nullable<Rcpp::CharacterVector>& p_area_method, Rcpp::Nullable<Rcpp::List>& include_cols, Rcpp::CharacterVector& src_names, Rcpp::Nullable<Rcpp::CharacterVector>& p_weights_names, bool warn_on_disaggregate, double grid_compat_tol);
-RcppExport SEXP _exactextractr_CPP_exact_extract(SEXP rastSEXP, SEXP rast_uncroppedSEXP, SEXP weightsSEXP, SEXP wkbSEXP, SEXP default_valueSEXP, SEXP default_weightSEXP, SEXP include_xySEXP, SEXP include_cell_numberSEXP, SEXP include_areaSEXP, SEXP area_weightsSEXP, SEXP coverage_areasSEXP, SEXP p_area_methodSEXP, SEXP include_colsSEXP, SEXP src_namesSEXP, SEXP p_weights_namesSEXP, SEXP warn_on_disaggregateSEXP, SEXP grid_compat_tolSEXP) {
+Rcpp::List CPP_exact_extract(Rcpp::S4& rast, const Rcpp::NumericVector& rast_ext, const Rcpp::NumericVector& rast_res, Rcpp::Nullable<Rcpp::S4>& rast_uncropped, Rcpp::Nullable<Rcpp::S4>& weights, const Rcpp::RawVector& wkb, double default_value, double default_weight, bool include_xy, bool include_cell_number, bool include_area, bool area_weights, bool coverage_areas, Rcpp::Nullable<Rcpp::CharacterVector>& p_area_method, Rcpp::Nullable<Rcpp::List>& include_cols, Rcpp::CharacterVector& src_names, Rcpp::Nullable<Rcpp::CharacterVector>& p_weights_names, bool warn_on_disaggregate, double grid_compat_tol);
+RcppExport SEXP _exactextractr_CPP_exact_extract(SEXP rastSEXP, SEXP rast_extSEXP, SEXP rast_resSEXP, SEXP rast_uncroppedSEXP, SEXP weightsSEXP, SEXP wkbSEXP, SEXP default_valueSEXP, SEXP default_weightSEXP, SEXP include_xySEXP, SEXP include_cell_numberSEXP, SEXP include_areaSEXP, SEXP area_weightsSEXP, SEXP coverage_areasSEXP, SEXP p_area_methodSEXP, SEXP include_colsSEXP, SEXP src_namesSEXP, SEXP p_weights_namesSEXP, SEXP warn_on_disaggregateSEXP, SEXP grid_compat_tolSEXP) {
 BEGIN_RCPP
     Rcpp::RObject rcpp_result_gen;
     Rcpp::RNGScope rcpp_rngScope_gen;
     Rcpp::traits::input_parameter< Rcpp::S4& >::type rast(rastSEXP);
+    Rcpp::traits::input_parameter< const Rcpp::NumericVector& >::type rast_ext(rast_extSEXP);
+    Rcpp::traits::input_parameter< const Rcpp::NumericVector& >::type rast_res(rast_resSEXP);
     Rcpp::traits::input_parameter< Rcpp::Nullable<Rcpp::S4>& >::type rast_uncropped(rast_uncroppedSEXP);
     Rcpp::traits::input_parameter< Rcpp::Nullable<Rcpp::S4>& >::type weights(weightsSEXP);
     Rcpp::traits::input_parameter< const Rcpp::RawVector& >::type wkb(wkbSEXP);
@@ -46,17 +48,19 @@ BEGIN_RCPP
     Rcpp::traits::input_parameter< Rcpp::Nullable<Rcpp::CharacterVector>& >::type p_weights_names(p_weights_namesSEXP);
     Rcpp::traits::input_parameter< bool >::type warn_on_disaggregate(warn_on_disaggregateSEXP);
     Rcpp::traits::input_parameter< double >::type grid_compat_tol(grid_compat_tolSEXP);
-    rcpp_result_gen = Rcpp::wrap(CPP_exact_extract(rast, rast_uncropped, weights, wkb, default_value, default_weight, include_xy, include_cell_number, include_area, area_weights, coverage_areas, p_area_method, include_cols, src_names, p_weights_names, warn_on_disaggregate, grid_compat_tol));
+    rcpp_result_gen = Rcpp::wrap(CPP_exact_extract(rast, rast_ext, rast_res, rast_uncropped, weights, wkb, default_value, default_weight, include_xy, include_cell_number, include_area, area_weights, coverage_areas, p_area_method, include_cols, src_names, p_weights_names, warn_on_disaggregate, grid_compat_tol));
     return rcpp_result_gen;
 END_RCPP
 }
 // CPP_stats
-Rcpp::NumericMatrix CPP_stats(Rcpp::S4& rast, Rcpp::Nullable<Rcpp::S4> weights, const Rcpp::RawVector& wkb, double default_value, double default_weight, bool coverage_areas, Rcpp::Nullable<Rcpp::CharacterVector>& p_area_method, const Rcpp::StringVector& stats, int max_cells_in_memory, double grid_compat_tol, const Rcpp::Nullable<Rcpp::NumericVector>& quantiles);
-RcppExport SEXP _exactextractr_CPP_stats(SEXP rastSEXP, SEXP weightsSEXP, SEXP wkbSEXP, SEXP default_valueSEXP, SEXP default_weightSEXP, SEXP coverage_areasSEXP, SEXP p_area_methodSEXP, SEXP statsSEXP, SEXP max_cells_in_memorySEXP, SEXP grid_compat_tolSEXP, SEXP quantilesSEXP) {
+Rcpp::NumericMatrix CPP_stats(Rcpp::S4& rast, const Rcpp::NumericVector& rast_ext, const Rcpp::NumericVector& rast_res, Rcpp::Nullable<Rcpp::S4> weights, const Rcpp::RawVector& wkb, double default_value, double default_weight, bool coverage_areas, Rcpp::Nullable<Rcpp::CharacterVector>& p_area_method, const Rcpp::StringVector& stats, int max_cells_in_memory, double grid_compat_tol, const Rcpp::Nullable<Rcpp::NumericVector>& quantiles);
+RcppExport SEXP _exactextractr_CPP_stats(SEXP rastSEXP, SEXP rast_extSEXP, SEXP rast_resSEXP, SEXP weightsSEXP, SEXP wkbSEXP, SEXP default_valueSEXP, SEXP default_weightSEXP, SEXP coverage_areasSEXP, SEXP p_area_methodSEXP, SEXP statsSEXP, SEXP max_cells_in_memorySEXP, SEXP grid_compat_tolSEXP, SEXP quantilesSEXP) {
 BEGIN_RCPP
     Rcpp::RObject rcpp_result_gen;
     Rcpp::RNGScope rcpp_rngScope_gen;
     Rcpp::traits::input_parameter< Rcpp::S4& >::type rast(rastSEXP);
+    Rcpp::traits::input_parameter< const Rcpp::NumericVector& >::type rast_ext(rast_extSEXP);
+    Rcpp::traits::input_parameter< const Rcpp::NumericVector& >::type rast_res(rast_resSEXP);
     Rcpp::traits::input_parameter< Rcpp::Nullable<Rcpp::S4> >::type weights(weightsSEXP);
     Rcpp::traits::input_parameter< const Rcpp::RawVector& >::type wkb(wkbSEXP);
     Rcpp::traits::input_parameter< double >::type default_value(default_valueSEXP);
@@ -67,7 +71,7 @@ BEGIN_RCPP
     Rcpp::traits::input_parameter< int >::type max_cells_in_memory(max_cells_in_memorySEXP);
     Rcpp::traits::input_parameter< double >::type grid_compat_tol(grid_compat_tolSEXP);
     Rcpp::traits::input_parameter< const Rcpp::Nullable<Rcpp::NumericVector>& >::type quantiles(quantilesSEXP);
-    rcpp_result_gen = Rcpp::wrap(CPP_stats(rast, weights, wkb, default_value, default_weight, coverage_areas, p_area_method, stats, max_cells_in_memory, grid_compat_tol, quantiles));
+    rcpp_result_gen = Rcpp::wrap(CPP_stats(rast, rast_ext, rast_res, weights, wkb, default_value, default_weight, coverage_areas, p_area_method, stats, max_cells_in_memory, grid_compat_tol, quantiles));
     return rcpp_result_gen;
 END_RCPP
 }
@@ -106,8 +110,8 @@ END_RCPP
 
 static const R_CallMethodDef CallEntries[] = {
     {"_exactextractr_CPP_coverage_fraction", (DL_FUNC) &_exactextractr_CPP_coverage_fraction, 3},
-    {"_exactextractr_CPP_exact_extract", (DL_FUNC) &_exactextractr_CPP_exact_extract, 17},
-    {"_exactextractr_CPP_stats", (DL_FUNC) &_exactextractr_CPP_stats, 11},
+    {"_exactextractr_CPP_exact_extract", (DL_FUNC) &_exactextractr_CPP_exact_extract, 19},
+    {"_exactextractr_CPP_stats", (DL_FUNC) &_exactextractr_CPP_stats, 13},
     {"_exactextractr_CPP_update_max_coverage", (DL_FUNC) &_exactextractr_CPP_update_max_coverage, 7},
     {"_exactextractr_CPP_resample", (DL_FUNC) &_exactextractr_CPP_resample, 6},
     {NULL, NULL, 0}

diff --git a/src/exact_extract.cpp b/src/exact_extract.cpp
@@ -44,6 +44,8 @@ using exactextract::RasterSource;
 
 // [[Rcpp::export]]
 Rcpp::List CPP_exact_extract(Rcpp::S4 & rast,
+                             const Rcpp::NumericVector & rast_ext,
+                             const Rcpp::NumericVector & rast_res,
                              Rcpp::Nullable<Rcpp::S4> & rast_uncropped,
                              Rcpp::Nullable<Rcpp::S4> & weights,
                              const Rcpp::RawVector & wkb,
@@ -68,7 +70,7 @@ Rcpp::List CPP_exact_extract(Rcpp::S4 & rast,
     auto weights_grid = exactextract::Grid<bounded_extent>::make_empty();
     auto common_grid = grid;
 
-    S4RasterSource rsrc(rast, default_value);
+    S4RasterSource rsrc(rast, rast_ext, rast_res, default_value);
     int src_nlayers = get_nlayers(rast);
 
     std::unique_ptr<S4RasterSource> rweights;
@@ -262,6 +264,8 @@ static int get_num_stats(const Rcpp::StringVector & stats,
 // Return a matrix with one row per stat and one row per raster layer
 // [[Rcpp::export]]
 Rcpp::NumericMatrix CPP_stats(Rcpp::S4 & rast,
+                              const Rcpp::NumericVector & rast_ext,
+                              const Rcpp::NumericVector & rast_res,
                               Rcpp::Nullable<Rcpp::S4> weights,
                               const Rcpp::RawVector & wkb,
                               double default_value,
@@ -281,7 +285,7 @@ Rcpp::NumericMatrix CPP_stats(Rcpp::S4 & rast,
 
     int nlayers = get_nlayers(rast);
 
-    S4RasterSource rsrc(rast, default_value);
+    S4RasterSource rsrc(rast, rast_ext, rast_res, default_value);
 
     std::unique_ptr<S4RasterSource> rweights;
     std::string area_method;

diff --git a/src/s4_raster_source.h b/src/s4_raster_source.h
@@ -36,6 +36,21 @@ class S4RasterSource {
     m_grid = make_grid(rast);
   }
 
+  S4RasterSource(SEXP rast,
+                 const Rcpp::NumericVector & ext,
+                 const Rcpp::NumericVector & res,
+                 double default_value = std::numeric_limits<double>::quiet_NaN()) :
+    m_grid(exactextract::Grid<exactextract::bounded_extent>::make_empty()),
+    m_rast(rast),
+    m_last_box(std::numeric_limits<double>::quiet_NaN(),
+               std::numeric_limits<double>::quiet_NaN(),
+               std::numeric_limits<double>::quiet_NaN(),
+               std::numeric_limits<double>::quiet_NaN()),
+    m_default_value(default_value)
+  {
+    m_grid = make_grid(ext, res);
+  }
+
   const exactextract::Grid<exactextract::bounded_extent> &grid() const {
     return m_grid;
   }