[FEATURE] Layout can be computed on minimizers.

include/chopper/configuration.hpp

@@ -40,6 +40,9 @@ struct configuration
     //!\brief The kmer size to hash the input sequences before computing a HyperLogLog sketch from them.
     uint8_t k{19};
 
+    //!\brief The window size to compute minimizers before computing a HyperLogLog sketch from them.
+    uint8_t window_size{19};
+
     //!\brief Do not write the sketches into a dedicated directory.
     bool disable_sketch_output{false};
 
@@ -83,6 +86,7 @@ struct configuration
         archive(CEREAL_NVP(debug));
         archive(CEREAL_NVP(sketch_directory));
         archive(CEREAL_NVP(k));
+        archive(CEREAL_NVP(window_size));
         archive(CEREAL_NVP(disable_sketch_output));
         archive(CEREAL_NVP(precomputed_files));
 

include/chopper/input_functor.hpp

@@ -37,6 +37,8 @@ struct input_functor
 
     uint8_t kmer_size{21};
 
+    uint8_t window_size{21};
+
     void operator()(size_t const num, seqan::hibf::insert_iterator it);
 };
 

src/chopper.cpp

@@ -56,7 +56,8 @@ int main(int argc, char const * argv[])
 
         chopper::sketch::check_filenames(filenames, config);
 
-        config.hibf_config.input_fn = chopper::input_functor{filenames, config.precomputed_files, config.k};
+        config.hibf_config.input_fn =
+            chopper::input_functor{filenames, config.precomputed_files, config.k, config.window_size};
         config.hibf_config.number_of_user_bins = filenames.size();
 
         exit_code |= chopper::layout::execute(config, filenames);

src/input_functor.cpp

@@ -43,7 +43,7 @@ void input_functor::operator()(size_t const num, seqan::hibf::insert_iterator it
 
         seqan3::shape shape = seqan3::ungapped{kmer_size};
         auto minimizer_view = seqan3::views::minimiser_hash(shape,
-                                                            seqan3::window_size{kmer_size},
+                                                            seqan3::window_size{window_size},
                                                             seqan3::seed{adjust_seed(shape.count())});
 
         for (auto && [seq] : fin)

src/set_up_parser.cpp

@@ -64,6 +64,19 @@ void set_up_parser(sharg::parser & parser, configuration & config)
                 "k-mer size might miss out on certain similarities. For DNA sequences, a k-mer size between "
                 "[16,32] has proven to work well."});
 
+    parser.add_option(
+        config.window_size,
+        sharg::config{
+            .short_id = '\0',
+            .long_id = "window",
+            .description =
+                "Setting this option will trigger the computation of (w,k)-minimizers instead of canonical kmers. "
+                "Minimizers can thin out the data, reduce the memory footpring of the resulting index and increase "
+                "runtime performance. On the other hand, it also decreases accuracy and might cause false negatives."
+                "For DNA sequences, a window size of 2-4 positions more than the kmers size, e.g. (42,20)-minimizers, "
+                "has proven to reduce the computational effort significantly while only slightly descreasing the "
+                "accuracy."});
+
     parser.add_option(
         config.hibf_config.tmax,
         sharg::config{

src/util/display_layout/general.cpp

@@ -297,7 +297,13 @@ int execute(config const & cfg)
 
             for (auto const & filename : filenames[user_bin.idx])
             {
-                process_file(filename, current_kmer_set, current_kmers, sketch, fill_current_kmers, chopper_config.k);
+                process_file(filename,
+                             current_kmer_set,
+                             current_kmers,
+                             sketch,
+                             fill_current_kmers,
+                             chopper_config.k,
+                             chopper_config.window_size);
             }
 
             // Compute set intersection: shared_kmers = shared_kmers ∩ current_kmers

src/util/display_layout/process_file.cpp

@@ -34,7 +34,8 @@ void process_file(std::string const & filename,
                   std::vector<uint64_t> & current_kmers,
                   seqan::hibf::sketch::hyperloglog & sketch,
                   bool const fill_current_kmers,
-                  uint8_t const kmer_size)
+                  uint8_t const kmer_size,
+                  uint8_t const window_size)
 {
     if (filename.ends_with(".minimiser"))
     {
@@ -68,7 +69,7 @@ void process_file(std::string const & filename,
 
         seqan3::shape shape{seqan3::ungapped{kmer_size}};
         auto minimizer_view = seqan3::views::minimiser_hash(shape,
-                                                            seqan3::window_size{kmer_size},
+                                                            seqan3::window_size{window_size},
                                                             seqan3::seed{chopper::adjust_seed(shape.count())});
         if (fill_current_kmers)
         {
@@ -96,7 +97,10 @@ void process_file(std::string const & filename,
     }
 }
 
-void process_file(std::string const & filename, std::vector<uint64_t> & current_kmers, uint8_t const kmer_size)
+void process_file(std::string const & filename,
+                  std::vector<uint64_t> & current_kmers,
+                  uint8_t const kmer_size,
+                  uint8_t const window_size)
 {
     if (filename.ends_with(".minimiser"))
     {
@@ -115,7 +119,7 @@ void process_file(std::string const & filename, std::vector<uint64_t> & current_
 
         seqan3::shape shape{seqan3::ungapped{kmer_size}};
         auto minimizer_view = seqan3::views::minimiser_hash(shape,
-                                                            seqan3::window_size{kmer_size},
+                                                            seqan3::window_size{window_size},
                                                             seqan3::seed{chopper::adjust_seed(shape.count())});
 
         for (auto && [seq] : fin)

src/util/display_layout/shared.hpp

@@ -25,11 +25,15 @@ struct config
 void execute_general(config const & cfg);
 void execute_sizes(config const & cfg);
 
-void process_file(std::string const & filename, std::vector<uint64_t> & current_kmers, uint8_t const kmer_size);
+void process_file(std::string const & filename,
+                  std::vector<uint64_t> & current_kmers,
+                  uint8_t const kmer_size,
+                  uint8_t const window_size);
 
 void process_file(std::string const & filename,
                   robin_hood::unordered_set<uint64_t> & current_kmer_set,
                   std::vector<uint64_t> & current_kmers,
                   seqan::hibf::sketch::hyperloglog & sketch,
                   bool const fill_current_kmers,
-                  uint8_t const kmer_size);
+                  uint8_t const kmer_size,
+                  uint8_t const window_size);

src/util/display_layout/sizes.cpp

@@ -301,7 +301,7 @@ void execute_general_stats(config const & cfg)
         if (filenames[user_bin_id].size() > 1)
             throw std::runtime_error{"No multi files accepted yet."};
 
-        process_file(filenames[user_bin_id][0], current_kmers, chopper_config.k);
+        process_file(filenames[user_bin_id][0], current_kmers, chopper_config.k, chopper_config.window_size);
 
         for (auto const kmer : current_kmers)
             it = kmer;

test/api/config_test.cpp

@@ -16,6 +16,7 @@ chopper::configuration generate_config()
     config.data_file = "/path/to/data.file";
     config.sketch_directory = "/path/to/sketch/directory";
     config.k = 20;
+    config.window_size = 24;
     config.disable_sketch_output = true;
     config.precomputed_files = true;
     config.output_filename = "file.layout";
@@ -47,6 +48,7 @@ bool operator==(chopper::configuration const & lhs, chopper::configuration const
            lhs.debug == rhs.debug &&                                                               //
            lhs.sketch_directory == rhs.sketch_directory &&                                         //
            lhs.k == rhs.k &&                                                                       //
+           lhs.window_size == rhs.window_size &&                                                   //
            lhs.disable_sketch_output == rhs.disable_sketch_output &&                               //
            lhs.precomputed_files == rhs.precomputed_files &&                                       //
            lhs.output_filename == rhs.output_filename &&                                           //
@@ -79,6 +81,7 @@ static constexpr std::string_view config_string_view{"@CHOPPER_CONFIG\n"
                                                      "@            \"value0\": \"/path/to/sketch/directory\"\n"
                                                      "@        },\n"
                                                      "@        \"k\": 20,\n"
+                                                     "@        \"window_size\": 24,\n"
                                                      "@        \"disable_sketch_output\": true,\n"
                                                      "@        \"precomputed_files\": true,\n"
                                                      "@        \"output_filename\": {\n"

test/api/layout/execute_layout_test.cpp

@@ -66,6 +66,7 @@ TEST(execute_test, few_ubs)
                                     "@            \"value0\": \"\"\n"
                                     "@        },\n"
                                     "@        \"k\": 19,\n"
+                                    "@        \"window_size\": 19,\n"
                                     "@        \"disable_sketch_output\": true,\n"
                                     "@        \"precomputed_files\": false,\n"
                                     "@        \"output_filename\": {\n"
@@ -275,6 +276,7 @@ TEST(execute_test, many_ubs)
                                     "@            \"value0\": \"\"\n"
                                     "@        },\n"
                                     "@        \"k\": 19,\n"
+                                    "@        \"window_size\": 19,\n"
                                     "@        \"disable_sketch_output\": true,\n"
                                     "@        \"precomputed_files\": false,\n"
                                     "@        \"output_filename\": {\n"

test/api/layout/execute_with_estimation_test.cpp

@@ -246,6 +246,7 @@ TEST(execute_estimation_test, many_ubs)
                                     "@            \"value0\": \"\"\n"
                                     "@        },\n"
                                     "@        \"k\": 19,\n"
+                                    "@        \"window_size\": 19,\n"
                                     "@        \"disable_sketch_output\": true,\n"
                                     "@        \"precomputed_files\": false,\n"
                                     "@        \"output_filename\": {\n"

test/cli/cli_chopper_pipeline_test.cpp

@@ -176,6 +176,7 @@ TEST_F(cli_test, chopper_layout2)
                                       "@            \"value0\": \"\"\n"
                                       "@        },\n"
                                       "@        \"k\": 19,\n"
+                                      "@        \"window_size\": 19,\n"
                                       "@        \"disable_sketch_output\": true,\n"
                                       "@        \"precomputed_files\": false,\n"
                                       "@        \"output_filename\": {\n"