Add make gapless input option

src/commands/convert/sync.cpp

@@ -101,6 +101,15 @@ void setup_sync( CLI::App& app )
     );
     options->gapless.option->group( "Settings" );
 
+    // For user convenience, we also check if the general settings option for making
+    // the stream gapless is given, to improve the option description here.
+    auto make_gapless_opt = sub->get_option_no_throw( "--make-gapless" );
+    if( make_gapless_opt ) {
+        auto descr = options->gapless.option->get_description();
+        descr += " Note: This option is an alias for the more general `--make-gapless` option.";
+        options->gapless.option->description( descr );
+    }
+
     // Guess Reference Base
     options->guess_ref_base.option = sub->add_flag(
         "--guess-reference-base",

src/options/variant_input.cpp

@@ -150,6 +150,26 @@ void VariantInputOptions::add_input_files_opts_to_app(
         CLI::IsMember( enum_map_keys( multi_file_contribution_type_map_ ), CLI::ignore_case )
     );
 
+    // Make gapless stream. If renamed, change it in the `sync` command as well,
+    // where we use the option name to imporove the option description there.
+    make_gapless_stream_.option = sub->add_flag(
+        "--make-gapless",
+        make_gapless_stream_.value,
+        "By default, we only operate on the positions for which there is data. "
+        "In particular, positions that are absent in the input are completely ignored; "
+        "they do not even show up in the `missing` column of output tables. "
+        "This is because for the statistics, data being absend or (marked as) missing "
+        "is merely a sementic distinction, but it does not change the results. "
+        "However, it might make processing with downstream tools easier if the output contains "
+        "all positions, for instance when using `single` windows. "
+        "With this option, all absent positions are filled in as missing data, so that they "
+        "show up in the `missing` column and as entries in single windows. "
+        "If a referene genome or dictionary is given, this might also include positions beyond "
+        "where there is input data, up until the length of each chromosome. "
+        "Note that this can lead to large ouput tables when processing single positions."
+    );
+    make_gapless_stream_.option->group( group );
+
     // Hidden options to set the Generic Input Stream block sizes for speed.
 
     // First for the main block size of the stream that is collecing all Variants,
@@ -680,6 +700,11 @@ void VariantInputOptions::conditionally_make_gapless_stream_() const
         return;
     }
 
+    // Also if the user provided the gapless option here.
+    if( make_gapless_stream_.value ) {
+        gapless_stream_ = true;
+    }
+
     // Now we check if we want to make a gapless stream, e.g., by demand from a command.
     // If so, we further check if any references are given, which we then want to use
     // for the stream in order to properly get the chromosome lengths.

src/options/variant_input.hpp

@@ -155,6 +155,16 @@ class VariantInputOptions
         gapless_stream_ = value;
     }
 
+    /**
+     * @brief Return whether the stream is made gapless due to some setting.
+     *
+     * This option is set to its correct value only after get_stream() has been called.
+     */
+    bool gapless_stream() const
+    {
+        return gapless_stream_;
+    }
+
     /**
      * @brief Transformations and filters for individual input sources.
      *
@@ -320,6 +330,7 @@ class VariantInputOptions
     // to avoid code repetition when adding and processing them here.
     std::vector<std::unique_ptr<VariantFileOptions>> input_files_;
     CliOption<std::string> multi_file_loci_set_ = "union";
+    CliOption<bool> make_gapless_stream_ = false;
 
     // Hidden options to set the Generic Input Stream block size for speed.
     CliOption<size_t> iterator_block_size_ = 8192;

src/options/window.cpp

@@ -326,7 +326,7 @@ std::unique_ptr<VariantWindowStream> WindowOptions::get_variant_window_stream(
         }
         case WindowType::kSingle: {
             result = genesis::utils::make_unique<VariantPositionWindowStream>(
-                get_variant_window_stream_single_( input_stream )
+                get_variant_window_stream_single_( input_stream, variant_input.gapless_stream() )
             );
             break;
         }
@@ -413,7 +413,7 @@ std::unique_ptr<VariantWindowViewStream> WindowOptions::get_variant_window_view_
         case WindowType::kSingle: {
             result = genesis::utils::make_unique<WindowViewStream>(
                 make_window_view_stream(
-                    get_variant_window_stream_single_( input_stream )
+                    get_variant_window_stream_single_( input_stream, variant_input.gapless_stream() )
                 )
             );
             break;
@@ -598,8 +598,19 @@ WindowOptions::get_variant_window_stream_queue_(
 
 WindowOptions::VariantPositionWindowStream
 WindowOptions::get_variant_window_stream_single_(
-    genesis::population::VariantInputStream& input
+    genesis::population::VariantInputStream& input,
+    bool is_gapless_stream
 ) const {
+    // If our input is meant to be gapless, we also do not want to filter for passing positions here.
+    // That will produce huge output, but well, it's want the user wants :-)
+    // As far as we can tell, this is the only type of window that needs this;
+    // all other window types behave exactly the same, except that they also are counting
+    // the missing positions properly when a gapless stream is given.
+    if( is_gapless_stream ) {
+        return genesis::population::make_default_position_window_stream(
+            input.begin(), input.end()
+        );
+    }
     return genesis::population::make_passing_variant_position_window_stream(
         input.begin(), input.end()
     );

src/options/window.hpp

@@ -243,7 +243,8 @@ class WindowOptions
     ) const;
 
     VariantPositionWindowStream get_variant_window_stream_single_(
-        genesis::population::VariantInputStream& input
+        genesis::population::VariantInputStream& input,
+        bool is_gapless_stream
     ) const;
 
     VariantRegionWindowStream get_variant_window_stream_regions_(