auto-magically import GFAz when a user provides a graph

Author: babyplutokurt

.gitignore

@@ -51,3 +51,4 @@ debug_main.cpp
 /*.fa
 /*.gfa
 .vscode/*
+.cache/*

src/algorithms/gfa_to_handle.cpp

@@ -1,5 +1,7 @@
 #include "gfa_to_handle.hpp"
+#include "gfaz_to_handle.hpp"
 #include "../path.hpp"
+#include "../utility.hpp"
 
 #include <gbwtgraph/utils.h>
 
@@ -325,19 +327,23 @@ static void add_path_listeners(GFAParser& parser, MutablePathMutableHandleGraph*
 }
 
 void gfa_to_handle_graph(const string& filename, MutableHandleGraph* graph,
-                         GFAIDMapInfo* translation) {
+                         GFAIDMapInfo* translation, int num_threads) {
+    if (filename != "-" && ends_with(filename, ".gfaz")) {
+        gfaz_to_handle_graph(filename, graph, translation, num_threads);
+        return;
+    }
 
     get_input_file(filename, [&](istream& in) {
        gfa_to_handle_graph(in, graph, translation);
     });
 }
 
 void gfa_to_handle_graph(const string& filename, MutableHandleGraph* graph,
-                         const string& translation_filename) {
+                         const string& translation_filename, int num_threads) {
 
 
     GFAIDMapInfo id_map_info;
-    gfa_to_handle_graph(filename, graph, &id_map_info);
+    gfa_to_handle_graph(filename, graph, &id_map_info, num_threads);
     write_gfa_translation(id_map_info, translation_filename);
 }
 
@@ -357,19 +363,23 @@ void gfa_to_handle_graph(istream& in, MutableHandleGraph* graph,
 
 void gfa_to_path_handle_graph(const string& filename, MutablePathMutableHandleGraph* graph,
                               GFAIDMapInfo* translation, int64_t max_rgfa_rank,
-                              unordered_set<PathSense>* ignore_sense) {
+                              unordered_set<PathSense>* ignore_sense, int num_threads) {
+    if (filename != "-" && ends_with(filename, ".gfaz")) {
+        gfaz_to_path_handle_graph(filename, graph, translation, max_rgfa_rank, ignore_sense, num_threads);
+        return;
+    }
 
     get_input_file(filename, [&](istream& in) {
-        gfa_to_path_handle_graph(in, graph, translation, max_rgfa_rank);
+        gfa_to_path_handle_graph(in, graph, translation, max_rgfa_rank, ignore_sense);
     });
 }
 
 void gfa_to_path_handle_graph(const string& filename, MutablePathMutableHandleGraph* graph,
                               int64_t max_rgfa_rank, const string& translation_filename,
-                              unordered_set<PathSense>* ignore_sense) {
+                              unordered_set<PathSense>* ignore_sense, int num_threads) {
 
     GFAIDMapInfo id_map_info;
-    gfa_to_path_handle_graph(filename, graph, &id_map_info, max_rgfa_rank);
+    gfa_to_path_handle_graph(filename, graph, &id_map_info, max_rgfa_rank, ignore_sense, num_threads);
     write_gfa_translation(id_map_info, translation_filename);
 
 }

src/algorithms/gfa_to_handle.hpp

@@ -60,33 +60,39 @@ struct GFAIDMapInfo : public NamedNodeBackTranslation {
 /// std::ios_base::failure if an IO operation fails. Throws invalid_argument if
 /// otherwise misused.
 /// Does not give max ID hints, and so might be very slow when loading into an ODGI graph.
+/// num_threads currently only affects GFAZ input, and is ignored for plain GFA.
 void gfa_to_handle_graph(const string& filename,
                          MutableHandleGraph* graph,
-                         GFAIDMapInfo* translation = nullptr);
+                         GFAIDMapInfo* translation = nullptr,
+                         int num_threads = 0);
 
 /// Overload which serializes its translation to a file internally.
 void gfa_to_handle_graph(const string& filename,
                          MutableHandleGraph* graph,
-                         const string& translation_filename);
+                         const string& translation_filename,
+                         int num_threads = 0);
 
 /// Load a GFA from a stream (assumed not to be seekable or reopenable) into a HandleGraph.
 void gfa_to_handle_graph(istream& in,
                          MutableHandleGraph* graph,
                          GFAIDMapInfo* translation = nullptr);
 
 /// Same as gfa_to_handle_graph but also adds path elements from the GFA to the graph.
+/// num_threads currently only affects GFAZ input, and is ignored for plain GFA.
 void gfa_to_path_handle_graph(const string& filename,
                               MutablePathMutableHandleGraph* graph,
                               GFAIDMapInfo* translation = nullptr,
                               int64_t max_rgfa_rank = numeric_limits<int64_t>::max(),
-                              unordered_set<PathSense>* ignore_sense = nullptr);
+                              unordered_set<PathSense>* ignore_sense = nullptr,
+                              int num_threads = 0);
 
 /// Overload which serializes its translation to a file internally.
 void gfa_to_path_handle_graph(const string& filename,
                               MutablePathMutableHandleGraph* graph,
                               int64_t max_rgfa_rank,
                               const string& translation_filename,
-                              unordered_set<PathSense>* ignore_sense = nullptr);
+                              unordered_set<PathSense>* ignore_sense = nullptr,
+                              int num_threads = 0);
 
 /// Load a GFA from a stream (assumed not to be seekable or reopenable) into a PathHandleGraph.
 void gfa_to_path_handle_graph(istream& in,

src/algorithms/gfaz_to_handle.cpp

@@ -205,7 +205,11 @@ parse_reference_samples(const GfaGraph &gfaz_graph) {
                    gbwtgraph::REFERENCE_SAMPLE_LIST_GFA_TAG.end(),
                    tag.begin()) &&
         tag[2] == ':' && tag[3] == 'Z' && tag[4] == ':') {
-      reference_samples = gbwtgraph::parse_reference_samples_tag(tag.substr(5));
+      auto parsed_reference_samples =
+          gbwtgraph::parse_reference_samples_tag(tag.substr(5));
+      reference_samples.clear();
+      reference_samples.insert(parsed_reference_samples.begin(),
+                               parsed_reference_samples.end());
     }
     start = end + 1;
   }