Merge pull request #2474 from vgteam/packed-call

[wip] Clean up vg augment

src/augment.cpp

@@ -15,7 +15,7 @@ void augment(MutablePathMutableHandleGraph* graph,
              istream& gam_stream,
              vector<Translation>* out_translations,
              ostream* gam_out_stream,
-             function<void(Path&)> save_path_fn,
+             bool embed_paths,
              bool break_at_ends,
              bool remove_softclips) {
 
@@ -32,7 +32,7 @@ void augment(MutablePathMutableHandleGraph* graph,
                  iterate_gam,
                  out_translations,
                  gam_out_stream,
-                 save_path_fn,
+                 embed_paths,
                  break_at_ends,
                  remove_softclips);
 }
@@ -41,7 +41,7 @@ void augment(MutablePathMutableHandleGraph* graph,
              vector<Path>& path_vector,
              vector<Translation>* out_translations,
              ostream* gam_out_stream,
-             function<void(Path&)> save_path_fn,
+             bool embed_paths,
              bool break_at_ends,
              bool remove_softclips) {
 
@@ -59,7 +59,7 @@ void augment(MutablePathMutableHandleGraph* graph,
                  iterate_gam,
                  out_translations,
                  gam_out_stream,
-                 save_path_fn,
+                 embed_paths,
                  break_at_ends,
                  remove_softclips);
 }
@@ -68,7 +68,7 @@ void augment_impl(MutablePathMutableHandleGraph* graph,
                   function<void(function<void(Alignment&)>, bool)> iterate_gam,
                   vector<Translation>* out_translations,
                   ostream* gam_out_stream,
-                  function<void(Path&)> save_path_fn,
+                  bool embed_paths,
                   bool break_at_ends,
                   bool remove_softclips) {
     // Collect the breakpoints
@@ -138,8 +138,8 @@ void augment_impl(MutablePathMutableHandleGraph* graph,
             // Copy over the name
             *added.mutable_name() = aln.name();
 
-            if (save_path_fn) {
-                save_path_fn(added);
+            if (embed_paths) {
+                add_path_to_graph(graph, added);
             }
 
             // something is off about this check.
@@ -198,6 +198,19 @@ void augment_impl(MutablePathMutableHandleGraph* graph,
     if (out_translations != nullptr) {
         *out_translations = make_translation(graph, node_translation, added_nodes, orig_node_sizes);
     }
+
+    VG* vg_graph = dynamic_cast<VG*>(graph);
+
+    // This code got run after augment in VG::edit, so we make sure it happens here too
+    if (vg_graph != nullptr) {
+        // Rebuild path ranks, aux mapping, etc. by compacting the path ranks
+        // Todo: can we just do this once?
+       vg_graph->paths.compact_ranks();
+
+       // execute a semi partial order sort on the nodes
+       vg_graph->sort();
+    }
+
 }
 
 
@@ -297,6 +310,15 @@ void find_breakpoints(const Path& path, unordered_map<id_t, set<pos_t>>& breakpo
 
 }
 
+path_handle_t add_path_to_graph(MutablePathHandleGraph* graph, const Path& path) {
+    path_handle_t path_handle = graph->create_path_handle(path.name(), path.is_circular());
+    for (int i = 0; i < path.mapping_size(); ++i) {
+        graph->append_step(path_handle, graph->get_handle(path.mapping(i).position().node_id(),
+                                                          path.mapping(i).position().is_reverse()));
+    }
+    return path_handle;
+}
+
 unordered_map<id_t, set<pos_t>> forwardize_breakpoints(const HandleGraph* graph,
                                                        const unordered_map<id_t, set<pos_t>>& breakpoints) {
     unordered_map<id_t, set<pos_t>> fwd;

src/augment.hpp

@@ -24,7 +24,7 @@ using namespace std;
 /// If out_translation is not null, a list of translations, one per node existing
 /// after the edit, describing
 /// how each new or conserved node is embedded in the old graph. 
-/// if save_path_fn is not null, this function will be run on each modified path
+/// if embed_paths is true, then the augmented alignemnents will be saved as embededed paths in the graph
 /// in order to add it back to the graph.
 /// If break_at_ends is true, nodes will be broken at
 /// the ends of paths that start/end woth perfect matches, so the paths can
@@ -35,7 +35,7 @@ void augment(MutablePathMutableHandleGraph* graph,
              istream& gam_stream,
              vector<Translation>* out_translation = nullptr,
              ostream* gam_out_stream = nullptr,
-             function<void(Path&)> save_path_fn = nullptr,
+             bool embed_paths = false,
              bool break_at_ends = false,
              bool remove_soft_clips = false);
 
@@ -45,7 +45,7 @@ void augment(MutablePathMutableHandleGraph* graph,
              vector<Path>& path_vector,
              vector<Translation>* out_translation = nullptr,
              ostream* gam_out_stream = nullptr,
-             function<void(Path&)> save_path_fn = nullptr,
+             bool embed_paths = false,
              bool break_at_ends = false,
              bool remove_soft_clips = false);
 
@@ -54,10 +54,14 @@ void augment_impl(MutablePathMutableHandleGraph* graph,
                   function<void(function<void(Alignment&)>, bool)> iterate_gam,
                   vector<Translation>* out_translation,
                   ostream* gam_out_stream,
-                  function<void(Path&)> save_path_fn,
+                  bool embed_paths,
                   bool break_at_ends,
                   bool remove_soft_clips);
 
+/// Add a path to the graph.  This is like VG::extend, and expects
+/// a path with no edits, and for all the nodes and edges in the path
+/// to exist exactly in the graph
+path_handle_t add_path_to_graph(MutablePathHandleGraph* graph, const Path& path);
 
 /// Find all the points at which a Path enters or leaves nodes in the graph. Adds
 /// them to the given map by node ID of sets of bases in the node that will need

src/io/register_loader_saver_hash_graph.cpp

@@ -17,7 +17,7 @@ using namespace std;
 using namespace vg::io;
 
 void register_loader_saver_hash_graph() {
-  Registry::register_bare_loader_saver<bdsg::HashGraph, PathHandleGraph, HandleGraph>("HashGraph", [](istream& input) -> void* {
+  Registry::register_bare_loader_saver<bdsg::HashGraph, MutablePathMutableHandleGraph, MutableHandleGraph, PathHandleGraph, HandleGraph>("HashGraph", [](istream& input) -> void* {
         // Allocate a HashGraph
          bdsg::HashGraph* hash_graph = new bdsg::HashGraph();
 

src/io/register_loader_saver_odgi.cpp

@@ -17,9 +17,9 @@ using namespace std;
 using namespace vg::io;
 
 void register_loader_saver_odgi() {
-  Registry::register_bare_loader_saver<bdsg::ODGI, PathHandleGraph, HandleGraph>("PackedGraph", [](istream& input) -> void* {
+    Registry::register_bare_loader_saver<bdsg::ODGI, MutablePathMutableHandleGraph, MutableHandleGraph, PathHandleGraph, HandleGraph>("PackedGraph", [](istream& input) -> void* {
         // Allocate a PackedGraph
-         bdsg::ODGI* odgi = new bdsg::ODGI();
+        bdsg::ODGI* odgi = new bdsg::ODGI();
 
         // Load it
         odgi->load(input);

src/io/register_loader_saver_packed_graph.cpp

@@ -17,7 +17,7 @@ using namespace std;
 using namespace vg::io;
 
 void register_loader_saver_packed_graph() {
-  Registry::register_bare_loader_saver<bdsg::PackedGraph, PathHandleGraph, HandleGraph>("PackedGraph", [](istream& input) -> void* {
+  Registry::register_bare_loader_saver<bdsg::PackedGraph, MutablePathMutableHandleGraph, MutableHandleGraph, PathHandleGraph, HandleGraph>("PackedGraph", [](istream& input) -> void* {
         // Allocate a PackedGraph
          bdsg::PackedGraph* packed_graph = new bdsg::PackedGraph();
 

src/io/register_loader_saver_vg.cpp

@@ -17,7 +17,7 @@ using namespace vg::io;
 
 void register_loader_saver_vg() {
     // We register for "" so we can handle untagged old-style vg files and make them into HandleGraphs
-    Registry::register_loader_saver<VG, PathHandleGraph, HandleGraph>(vector<string>{"VG", ""},
+    Registry::register_loader_saver<VG, MutablePathMutableHandleGraph, MutableHandleGraph, PathHandleGraph, HandleGraph>(vector<string>{"VG", ""},
         [](const message_sender_function_t& for_each_message) -> void* {
         // We have a bit of a control problem.
         // The source function wants to drive; we give it a function of strings, and it calls it with all the strings in turn.