add vg validate (resolves #126)

main.cpp

@@ -25,6 +25,106 @@ using namespace std;
 using namespace google::protobuf;
 using namespace vg;
 
+void help_validate(char** argv) {
+    cerr << "usage: " << argv[0] << " validate [options] graph" << endl
+         << "Validate the graph." << endl
+         << endl
+         << "options:" << endl
+         << "    default: check all aspects of the graph, if options are specified do only those" << endl
+         << "    -n, --nodes    verify that we have the expected number of nodes" << endl
+         << "    -e, --edges    verify that the graph contains all nodes that are referred to by edges" << endl
+         << "    -p, --paths    verify that contiguous path segments are connected by edges" << endl
+         << "    -o, --orphans  verify that all nodes have edges" << endl;
+}
+
+int main_validate(int argc, char** argv) {
+
+    if (argc <= 2) {
+        help_validate(argv);
+        return 1;
+    }
+
+    bool check_nodes = false;
+    bool check_edges = false;
+    bool check_orphans = false;
+    bool check_paths = false;
+
+    int c;
+    optind = 2; // force optind past command positional argument
+    while (true) {
+        static struct option long_options[] =
+            {
+                {"help", no_argument, 0, 'h'},
+                {"nodes", no_argument, 0, 'n'},
+                {"edges", no_argument, 0, 'e'},
+                {"paths", no_argument, 0, 'o'},
+                {"orphans", no_argument, 0, 'p'},
+                {0, 0, 0, 0}
+            };
+
+        int option_index = 0;
+        c = getopt_long (argc, argv, "hneop",
+                         long_options, &option_index);
+
+        // Detect the end of the options.
+        if (c == -1)
+            break;
+
+        switch (c)
+        {
+
+        case 'n':
+            check_nodes = true;
+            break;
+
+        case 'e':
+            check_edges = true;
+            break;
+
+        case 'o':
+            check_orphans = true;
+            break;
+
+        case 'p':
+            check_paths = true;
+            break;
+
+        case 'h':
+        case '?':
+            help_validate(argv);
+            exit(1);
+            break;
+
+        default:
+            abort ();
+        }
+    }
+
+    VG* graph;
+    string file_name = argv[optind];
+    if (file_name == "-") {
+        graph = new VG(std::cin);
+    } else {
+        ifstream in;
+        in.open(file_name.c_str());
+        graph = new VG(in);
+    }
+
+    // if we chose a specific subset, do just them
+    if (check_nodes || check_edges || check_orphans || check_paths) {
+        if (graph->is_valid(check_nodes, check_edges, check_orphans, check_paths)) {
+            return 0;
+        } else {
+            return 1;
+        }
+    // otherwise do everything
+    } else if (graph->is_valid()) {
+        return 0;
+    } else {
+        return 1;
+    }
+}
+
 void help_compare(char** argv) {
     cerr << "usage: " << argv[0] << " compare [options] graph1 graph2" << endl
          << "Compare kmer sets of two graphs" << endl
@@ -900,7 +1000,7 @@ int main_msga(int argc, char** argv) {
         // use this step to simplify the graph so we can efficiently normalize it
         graph->remove_non_path();
         graph->paths.clear();
-        graph->graph.clear_path();
+        graph->graph.clear_path(); // paths.clear() should do this too
         graph->normalize();
         graph->dice_nodes(node_max);
         graph->sort();
@@ -4674,7 +4774,8 @@ void vg_help(char** argv) {
          << "  -- msga          multiple sequence graph alignment" << endl
          << "  -- pileup        build a pileup from a set of alignments" << endl
          << "  -- call          prune the graph by genotyping a pileup" << endl
-         << "  -- compare       compare the kmer space of two graphs" << endl;
+         << "  -- compare       compare the kmer space of two graphs" << endl
+         << "  -- validate      validate the semantics of a graph" << endl;
 }
 
 int main(int argc, char *argv[])
@@ -4726,6 +4827,8 @@ int main(int argc, char *argv[])
         return main_call(argc, argv);
     } else if (command == "compare") {
         return main_compare(argc, argv);
+    } else if (command == "validate") {
+        return main_validate(argc, argv);
     } else {
         cerr << "error:[vg] command " << command << " not found" << endl;
         vg_help(argv);