tensor_network_graphs can use V1 and V2 TensorNetworks

examples/tensor_network_graphs/tensor_network_graphs.cpp

@@ -3,6 +3,7 @@
 #include <SeQuant/core/runtime.hpp>
 #include <SeQuant/core/tensor.hpp>
 #include <SeQuant/core/tensor_network.hpp>
+#include <SeQuant/core/tensor_network_v2.hpp>
 #include <SeQuant/domain/mbpt/context.hpp>
 #include <SeQuant/domain/mbpt/convention.hpp>
 
@@ -36,6 +37,22 @@ std::optional<TensorNetwork> to_network(const ExprPtr &expr) {
   }
 }
 
+std::optional<TensorNetworkV2> to_network_v2(const ExprPtr &expr) {
+  if (expr.is<Tensor>()) {
+    return TensorNetworkV2({expr});
+  } else if (expr.is<Product>()) {
+    for (const ExprPtr &factor : expr.as<Product>().factors()) {
+      if (!factor.is<Tensor>()) {
+        return {};
+      }
+    }
+
+    return TensorNetworkV2(expr.as<Product>().factors());
+  } else {
+    return {};
+  }
+}
+
 void print_help() {
   std::wcout << "Helper to generate dot (GraphViz) representations of tensor "
                 "network graphs.\n";
@@ -44,6 +61,7 @@ void print_help() {
       << "  <exe> [options] <network 1> [<network 2> [... [<network N>] ] ]\n";
   std::wcout << "Options:\n";
   std::wcout << "  --help     Shows this help message\n";
+  std::wcout << "  --v2       Use TensorNetworkV2\n";
   std::wcout << "  --no-named Treat all indices as unnamed (even if they are "
                 "external)\n";
 }
@@ -55,6 +73,7 @@ int main(int argc, char **argv) {
       BraKetSymmetry::conjugate, SPBasis::spinorbital));
 
   bool use_named_indices = true;
+  bool use_tnv2 = false;
   const TensorNetwork::named_indices_t empty_named_indices;
 
   if (argc <= 1) {
@@ -70,6 +89,9 @@ int main(int argc, char **argv) {
     } else if (current == L"--no-named") {
       use_named_indices = false;
       continue;
+    } else if (current == L"--v2") {
+      use_tnv2 = true;
+      continue;
     }
 
     ExprPtr expr;
@@ -82,16 +104,30 @@ int main(int argc, char **argv) {
     }
     assert(expr);
 
-    std::optional<TensorNetwork> network = to_network(expr);
-    if (!network.has_value()) {
-      std::wcout << "Failed to construct tensor network for input '" << current
-                 << "'" << std::endl;
-      return 2;
-    }
+    if (!use_tnv2) {
+      std::optional<TensorNetwork> network = to_network(expr);
+      if (!network.has_value()) {
+        std::wcout << "Failed to construct tensor network for input '"
+                   << current << "'" << std::endl;
+        return 2;
+      }
+
+      auto [graph, vlabels, vcolors, vtypes] = network->make_bliss_graph(
+          use_named_indices ? nullptr : &empty_named_indices);
+      std::wcout << "Graph for '" << current << "'\n";
+      graph->write_dot(std::wcout, vlabels);
+    } else {
+      std::optional<TensorNetworkV2> network = to_network_v2(expr);
+      if (!network.has_value()) {
+        std::wcout << "Failed to construct tensor network for input '"
+                   << current << "'" << std::endl;
+        return 2;
+      }
 
-    TensorNetwork::Graph graph = network->create_graph(
-        use_named_indices ? nullptr : &empty_named_indices);
-    std::wcout << "Graph for '" << current << "'\n";
-    graph.bliss_graph->write_dot(std::wcout, graph.vertex_labels);
+      TensorNetworkV2::Graph graph = network->create_graph(
+          use_named_indices ? nullptr : &empty_named_indices);
+      std::wcout << "Graph for '" << current << "'\n";
+      graph.bliss_graph->write_dot(std::wcout, graph.vertex_labels);
+    }
   }
 }