From 4ea67c70a58acc55a8eee44af20f8bf0887577b7 Mon Sep 17 00:00:00 2001
From: David Cosgrove <davidacosgroveaz@gmail.com>
Date: Mon, 14 Apr 2025 17:03:50 +0100
Subject: [PATCH] Support pickling Shape inputs (#8434)

* Stash for transfer.

* Basic serialization.

* Serialization of ShapeInput.

* Default c'tor and operator=.

* Response to review.

* Add cmake to macOS CI image.

* Add make, not cmake.

* Conditional serialization test.

---------

Co-authored-by: David Cosgrove <david@cozchemix.co.uk>
---
 .azure-pipelines/mac_build.yml          |  2 +-
 .azure-pipelines/mac_build_java.yml     |  2 +-
 External/pubchem_shape/PubChemShape.hpp | 56 +++++++++++++++++++++++--
 External/pubchem_shape/test.cpp         | 21 ++++++++++
 4 files changed, 76 insertions(+), 5 deletions(-)

diff --git a/.azure-pipelines/mac_build.yml b/.azure-pipelines/mac_build.yml
index d46140deb..8d68d379e 100644
--- a/.azure-pipelines/mac_build.yml
+++ b/.azure-pipelines/mac_build.yml
@@ -27,7 +27,7 @@ steps:
         libboost-python=$(boost_version) libboost-python-devel=$(boost_version) \
         qt \
         numpy matplotlib=3.8 cairo pillow eigen pandas=2.1  \
-        jupyter=1.0 ipython=8.20 sphinx myst-parser pytest nbval
+        jupyter=1.0 ipython=8.20 sphinx myst-parser pytest nbval make
   displayName: Setup build environment
 - bash: |
     export CONDA=/tmp/miniforge
diff --git a/.azure-pipelines/mac_build_java.yml b/.azure-pipelines/mac_build_java.yml
index 6ffc8c36f..9d2bceaf6 100644
--- a/.azure-pipelines/mac_build_java.yml
+++ b/.azure-pipelines/mac_build_java.yml
@@ -23,7 +23,7 @@ steps:
         libcxx=$(compiler_version) cmake=3.26 \
         libboost=$(boost_version) \
         libboost-devel=$(boost_version) \
-        cairo eigen swig=4.2
+        cairo eigen swig=4.2 make
     conda activate rdkit_build
   displayName: Setup build environment
 - bash: |
diff --git a/External/pubchem_shape/PubChemShape.hpp b/External/pubchem_shape/PubChemShape.hpp
index a55064f9d..19090167b 100644
--- a/External/pubchem_shape/PubChemShape.hpp
+++ b/External/pubchem_shape/PubChemShape.hpp
@@ -1,12 +1,62 @@
+#ifndef RDKIT_PUBCHEMSHAPE_GUARD
+#define RDKIT_PUBCHEMSHAPE_GUARD
+
 #include <GraphMol/ROMol.h>
 #include <map>
 #include <vector>
 
-#ifndef RDKIT_PUBCHEMSHAPE_GUARD
-#define RDKIT_PUBCHEMSHAPE_GUARD
+#ifdef RDK_USE_BOOST_SERIALIZATION
+#include <RDGeneral/BoostStartInclude.h>
+#include <boost/archive/text_oarchive.hpp>
+#include <boost/archive/text_iarchive.hpp>
+#include <boost/serialization/map.hpp>
+#include <boost/serialization/vector.hpp>
+#include <RDGeneral/BoostEndInclude.h>
+#endif
 
 //! The input for the pubchem shape alignment code
 struct RDKIT_PUBCHEMSHAPE_EXPORT ShapeInput {
+  ShapeInput() = default;
+  ShapeInput(const std::string &str) {
+#ifndef RDK_USE_BOOST_SERIALIZATION
+    PRECONDITION(0, "Boost SERIALIZATION is not enabled")
+#else
+    std::stringstream ss(str);
+    boost::archive::text_iarchive ia(ss);
+    ia &*this;
+#endif
+  }
+  ShapeInput(const ShapeInput &other) = default;
+  ShapeInput(ShapeInput &&other) = default;
+  ShapeInput &operator=(const ShapeInput &other) = default;
+  ShapeInput &operator=(ShapeInput &&other) = default;
+  ~ShapeInput() = default;
+
+  std::string toString() const {
+#ifndef RDK_USE_BOOST_SERIALIZATION
+    PRECONDITION(0, "Boost SERIALIZATION is not enabled")
+#else
+    std::stringstream ss;
+    boost::archive::text_oarchive oa(ss);
+    oa &*this;
+    return ss.str();
+#endif
+  }
+
+#ifdef RDK_USE_BOOST_SERIALIZATION
+  template <class Archive>
+  void serialize(Archive &ar, const unsigned int) {
+    ar & coord;
+    ar & alpha_vector;
+    ar & atom_type_vector;
+    ar & volumeAtomIndexVector;
+    ar & colorAtomType2IndexVectorMap;
+    ar & shift;
+    ar & sov;
+    ar & sof;
+  }
+#endif
+
   std::vector<float> coord;
   std::vector<double> alpha_vector;
   std::vector<unsigned int> atom_type_vector;
@@ -70,4 +120,4 @@ RDKIT_PUBCHEMSHAPE_EXPORT std::pair<double, double> AlignMolecule(
     double opt_param = 1.0, unsigned int max_preiters = 10u,
     unsigned int max_postiters = 30u);
 
-#endif
\ No newline at end of file
+#endif
diff --git a/External/pubchem_shape/test.cpp b/External/pubchem_shape/test.cpp
index 737b09415..3d3193082 100644
--- a/External/pubchem_shape/test.cpp
+++ b/External/pubchem_shape/test.cpp
@@ -282,3 +282,24 @@ TEST_CASE("Github #8096") {
     CHECK_THAT(nbr_ct, Catch::Matchers::WithinAbs(1.0, 0.005));
   }
 }
+
+#ifdef RDK_USE_BOOST_SERIALIZATION
+TEST_CASE("Serialization") {
+  auto m1 =
+      "[H]c1c([H])c([H])c([H])c([H])c1[H] |(-2.06264,-0.844763,-0.0261403;-1.04035,-0.481453,-0.0114878;-0.00743655,-1.41861,-0.0137121;-0.215455,-2.47997,-0.0295909;1.29853,-0.949412,0.00507497;2.12524,-1.65277,0.00390664;1.58501,0.395878,0.0254188;2.61997,0.704365,0.0394811;0.550242,1.31385,0.0273741;0.783172,2.37039,0.0434262;-0.763786,0.88847,0.00908113;-1.60557,1.58532,0.0100194)|"_smiles;
+  REQUIRE(m1);
+  auto shape = PrepareConformer(*m1);
+  auto istr = shape.toString();
+
+  ShapeInput shape2(istr);
+  CHECK(shape2.coord == shape.coord);
+  CHECK(shape2.alpha_vector == shape.alpha_vector);
+  CHECK(shape2.atom_type_vector == shape.atom_type_vector);
+  CHECK(shape2.volumeAtomIndexVector == shape.volumeAtomIndexVector);
+  CHECK(shape2.colorAtomType2IndexVectorMap ==
+        shape.colorAtomType2IndexVectorMap);
+  CHECK(shape2.shift == shape.shift);
+  CHECK_THAT(shape2.sov, Catch::Matchers::WithinAbs(253.764, 0.005));
+  CHECK_THAT(shape2.sof, Catch::Matchers::WithinAbs(5.074, 0.005));
+}
+#endif
\ No newline at end of file