Merge branch 'feature/provenance' into develop

This adds basic support for provenance tracking to
IMP and RMF. Relates #976. Operations that alter parts
of the structure (such as reading in from a PDB file,
sampling, filtering, clustering) can now be recorded
directly in the Model itself by means of provenance
decorators, attached to atom::Hierarchy nodes. This
provenance information is also stored in RMF files.
Where possible, IMP and PMI should fill in this information
automatically.

modules/atom/include/Hierarchy.h

@@ -451,7 +451,7 @@ IMPATOMEXPORT bool get_is_heterogen(Hierarchy h);
 
 //! Clone the Hierarchy
 /** This method copies the Bond, Bonded, Atom,
-    Residue, and Domain data and the particle name to the
+    Residue, Domain, and provenance data and the particle name to the
     new copies in addition to the Hierarchy relationships.
 
     \relates Hierarchy
@@ -460,7 +460,7 @@ IMPATOMEXPORT Hierarchy create_clone(Hierarchy d);
 
 //! Clone the node in the Hierarchy
 /** This method copies the Atom,
-    Residue, Chain and Domain data and the particle name.
+    Residue, Chain, Domain, and provenance data and the particle name.
 
     \relates Hierarchy
 */
@@ -470,7 +470,8 @@ IMPATOMEXPORT Hierarchy create_clone_one(Hierarchy d);
 /** All bonds connecting to these atoms are destroyed as are
     hierarchy links in the Hierarchy and the particles are
     removed from the Model. If this particle has a parent, it is
-    removed from the parent.
+    removed from the parent. Any provenance information for this
+    Hierarchy is also removed.
     \relates Hierarchy
 */
 IMPATOMEXPORT void destroy(Hierarchy d);

modules/atom/src/Hierarchy.cpp

@@ -21,6 +21,7 @@
 #include <IMP/atom/Molecule.h>
 #include <IMP/algebra/Sphere3D.h>
 #include <IMP/atom/hierarchy_tools.h>
+#include <IMP/core/provenance.h>
 #include <IMP/algebra/geometric_alignment.h>
 #include <IMP/core/rigid_bodies.h>
 
@@ -437,6 +438,11 @@ Hierarchy clone_internal(Hierarchy d,
   if (Representation::get_is_setup(d.get_particle())) {
     nd = Representation::setup_particle(p, Representation(d.get_particle()));
   }
+  if (core::Provenanced::get_is_setup(d.get_particle())) {
+    core::Provenanced pd(d.get_particle());
+    core::Provenance prov = core::create_clone(pd.get_provenance());
+    core::Provenanced::setup_particle(p, prov);
+  }
 
   if (nd == Hierarchy()) nd = Hierarchy::setup_particle(p);
   using core::XYZ;
@@ -531,6 +537,14 @@ void destroy(Hierarchy d) {
         destroy_bond(b.get_bond(b.get_number_of_bonds() - 1));
       }
     }
+    if (core::Provenanced::get_is_setup(all[i])) {
+      core::Provenance prov = core::Provenanced(all[i]).get_provenance();
+      while (prov) {
+        core::Provenance previous = prov.get_previous();
+        prov.get_model()->remove_particle(prov.get_particle_index());
+        prov = previous;
+      }
+    }
     Hierarchy hc(all[i]);
     while (hc.get_number_of_children() > 0) {
       hc.remove_child(hc.get_child(hc.get_number_of_children() - 1));

modules/atom/src/pdb.cpp

@@ -16,6 +16,7 @@
 #include <IMP/atom/CHARMMAtom.h>
 #include <IMP/atom/CHARMMParameters.h>
 #include <IMP/atom/charmm_segment_topology.h>
+#include <IMP/core/provenance.h>
 #include <IMP/core/Hierarchy.h>
 #include <IMP/core/rigid_bodies.h>
 #include <boost/algorithm/string.hpp>
@@ -244,11 +245,18 @@ Particle* residue_particle(Model* m, const std::string& pdb_line) {
   return p;
 }
 
-Particle* chain_particle(Model* m, char chain_id) {
+Particle* chain_particle(Model* m, char chain_id, std::string filename) {
   Particle* p = new Particle(m);
   Chain::setup_particle(p, chain_id);
   p->set_name(std::string("Chain " + std::string(1, chain_id)));
   Molecule::setup_particle(p);
+
+  // Set provenance of this chain
+  core::StructureProvenance sp
+          = core::StructureProvenance::setup_particle(new Particle(m),
+                                          filename, std::string(1, chain_id));
+  core::add_provenance(m, p->get_index(), sp);
+
   return p;
 }
 }
@@ -350,7 +358,7 @@ Hierarchies read_pdb(std::istream& in, std::string name, Model* model,
         if (cp == nullptr || chain != curr_chain) {
           curr_chain = chain;
           // create new chain particle
-          cp = chain_particle(model, chain);
+          cp = chain_particle(model, chain, name);
           chain_name_set = false;
           Hierarchy(root_p).add_child(Chain(cp));
           rp = nullptr;  // make sure we get a new residue

modules/atom/test/test_pdb.py

@@ -3,6 +3,7 @@
 import IMP
 import IMP.test
 import IMP.atom
+import IMP.core
 
 
 class Tests(IMP.test.TestCase):
@@ -46,7 +47,7 @@ def test_read(self):
         #! read PDB
         mp = IMP.atom.read_pdb(self.open_input_file("input.pdb"),
                                m, IMP.atom.NonWaterPDBSelector())
-        self.assertEqual(len(m.get_particle_indexes()), 1132)
+        self.assertEqual(len(m.get_particle_indexes()), 1133)
         # IMP.atom.show_molecular_hierarchy(mp)
         IMP.atom.show(mp)
         IMP.atom.add_bonds(mp)
@@ -58,7 +59,7 @@ def test_read(self):
         m2 = IMP.Model()
         mp = IMP.atom.read_pdb(self.open_input_file("input.pdb"),
                                m2, IMP.atom.CAlphaPDBSelector())
-        self.assertEqual(len(m2.get_particle_indexes()), 260)
+        self.assertEqual(len(m2.get_particle_indexes()), 261)
         ps = IMP.atom.get_by_type(mp, IMP.atom.ATOM_TYPE)
         self.assertEqual(len(ps), 129)
         IMP.atom.add_bonds(mp)
@@ -218,5 +219,21 @@ def test_indexes(self):
         lvs = IMP.atom.get_leaves(hp)
         self.assertEqual(IMP.atom.Atom(lvs[2]).get_input_index(), 3)
 
+    def test_provenance(self):
+        """Test that StructureProvenance is set"""
+        m = IMP.Model()
+        fname = self.get_input_file_name("hydrogen.pdb")
+        mp = IMP.atom.read_pdb(fname, m)
+        chains = IMP.atom.get_by_type(mp, IMP.atom.CHAIN_TYPE)
+        self.assertEqual(len(chains), 5)
+        for c in chains:
+            self.assertTrue(IMP.core.Provenanced.get_is_setup(c))
+            p = IMP.core.Provenanced(c).get_provenance()
+            self.assertTrue(IMP.core.StructureProvenance.get_is_setup(p))
+            sp = IMP.core.StructureProvenance(p)
+            self.assertEqual(sp.get_filename(), fname)
+            self.assertEqual(sp.get_chain_id(), IMP.atom.Chain(c).get_id())
+
+
 if __name__ == '__main__':
     IMP.test.main()