Debugging and fixing production

HiPRGen/mol_entry.py

@@ -46,9 +46,12 @@ def find_fragment_atom_mappings(fragment_1, fragment_2, return_one=True):
             hot_f2_indices.append(fragment_2.hot_atoms[hot_nbh_hash])
         hot_f1_indices.append(fragment_1.hot_atoms[hot_nbh_hash])
 
+    #print("hot_f1_indices:", hot_f1_indices)
+    #print("hot_f2_indices:", hot_f2_indices)
 
     for left_index in fragment_1.compressed_graph.nodes():
-
+        #print("left_index:", left_index)
+
         neighborhood_hash = fragment_1.neighborhood_hashes[left_index]
         if neighborhood_hash in match_hot_atoms and left_index in hot_f1_indices:
             neighborhood_hash = neighborhood_hash + "hot"
@@ -57,27 +60,40 @@ def find_fragment_atom_mappings(fragment_1, fragment_2, return_one=True):
 
         groups_by_hash[neighborhood_hash][0].append(left_index)
 
+    #print("groups_by_hash:", groups_by_hash)
 
     for right_index in fragment_2.compressed_graph.nodes():
+        #print("right_index:", right_index)
 
         neighborhood_hash = fragment_2.neighborhood_hashes[right_index]
         if neighborhood_hash in match_hot_atoms and right_index in hot_f2_indices:
             neighborhood_hash = neighborhood_hash + "hot"
         if neighborhood_hash not in groups_by_hash:
+            #print("NEW HASH HOW IS THIS POSSIBLE?!!!??!?!")
+            #print(huh)
             groups_by_hash[neighborhood_hash] = ([],[])
 
         groups_by_hash[neighborhood_hash][1].append(right_index)
 
+    #print("groups_by_hash:", groups_by_hash)
+
     groups = list(groups_by_hash.values())
+    #print("groups:", groups)
 
     product_sym_iterator = product(*[sym_iterator(len(p[0])) for p in groups])
 
+    #print("product_sym_iterator:", product_sym_iterator)
     mappings = []
 
     for product_perm in product_sym_iterator:
+        #print("product_perm:", product_perm)
         mapping = {}
         for perm, vals in zip(product_perm, groups):
+            #print("perm:", perm)
+            #print("vals:", vals)
             for i, j in enumerate(perm):
+                #print("i:", i)
+                #print("j:", j)
                 mapping[vals[0][i]] = vals[1][j]
 
         isomorphism = True

HiPRGen/reaction_questions.py

@@ -60,20 +60,20 @@ class Terminal(Enum): KEEP = 1 DISCARD = -1
 
 hydrogen_graph = nx.MultiGraph()
 hydrogen_graph.add_node(0, specie="H")
-hydrogen_hash = weisfeiler_lehman_graph_hash(hydrogen_graph, node_attr="specie")
+hydrogen_hash = weisfeiler_lehman_graph_hash(hydrogen_graph, node_attr="specie", iterations=4)
 
 fluorine_graph = nx.MultiGraph()
 fluorine_graph.add_node(0, specie="F")
-fluorine_hash = weisfeiler_lehman_graph_hash(fluorine_graph, node_attr="specie")
+fluorine_hash = weisfeiler_lehman_graph_hash(fluorine_graph, node_attr="specie", iterations=4)
 
 carbon_graph = nx.MultiGraph()
 carbon_graph.add_node(0, specie="C")
-carbon_hash = weisfeiler_lehman_graph_hash(carbon_graph, node_attr="specie")
+carbon_hash = weisfeiler_lehman_graph_hash(carbon_graph, node_attr="specie", iterations=4)
 
 co3_mol = Molecule.from_file("xyz_files/co3.xyz")
 co3_mg = MoleculeGraph.with_local_env_strategy(co3_mol, OpenBabelNN())
 co3_g = co3_mg.graph.to_undirected()
-co3_hash = weisfeiler_lehman_graph_hash(co3_g, node_attr="specie")
+co3_hash = weisfeiler_lehman_graph_hash(co3_g, node_attr="specie", iterations=4)
 
 
 def run_decision_tree(
@@ -1939,27 +1939,27 @@ def __call__(self, reaction, mol_entries, params):
     (dG_above_threshold(0.0, "free_energy", 0.0), Terminal.DISCARD),
     (reactants_are_both_anions_or_both_cations(), Terminal.DISCARD),
     (reaction_is_charge_transfer(), Terminal.KEEP),
-    # (reaction_is_covalent_charge_decomposable(), Terminal.DISCARD),
-    # (reaction_is_coupled_electron_fragment_transfer(), Terminal.DISCARD),
-    # (star_count_diff_above_threshold(6), Terminal.DISCARD),
-    # (
-    #     fragment_matching_found(),
-    #     [
-    #         (single_reactant_single_product_not_atom_transfer(), Terminal.DISCARD),
-    #         (single_reactant_double_product_ring_close(), Terminal.DISCARD),
-    #         (reaction_is_hindered(), Terminal.DISCARD),
-    #         (
-    #             reaction_is_covalent_decomposable(),
-    #             [
-    #                 (fragments_are_not_2A_B(), Terminal.DISCARD),
-    #                 (mapping_with_reaction_center_not_found(), Terminal.DISCARD),
-    #                 (reaction_default_true(), Terminal.KEEP),
-    #             ],
-    #         ),
-    #         (mapping_with_reaction_center_not_found(), Terminal.DISCARD),
-    #         (reaction_default_true(), Terminal.KEEP),
-    #     ],
-    # ),
+    (reaction_is_covalent_charge_decomposable(), Terminal.DISCARD),
+    (reaction_is_coupled_electron_fragment_transfer(), Terminal.DISCARD),
+    (star_count_diff_above_threshold(6), Terminal.DISCARD),
+    (
+        fragment_matching_found(),
+        [
+            (single_reactant_single_product_not_atom_transfer(), Terminal.DISCARD),
+            (single_reactant_double_product_ring_close(), Terminal.DISCARD),
+            (reaction_is_hindered(), Terminal.DISCARD),
+            (
+                reaction_is_covalent_decomposable(),
+                [
+                    (fragments_are_not_2A_B(), Terminal.DISCARD),
+                    (mapping_with_reaction_center_not_found(), Terminal.DISCARD),
+                    (reaction_default_true(), Terminal.KEEP),
+                ],
+            ),
+            (mapping_with_reaction_center_not_found(), Terminal.DISCARD),
+            (reaction_default_true(), Terminal.KEEP),
+        ],
+    ),
     (reaction_default_true(), Terminal.DISCARD),
 ]
 
@@ -1990,4 +1990,4 @@ def __call__(self, reaction, mol_entries, params):
         ],
     ),
     (reaction_default_true(), Terminal.DISCARD),
-]
+]

HiPRGen/species_filter.py

@@ -213,14 +213,28 @@ def collapse_isomorphism_group(g):
     log_message("mapping fragments")
     fragment_dict = {}
     for mol in mol_entries:
-        # print(mol.entry_id)
+        log_message("")
+        log_message(mol.entry_id)
+        f = open("mapping_record", "a")
+        f.write("Mapping " + mol.entry_id + "\n")
+        f.close()
         for fragment_complex in mol.fragment_data:
+            log_message("new fragment complex")
+            log_message("bonds broken:" + str(fragment_complex.bonds_broken))
             for ii, fragment in enumerate(fragment_complex.fragment_objects):
+                log_message(str(ii))
                 hot_nbh_hashes = list(fragment.hot_atoms.keys())
+                log_message("hot_nbh_hashes:" + str(hot_nbh_hashes))
                 assert len(hot_nbh_hashes) == 0 or len(hot_nbh_hashes) == 1 or len(hot_nbh_hashes) == 2
                 assert fragment.fragment_hash == fragment_complex.fragment_hashes[ii]
                 if fragment.fragment_hash not in fragment_dict:
+                    log_message("Adding new fragment! Hash:" + fragment.fragment_hash)
                     fragment_dict[fragment.fragment_hash] = copy.deepcopy(fragment)
+                else:
+                    log_message("Fragment hash already in dict")
+                log_message("fragment hash:" + fragment.fragment_hash)
+                #print("fragment:", fragment)
+                #print("fragment_dict[fragment.fragment_hash]):", fragment_dict[fragment.fragment_hash])
                 all_mappings = find_fragment_atom_mappings(
                     fragment,
                     fragment_dict[fragment.fragment_hash])
@@ -407,4 +421,4 @@ def add_electron_species(
 def clean(input):
     return "".join([i for i in input if not i.isdigit()])
 def clean_op(input):
-    return "".join([i for i in input if i.isdigit()])
+    return "".join([i for i in input if i.isdigit()])

HiPRGen/species_questions.py

@@ -123,7 +123,7 @@ def __call__(self, mol):
                 )                                             #covalently bonded to
 
                 mol.star_hashes[i] = weisfeiler_lehman_graph_hash( #star_hashes is a dictionary, and this adds an entry to it
-                    neighborhood, node_attr="specie"               #with the atom index, i, as the key and a graph_hash (string)
+                    neighborhood, node_attr="specie", iterations=6               #with the atom index, i, as the key and a graph_hash (string)
                 )                                                  #as the value
 
         return False
@@ -151,7 +151,9 @@ def __call__(self, mol):
 
                     neighborhood_hash = weisfeiler_lehman_graph_hash(
                         neighborhood,
-                        node_attr='specie')
+                        node_attr='specie',
+                        iterations=6,
+                    )
 
                     hash_list.append(neighborhood_hash)
 
@@ -196,7 +198,7 @@ def __call__(self, mol):
                 subgraph = h.subgraph(c) #generates a subgraph from one set of nodes (this is a fragment graph)
 
                 fragment_hash = weisfeiler_lehman_graph_hash( #saves the hash of this graph
-                    subgraph, node_attr="specie"
+                    subgraph, node_attr="specie", iterations=6
                 )
 
                 tmp["c"] = copy.deepcopy(c)
@@ -237,7 +239,9 @@ def __call__(self, mol):
 
                                     neighborhood_hash = weisfeiler_lehman_graph_hash(
                                         neighborhood,
-                                        node_attr='specie')
+                                        node_attr='specie',
+                                        iterations=6,
+                                    )
 
                                     hash_list.append(neighborhood_hash)
 
@@ -342,7 +346,7 @@ def __call__(self, mol):
                             subgraph = h.subgraph(c)
 
                             fragment_hash = weisfeiler_lehman_graph_hash(
-                                subgraph, node_attr="specie"
+                                subgraph, node_attr="specie", iterations=6
                             )
 
                             fragments.append(fragment_hash)
@@ -389,6 +393,14 @@ def __call__(self, mol):
         return mol.formula == "H1 O1" and mol.charge == 1
 
 
+class formula_filter(MSONable):
+    def __init__(self, formula):
+        self.formula = formula
+
+    def __call__(self, mol):
+        return mol.formula == self.formula
+
+
 class fix_hydrogen_bonding(MSONable):
     def __init__(self):
         pass
@@ -514,10 +526,10 @@ def __call__(self, mol):
 
 
 def compute_graph_hashes(mol):
-    mol.total_hash = weisfeiler_lehman_graph_hash(mol.graph, node_attr="specie")
+    mol.total_hash = weisfeiler_lehman_graph_hash(mol.graph, node_attr="specie", iterations=6)
 
     mol.covalent_hash = weisfeiler_lehman_graph_hash(
-        mol.covalent_graph, node_attr="specie"
+        mol.covalent_graph, node_attr="specie", iterations=6
     )
 
     return False
@@ -608,6 +620,7 @@ def __call__(self, mol):
     (fix_hydrogen_bonding(), Terminal.KEEP),
     (h_atom_filter(), Terminal.DISCARD),
     (oh_plus_filter(), Terminal.DISCARD),
+    (formula_filter("C36 H28 S2"), Terminal.DISCARD),
     (compute_graph_hashes, Terminal.KEEP),
     (add_star_hashes(), Terminal.KEEP),
     (add_unbroken_fragment(neighborhood_width=width), Terminal.KEEP),

test.py

@@ -710,8 +710,23 @@ def euvl_phase2_test():
     folder = "./scratch/euvl_phase2_test"
     subprocess.run(["mkdir", folder])
 
+    bondnet_test_json = "./scratch/euvl_phase2_test/reaction_networks_graphs"
+    lmdbs_path_mol    = "./scratch/euvl_phase2_test/lmdbs/mol"
+    lmdbs_path_reaction = "./scratch/euvl_phase2_test/lmdbs/reaction"
+    subprocess.run(["mkdir", bondnet_test_json])
+    subprocess.run(["mkdir", "-p",lmdbs_path_mol])
+    subprocess.run(["mkdir", "-p",lmdbs_path_reaction])
+
     mol_json = "./data/euvl_test_set.json"
+    #mol_json = "./data/problem_entries.json"
     database_entries = loadfn(mol_json)
+    #database_entries = [database_entries[2], database_entries[4]]
+    #database_entries = [database_entries[3], database_entries[4]]
+
+    problem_entries = loadfn("./data/big_entries.json")
+    for entry in problem_entries:
+        database_entries.append(entry)
+
 
     species_decision_tree = euvl_species_decision_tree
 
@@ -720,14 +735,17 @@ def euvl_phase2_test():
         "electron_free_energy": 0.0,
     }
 
-    mol_entries = species_filter(
+    mol_entries, dgl_molecules_dict = species_filter(
         database_entries,
         mol_entries_pickle_location=folder + "/mol_entries.pickle",
+        dgl_mol_grphs_pickle_location = folder + "/dgl_mol_graphs.pickle",
+        grapher_features_pickle_location= folder + "/grapher_features.pickle",
         species_report=folder + "/unfiltered_species_report.tex",
         species_decision_tree=species_decision_tree,
         coordimer_weight=lambda mol: (mol.get_free_energy(params["temperature"])),
         species_logging_decision_tree=species_decision_tree,
         generate_unfiltered_mol_pictures=False,
+        mol_lmdb_path = folder + "/lmdbs/mol/mol.lmdb",
     )
 
     print(len(mol_entries), "initial mol entries")
@@ -740,6 +758,7 @@ def euvl_phase2_test():
         folder + "/buckets.sqlite",
         folder + "/rn.sqlite",
         folder + "/reaction_report.tex",
+        bondnet_test_json + "/test.json"
     )
 
     worker_payload = WorkerPayload(
@@ -763,6 +782,9 @@ def euvl_phase2_test():
             folder + "/mol_entries.pickle",
             folder + "/dispatcher_payload.json",
             folder + "/worker_payload.json",
+            folder + "/dgl_mol_graphs.pickle",
+            folder + "/grapher_features.pickle",
+            folder + "/lmdbs/reaction/reaction.lmdb"
         ]
     )
 
@@ -869,7 +891,7 @@ def euvl_phase2_test():
         tests_passed = False
 
     print("Number of reactions:", network_loader.number_of_reactions)
-    if network_loader.number_of_reactions == 3912:
+    if network_loader.number_of_reactions == 3905:
         print(bcolors.PASS + "euvl_phase_2_test: correct number of reactions" + bcolors.ENDC)
     else:
         print(bcolors.FAIL + "euvl_phase_2_test: correct number of reactions" + bcolors.ENDC)
@@ -1007,8 +1029,8 @@ def euvl_bondnet_test():
     # flicho_test,
     # co2_test,
     # euvl_phase1_test,
-    # euvl_phase2_test,
-    euvl_bondnet_test
+    euvl_phase2_test,
+    # euvl_bondnet_test
 ]
 
 for test in tests: