Merge pull request #175 from hjkgrp/MOF_update

clarifications to MOF code

molSimplify/Informatics/MOF/MOF_RAC_example.ipynb

@@ -365,8 +365,12 @@
     "# You don't want to set it much higher than that, since the code may start to assign bonds where none exist.\n",
     "wiggle_room = 1\n",
     "\n",
+    "# Note: if get_primitive does not work for your MOF, you can skip that step.\n",
+    "# But if your MOF does not have P1 symmetry, you will need to make it have P1 symmetry before RAC featurization with get_MOF_descriptors.\n",
+    "\t# One option for doing that is using Vesta.\n",
     "get_primitive(f'{working_path}/{cif_name}.cif', f'{working_path}/{cif_name}_primitive.cif')\n",
-    "overlap_removal(f'{working_path}/{cif_name}_primitive.cif', f'{working_path}/{cif_name}_no_overlap.cif', wiggle_room=wiggle_room)\n",
+    "\n",
+    "overlap_removal(f'{working_path}/{cif_name}_primitive.cif', f'{working_path}/{cif_name}_no_overlap.cif')\n",
     "solvent_removal(f'{working_path}/{cif_name}_no_overlap.cif', f'{working_path}/{cif_name}_no_solvent.cif', wiggle_room=wiggle_room)\n",
     "\n",
     "# Now you can featurize your \"clean\" MOF.\n",

molSimplify/Informatics/MOF/MOF_descriptors.py

@@ -851,7 +851,7 @@ def detect_1D_rod(SBU_list, molcif, allatomtypes, cell_v, logpath, name):
 def get_MOF_descriptors(data, depth, path=False, xyzpath=False, graph_provided=False, wiggle_room=1,
     max_num_atoms=2000, get_sbu_linker_bond_info=False, surrounded_sbu_file_generation=False, detect_1D_rod_sbu=False):
     """
-    Generates RAC descriptors on a MOF.
+    Generates RAC descriptors on a MOF, assuming it has P1 symmetry.
     Writes three files: sbu_descriptors.csv, linker_descriptors.csv, and lc_descriptors.csv
     These files contain the RAC descriptors of the MOF.
 
@@ -926,9 +926,9 @@ def get_MOF_descriptors(data, depth, path=False, xyzpath=False, graph_provided=F
     Getting the adjacency matrix.
     """""""""
     if not graph_provided: # Make the adjacency matrix.
-        distance_mat = compute_distance_matrix3(cell_v,cart_coords)
+        distance_mat = compute_distance_matrix3(cell_v, cart_coords)
         try:
-            adj_matrix, _ = compute_adj_matrix(distance_mat,allatomtypes,wiggle_room)
+            adj_matrix, _ = compute_adj_matrix(distance_mat, allatomtypes, wiggle_room)
         except NotImplementedError:
             failure_str = f"Failed to featurize {name}: atomic overlap\n"
             full_names, full_descriptors = failure_response(path, failure_str)

molSimplify/Informatics/MOF/PBC_functions.py

@@ -1122,7 +1122,7 @@ def remove_undesired_atoms(undesired_indices, allatomtypes, fcoords):
     return allatomtypes_trim, fcoords_trim
 
 
-def overlap_removal(cif_path, new_cif_path, wiggle_room=1):
+def overlap_removal(cif_path, new_cif_path):
     """
     Reads a cif file, removes overlapping atoms, and writes the cif to the provided path.
     For a new CIF, recommended to remove symmetry (either with Vesta or with get_primitive), then run overlap_removal, then run solvent_removal.
@@ -1133,9 +1133,6 @@ def overlap_removal(cif_path, new_cif_path, wiggle_room=1):
         The path of the cif file to be read.
     new_cif_path : str
         The path to which the modified cif file will be written.
-    wiggle_room : float
-        A multiplier that allows for more or less strict bond distance cutoffs.
-        Useful for some trouble CIFs with long bonds.
 
     Returns
     -------
@@ -1155,7 +1152,7 @@ def overlap_removal(cif_path, new_cif_path, wiggle_room=1):
 
     # Assuming that the cif does not have graph information of the structure.
     distance_mat = compute_distance_matrix3(cell_v,cart_coords)
-    adj_matrix, overlap_atoms = compute_adj_matrix(distance_mat, allatomtypes, wiggle_room=wiggle_room, handle_overlap=True)
+    adj_matrix, overlap_atoms = compute_adj_matrix(distance_mat, allatomtypes, handle_overlap=True)
 
     # Dealing with the case of overlapping atoms.
     if len(overlap_atoms) != 0: