feat: return more info about the fragmentation fallbacks

- log if we find a >0d node 
- implement interface for branching site structure

.bumpversion.cfg

@@ -1,20 +1,20 @@
 [bumpversion]
-current_version = 0.0.5-dev
+current_version = 0.0.6-dev
 commit = True
 tag = False
 parse = (?P<major>\d+)\.(?P<minor>\d+)\.(?P<patch>\d+)(?:-(?P<release>[0-9A-Za-z-]+(?:\.[0-9A-Za-z-]+)*))?(?:\+(?P<build>[0-9A-Za-z-]+(?:\.[0-9A-Za-z-]+)*))?
 serialize = 
-    {major}.{minor}.{patch}-{release}+{build}
-    {major}.{minor}.{patch}+{build}
-    {major}.{minor}.{patch}-{release}
-    {major}.{minor}.{patch}
+	{major}.{minor}.{patch}-{release}+{build}
+	{major}.{minor}.{patch}+{build}
+	{major}.{minor}.{patch}-{release}
+	{major}.{minor}.{patch}
 
 [bumpversion:part:release]
 optional_value = production
 first_value = dev
 values = 
-    dev
-    production
+	dev
+	production
 
 [bumpverion:part:build]
 values = [0-9A-Za-z-]+

docs/source/conf.py

@@ -24,7 +24,7 @@
 author = "Kevin Maik Jablonka"
 
 # The full version, including alpha/beta/rc tags
-release = "0.0.5-dev"
+release = "0.0.6-dev"
 
 
 # -- General configuration ---------------------------------------------------

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = moffragmentor
-version = 0.0.5-dev
+version = 0.0.6-dev
 description = Splits MOFs into metal nodes and linkers.
 author = Kevin Maik Jablonka
 author_email = mail@kjablonka.com

src/moffragmentor/fragmentor/__init__.py

@@ -49,7 +49,7 @@ def run_fragmentation(
         try:
             not_node = []
             logger.debug(f"Fragmenting MOF for the {counter} time")
-            node_result, node_collection = find_nodes(
+            node_result, node_collection, has_1d_sbu = find_nodes(
                 mof,
                 unbound_solvent,
                 forbidden_indices,

src/moffragmentor/fragmentor/nodelocator.py

@@ -215,10 +215,12 @@ def create_single_metal_nodes(mof, node_result):
 def break_rod_nodes(mof, node_result):
     """Break rod nodes into smaller pieces."""
     new_nodes = []
+    broke_nodes = False
     for node in node_result.nodes:
         if get_sbu_dimensionality(mof, node) >= 1:
             logger.debug("Found 1- or 2-dimensional node. Will break into isolated metals.")
             new_nodes.extend(break_rod_node(mof, node))
+            broke_nodes = True
         else:
             new_nodes.append(node)
     new_node_result = NodelocationResult(
@@ -228,7 +230,7 @@ def break_rod_nodes(mof, node_result):
         node_result.binding_indices,
         node_result.to_terminal_from_branching,
     )
-    return new_node_result
+    return new_node_result, broke_nodes
 
 
 def check_node(node_indices, branching_indices, mof) -> bool:
@@ -300,6 +302,7 @@ def find_nodes(
     Returns:
         NodeCollection: collection of nodes
     """
+    broke_rod_nodes = None
     if forbidden_indices is None:
         forbidden_indices = []
 
@@ -311,14 +314,14 @@ def find_nodes(
         node_result = create_single_metal_nodes(mof, node_result)
     if check_dimensionality:
         # If we have nodes with dimensionality >0, we change these nodes to only contain the metal
-        node_result = break_rod_nodes(mof, node_result)
+        node_result, broke_rod_nodes = break_rod_nodes(mof, node_result)
     if break_organic_nodes_:
         # ToDo: This, of course, would also break prophyrin ...
         node_result = break_organic_nodes(node_result, mof)
 
     node_collection = create_node_collection(mof, node_result)
 
-    return node_result, node_collection
+    return node_result, node_collection, broke_rod_nodes
 
 
 def metal_and_branching_coplanar(node_idx, all_branching_idx, mof, tol=0.1):

src/moffragmentor/net.py

@@ -621,8 +621,9 @@ def build_net(
             edge_selection.append(edges[0][0])
         else:
             # sort ascending by second element in tuple in the list
-            edge_selection.append(sorted(edges, key=lambda x: x[1])[0][0])
-
+            selected_edges = [e[0] for e in sorted(edges, key=lambda x: x[1])]
+            edge_selection.extend(selected_edges)  # sorted(edges, key=lambda x: x[1])[0][0])
+            # edge_selection.extend(sorted(edges, key=lambda x: x[1])[0][0])
     for edge in edge_selection:
         if not contains_edge(edge, found_edges):
             found_edges.append(edge)

src/moffragmentor/sbu/linker.py

@@ -2,7 +2,7 @@
 """Describing the organic building blocks, i.e., linkers."""
 from copy import deepcopy
 
-from pymatgen.core import Molecule
+from pymatgen.core import Molecule, PeriodicSite, Structure
 from rdkit.Chem import AllChem
 
 from .sbu import SBU
@@ -15,6 +15,21 @@ def __repr__(self) -> str:
         """Return a string representation of the linker."""
         return f"Linker ({self.composition})"
 
+    def _get_branching_sites_structure(self) -> Structure:
+        new_sites = []
+        s = self._get_boxed_structure()
+        for i, site in enumerate(s):
+            if set([self.mapping_to_original_indices[i]]) & self.original_graph_branching_indices:
+                if "original_species" in site.properties and site.properties["original_species"]:
+                    species = site.properties["original_species"]
+                else:
+                    species = site.species
+
+                new_site = PeriodicSite(species, site.coords, site.lattice)
+                new_sites.append(new_site)
+
+        return Structure.from_sites(new_sites)
+
 
 def _get_edge_dict_from_rdkit_mol(mol):
     edges = {}

src/moffragmentor/sbu/sbu.py

@@ -10,7 +10,7 @@
 from backports.cached_property import cached_property
 from loguru import logger
 from pymatgen.analysis.graphs import MoleculeGraph
-from pymatgen.core import Molecule, Structure
+from pymatgen.core import Molecule, PeriodicSite, Structure
 from pymatgen.io.babel import BabelMolAdaptor
 from rdkit import Chem
 from scipy.spatial.distance import pdist
@@ -378,23 +378,40 @@ def smiles(self) -> str:
         except Exception:
             return smiles
 
-    def _get_boxed_structure(self):
-        max_size = _get_max_sep(self.molecule.cart_coords)
-        structure = self.molecule.get_boxed_structure(
+    def _get_boxed_structure(self, mol=None) -> Structure:
+        if mol is None:
+            mol = self.molecule
+        max_size = _get_max_sep(mol.cart_coords)
+        structure = mol.get_boxed_structure(
             max_size + 0.1 * max_size,
             max_size + 0.1 * max_size,
             max_size + 0.1 * max_size,
             reorder=False,
         )
         return structure
 
-    def _get_binding_sites_structure(self):
+    def _get_binding_sites_structure(self) -> Structure:
         sites = []
         s = self._get_boxed_structure()
         for i in self.binding_indices:
             sites.append(s[i])
         return Structure.from_sites(sites)
 
+    def _get_branching_sites_structure(self) -> Structure:
+        new_sites = []
+        s = self._get_boxed_structure(self._dummy_molecule)
+        for i, site in enumerate(s):
+            if set(self._dummy_molecule_indices_mapping[i]) & self._dummy_branching_indices:
+                if "original_species" in site.properties and site.properties["original_species"]:
+                    species = site.properties["original_species"]
+                else:
+                    species = site.species
+
+                new_site = PeriodicSite(species, site.coords, site.lattice)
+                new_sites.append(new_site)
+
+        return Structure.from_sites(new_sites)
+
     @cached_property
     def descriptors(self):
         return self._get_descriptors()

src/moffragmentor/version.py

@@ -14,7 +14,7 @@
     "get_git_hash",
 ]
 
-VERSION = "0.0.5-dev"
+VERSION = "0.0.6-dev"
 
 
 def get_git_hash() -> str: