Merge pull request #245 from JaGeo/enhance_batch_sg

Add cutoff args to batch structure graphs featurizer

lobsterpy/featurize/batch.py

@@ -573,6 +573,8 @@ class BatchStructureGraphs:
     :param path_to_lobster_calcs: path to root directory consisting of all lobster calc
     :param add_additional_data_sg: bool indicating whether to include `icoop` and `icobi` data as edge properties
     :param which_bonds: selects which kind of bonds are analyzed. "all" is the default
+    :param cutoff_icohp: only bonds that are stronger than cutoff_icohp * strongest ICOHP will be considered.
+    :param noise_cutoff: if provided hardcodes the lower limit of icohps considered.
     :param start: start energy for bonding antibonding percent integration
     :param n_jobs: parallel processes to run
 
@@ -583,6 +585,8 @@ def __init__(
         path_to_lobster_calcs: str | Path,
         add_additional_data_sg: bool = True,
         which_bonds: str = "all",
+        cutoff_icohp: float = 0.10,
+        noise_cutoff: float = 0.1,
         start: float | None = None,
         n_jobs: int = 4,
     ):
@@ -592,13 +596,17 @@ def __init__(
         :param path_to_lobster_calcs: path to root directory consisting of all lobster calc
         :param add_additional_data_sg: bool indicating whether to include `icoop` and `icobi` data as edge properties
         :param which_bonds: selects which kind of bonds are analyzed. "all" is the default
+        :param cutoff_icohp: only bonds that are stronger than cutoff_icohp * strongest ICOHP will be considered.
+        :param noise_cutoff: if provided hardcodes the lower limit of icohps considered.
         :param start: start energy for bonding antibonding percent integration
         :param n_jobs: parallel processes to run
 
         """
         self.path_to_lobster_calcs = path_to_lobster_calcs
         self.add_additional_data_sg = add_additional_data_sg
         self.which_bonds = which_bonds
+        self.cutff_icohp = cutoff_icohp
+        self.noise_cutoff = noise_cutoff
         self.start = start
         self.n_jobs = n_jobs
 
@@ -627,6 +635,8 @@ def _get_sg_df(self, path_to_lobster_calc: str | Path) -> pd.DataFrame:
             path_to_icobilist=str(file_paths.get("icobilist")),
             path_to_madelung=str(file_paths.get("madelung")),
             which_bonds=self.which_bonds,
+            cutoff_icohp=self.cutff_icohp,
+            noise_cutoff=self.noise_cutoff,
             start=self.start,
         )
 

lobsterpy/structuregraph/graph.py

@@ -39,6 +39,7 @@ class LobsterGraph:
     :param path_to_icobilist: path to ICOBILIST.lobster (e.g., "ICOBILIST.lobster").
     :param path_to_madelung: path to MadelungEnergies.lobster (e.g., "MadelungEnergies.lobster")
     :param cutoff_icohp: only bonds that are stronger than cutoff_icohp * strongest ICOHP will be considered.
+    :param noise_cutoff: if provided hardcodes the lower limit of icohps considered.
     :param add_additional_data_sg: if True will add the information from ICOOPLIST.lobster
         and ICOBILIST.lobster based on ICOHPLIST.lobster relevant bond.
     :param which_bonds: selects which kind of bonds are analyzed. "all" is the default.
@@ -57,6 +58,7 @@ def __init__(
         path_to_icobilist: str | Path | None = None,
         which_bonds: str = "all",
         cutoff_icohp: float = 0.10,
+        noise_cutoff: float = 0.1,
         start: float | None = None,
     ):
         """Initialize and return a structure graph object."""
@@ -79,6 +81,7 @@ def __init__(
         self.path_to_madelung = path_to_madelung
         self.which_bonds = which_bonds
         self.cutoff_icohp = cutoff_icohp
+        self.noise_cutoff = noise_cutoff
 
         if self.which_bonds == "all":
             self.additional_condition = 0
@@ -115,6 +118,7 @@ def get_decorated_sg(self):
                 id_blist_sg2="ICOOP",
                 valences_from_charges=True,
                 adapt_extremum_to_add_cond=True,
+                noise_cutoff=self.noise_cutoff,
             )
 
         else:
@@ -128,6 +132,7 @@ def get_decorated_sg(self):
                 add_additional_data_sg=self.add_additional_data_sg,
                 valences_from_charges=True,
                 adapt_extremum_to_add_cond=True,
+                noise_cutoff=self.noise_cutoff,
             )
 
         # Adds Mulliken and Löwdin charges as site properties to structure object (node properties)