PFASGroups 3.4.2__tar.gz → 3.4.3__tar.gz

{pfasgroups-3.4.2 → pfasgroups-3.4.3}/PFASGroups/PFASDefinitionModel.py

@@ -177,36 +177,49 @@ class PFASDefinition:
     
     def _check_fluorine_ratio(self, formula: Dict[str, int], include_hydrogen: bool) -> bool:
         """Check if the fluorine ratio in a molecular formula meets the threshold.
-        
+
+        Mass-weighted (F mass / total mass), matching the zeropmdb production
+        database's `fluorine_ratio` field (see `Update_CFHNOPS.sql`):
+        `fluorine_ratio = F_mass / (molecular_weight - H_mass)`. This is NOT an
+        atom-count ratio -- a molecule with heavy non-F/non-H atoms (e.g. Cl, Br)
+        will have a lower ratio per atom than per unit mass would suggest, and
+        vice versa for light atoms (C, N, O), so the two formulas can disagree
+        on which side of a threshold like 0.3 a given molecule falls.
+
         Parameters
         ----------
         formula : Dict[str, int]
             Molecular formula as {element: count} dictionary
         include_hydrogen : bool
-            If True, includes hydrogen atoms in total atom count.
-            If False, excludes hydrogen from total (heavy atoms only)
-        
+            If True, includes hydrogen mass in the total mass.
+            If False, excludes hydrogen mass from total (heavy atoms only) --
+            this is the "heavy atoms only" mode PFASSTRUCTv5 uses.
+
         Returns
         -------
         bool
-            True if (F_count / total_atoms) >= self.fluorineRatio, False otherwise
-        
+            True if (F_mass / total_mass) >= self.fluorineRatio, False otherwise
+
         Notes
         -----
-        - Returns False if total_atoms is 0
+        - Returns False if total_mass is 0
         - Formula with no fluorine (F_count=0) will fail unless fluorineRatio=0
         """
+        from rdkit.Chem import GetPeriodicTable
+        pt = GetPeriodicTable()
+
         f_count = formula.get('F', 0)
-        
+        f_mass = f_count * pt.GetAtomicWeight('F')
+
         if include_hydrogen:
-            total_atoms = sum(formula.values())
+            total_mass = sum(count * pt.GetAtomicWeight(sym) for sym, count in formula.items())
         else:
-            total_atoms = sum(v for k, v in formula.items() if k != 'H')
-        
-        if total_atoms == 0:
+            total_mass = sum(count * pt.GetAtomicWeight(sym) for sym, count in formula.items() if sym != 'H')
+
+        if total_mass == 0:
             return False
-        
-        ratio = f_count / total_atoms
+
+        ratio = f_mass / total_mass
         return ratio >= self.fluorineRatio
     
     def test(self, test_data=None):

{pfasgroups-3.4.2 → pfasgroups-3.4.3}/PFASGroups/data/PFAS_definitions_smarts.json

@@ -134,11 +134,13 @@
         ],
         "fluorineRatio": 0.3,
         "includeHydrogen": false,
-        "description": "Custom structural patterns or F ratio ≥ 0.3 (heavy atoms only)",
+        "description": "Custom structural patterns or F ratio >= 0.3, mass-weighted (F mass / heavy-atom mass), matching the zeropmdb production database's fluorine_ratio field. Not an atom-count ratio: molecules with heavier non-F/non-H atoms (Cl, Br, S...) score lower per unit mass than per atom, and vice versa for lighter atoms (C, N, O).",
         "test": {
             "category": "definition",
             "examples": {
                 "positives": [
+                    "C1=COC(=C1)C=C(S(=O)(=O)C(F)(F)F)S(=O)(=O)C(F)(F)F",
+                    "COC(=C(F)Cl)F",
                     "C(=O)(C(F)(F)Cl)C(F)(F)Cl",
                     "COC(C(OC(C(=O)C(F)(F)F)(F)F)F)(F)F",
                     "C(C(=O)CC(C(=O)C(F)(F)F)(F)F)C(C(=O)C(F)(F)F)(F)F",
@@ -160,9 +162,7 @@
                 ],
                 "negatives": [
                     "CCCCC(CC(C(F)(F)Br)(F)Cl)Br",
-                    "C(C(OC(Cl)(Cl)Cl)(F)F)(F)Cl",
-                    "C1=COC(=C1)C=C(S(=O)(=O)C(F)(F)F)S(=O)(=O)C(F)(F)F",
-                    "COC(=C(F)Cl)F"
+                    "C(C(OC(Cl)(Cl)Cl)(F)F)(F)Cl"
                 ]
             }
         }

{pfasgroups-3.4.2 → pfasgroups-3.4.3/PFASGroups.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: PFASGroups
-Version: 3.4.2
+Version: 3.4.3
 Summary: A comprehensive cheminformatics package for automated detection, classification, and analysis of halogenated substances, with a focus on per- and polyfluoroalkyl substances (PFAS). Combines SMARTS pattern matching, molecular formula constraints, and graph-based pathfinding to identify 119 groups. Supports creations of embeddings with graph metrics for Machine Learning workflows.
 Author: Luc T. Miaz
 Author-email: "Luc T. Miaz" <luc@miaz.ch>
@@ -846,6 +846,8 @@ See [USER_GUIDE.md](USER_GUIDE.md) for comprehensive examples including:
 - Integration with pandas and scikit-learn
 
 ## Summary of changes by version
+- **Version 3.4.3**: Fixed ratio error, specific to PFASSTRUCTv5
+
 - **Version 3.4.2**: Fixed ambiguity with OPPT 2023.
 
 - **Version 3.4.1**: Added generic embedding for non-halogenated specific compounds which includes wildcard-components functional groups and halogen component size. Fixed H components and added cyclic and aryl H components.

{pfasgroups-3.4.2/PFASGroups.egg-info → pfasgroups-3.4.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: PFASGroups
-Version: 3.4.2
+Version: 3.4.3
 Summary: A comprehensive cheminformatics package for automated detection, classification, and analysis of halogenated substances, with a focus on per- and polyfluoroalkyl substances (PFAS). Combines SMARTS pattern matching, molecular formula constraints, and graph-based pathfinding to identify 119 groups. Supports creations of embeddings with graph metrics for Machine Learning workflows.
 Author: Luc T. Miaz
 Author-email: "Luc T. Miaz" <luc@miaz.ch>
@@ -846,6 +846,8 @@ See [USER_GUIDE.md](USER_GUIDE.md) for comprehensive examples including:
 - Integration with pandas and scikit-learn
 
 ## Summary of changes by version
+- **Version 3.4.3**: Fixed ratio error, specific to PFASSTRUCTv5
+
 - **Version 3.4.2**: Fixed ambiguity with OPPT 2023.
 
 - **Version 3.4.1**: Added generic embedding for non-halogenated specific compounds which includes wildcard-components functional groups and halogen component size. Fixed H components and added cyclic and aryl H components.

{pfasgroups-3.4.2 → pfasgroups-3.4.3}/README.md

@@ -798,6 +798,8 @@ See [USER_GUIDE.md](USER_GUIDE.md) for comprehensive examples including:
 - Integration with pandas and scikit-learn
 
 ## Summary of changes by version
+- **Version 3.4.3**: Fixed ratio error, specific to PFASSTRUCTv5
+
 - **Version 3.4.2**: Fixed ambiguity with OPPT 2023.
 
 - **Version 3.4.1**: Added generic embedding for non-halogenated specific compounds which includes wildcard-components functional groups and halogen component size. Fixed H components and added cyclic and aryl H components.

{pfasgroups-3.4.2 → pfasgroups-3.4.3}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "PFASGroups"
-version = "3.4.2"
+version = "3.4.3"
 description = "A comprehensive cheminformatics package for automated detection, classification, and analysis of halogenated substances, with a focus on per- and polyfluoroalkyl substances (PFAS). Combines SMARTS pattern matching, molecular formula constraints, and graph-based pathfinding to identify 119 groups. Supports creations of embeddings with graph metrics for Machine Learning workflows."
 authors = [
     { name = "Luc T. Miaz", email = "luc@miaz.ch" }

{pfasgroups-3.4.2 → pfasgroups-3.4.3}/setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = PFASGroups
-version = 3.4.2
+version = 3.4.3
 description = A comprehensive cheminformatics package for automated detection, classification, and analysis of halogenated substances, with a focus on per- and polyfluoroalkyl substances (PFAS). Supports multi-halogen fingerprinting (F, Cl, Br, I) with per/poly saturation filtering and stackable fingerprint vectors.
 long_description = file: README.md
 long_description_content_type = text/markdown