update default fp functions values

chembl · Dec 26, 2024 · 16117e9 · 16117e9
1 parent d9f40bd
commit 16117e9
Show file tree

Hide file tree

Showing 4 changed files with 22 additions and 22 deletions.
diff --git a/FPSim2/io/backends/pytables.py b/FPSim2/io/backends/pytables.py
@@ -55,6 +55,11 @@ def create_db_file(
 
     if not fp_params:
         fp_params = FP_FUNC_DEFAULTS[fp_type]
+    else:
+        if "fpSize" not in fp_params:
+            if "fpSize" in FP_FUNC_DEFAULTS[fp_type]:
+                fp_params["fpSize"] = FP_FUNC_DEFAULTS[fp_type]["fpSize"]
+
     supplier = get_mol_supplier(mols_source)
     fp_length = get_fp_length(fp_type, fp_params)
     filters = tb.Filters(complib="blosc2", complevel=9, fletcher32=False)

diff --git a/FPSim2/io/chem.py b/FPSim2/io/chem.py
@@ -22,8 +22,7 @@
 }
 
 FP_FUNCS = {
-    "MACCSKeys": lambda m,
-    **kwargs: rdMolDescriptors.GetMACCSKeysFingerprint(m),
+    "MACCSKeys": lambda m, **kwargs: rdMolDescriptors.GetMACCSKeysFingerprint(m),
     "Morgan": lambda m, **kwargs: rdFingerprintGenerator.GetMorganGenerator(
         **kwargs
     ).GetFingerprint(m),
@@ -37,50 +36,46 @@
     "RDKit": lambda m, **kwargs: rdFingerprintGenerator.GetRDKitFPGenerator(
         **kwargs
     ).GetFingerprint(m),
-    "Pattern": lambda m, **kwargs: Chem.PatternFingerprint(m),
+    "RDKitPattern": lambda m, **kwargs: Chem.PatternFingerprint(m),
 }
 
 FP_FUNC_DEFAULTS = {
     "MACCSKeys": {},
     "Morgan": {
+        "radius": 3,
+        "fpSize": 2048,
+        "countSimulation": False,
         "includeChirality": False,
         "useBondTypes": True,
-        "useCountSimulation": False,
-        "countBounds": None,
-        "fpSize": 2048,
-        "radius": 2,
+        "onlyNonzeroInvariants": False,
+        "includeRingMembership": True,
+        "includeRedundantEnvironments": False,
     },
     "TopologicalTorsion": {
-        "includeChirality": False,
         "fpSize": 2048,
-        "countSimulation": False,
-        "countBounds": None,
+        "includeChirality": False,
         "torsionAtomCount": 4,
+        "countSimulation": True,
     },
     "AtomPair": {
-        "includeChirality": False,
         "fpSize": 2048,
-        "countSimulation": False,
-        "countBounds": None,
         "minDistance": 1,
         "maxDistance": 30,
+        "includeChirality": False,
+        "use2D": True,
+        "countSimulation": True,
     },
     "RDKit": {
         "fpSize": 2048,
         "minPath": 1,
         "maxPath": 7,
-        "countSimulation": False,
-        "countBounds": None,
         "useHs": True,
         "branchedPaths": True,
         "useBondOrder": True,
-    },
-    "Pattern": {
-        "fpSize": 2048,
         "countSimulation": False,
-        "countBounds": None,
-        "tautomerFingerprints": False,
+        "numBitsPerFeature": 2,
     },
+    "RDKitPattern": {"fpSize": 2048, "tautomerFingerprints": False},
 }
 
 

diff --git a/docs/user_guide/create_db_file.md b/docs/user_guide/create_db_file.md
@@ -12,7 +12,7 @@ The fingerprints are calculated with [RDKit](https://www.rdkit.org/). Fingerprin
 - [TopologicalTorsion](https://www.rdkit.org/docs/source/rdkit.Chem.rdFingerprintGenerator.html#rdkit.Chem.rdFingerprintGenerator.GetTopologicalTorsionGenerator)
 - [AtomPair](https://www.rdkit.org/docs/source/rdkit.Chem.rdFingerprintGenerator.html#rdkit.Chem.rdFingerprintGenerator.GetAtomPairGenerator)
 - [RDKit](https://www.rdkit.org/docs/source/rdkit.Chem.rdFingerprintGenerator.html#rdkit.Chem.rdFingerprintGenerator.GetRDKitFPGenerator)
-- [RDKPatternFingerprint](https://www.rdkit.org/docs/source/rdkit.Chem.rdmolops.html#rdkit.Chem.rdmolops.PatternFingerprint)
+- [RDKitPattern](https://www.rdkit.org/docs/source/rdkit.Chem.rdmolops.html#rdkit.Chem.rdmolops.PatternFingerprint)
 
 ## From a .sdf file
 ```python

diff --git a/docs/user_guide/subs_screenout.md b/docs/user_guide/subs_screenout.md
@@ -5,7 +5,7 @@
 Use the `FPSim2.FPSim2Engine.substructure` function to run an optimised Tversky (*a=1*, *b=0*, *threshold=1.0*) substructure screenout. Bear in mind that this is not a full substructure (i.e. with subgraph isomorphism) search.
 
 !!! tip
-    It's recommended to use **RDKPatternFingerprint** fingerprint type with this kind of searches.
+    It's recommended to use **RDKitPattern** fingerprint type with this kind of searches.
 
 ```python
 from FPSim2 import FPSim2Engine