Adding synthetic_complexity_score

molgenevalmetric.py

@@ -43,6 +43,7 @@ This is a standalone, importable SCScorer model. It does not have tensorflow as
 dependency and is a more attractive option for deployment. The calculations are
 fast enough that there is no real reason to use GPUs (via tf) instead of CPUs (via np)
 '''
+
 import numpy as np
 import time
 import rdkit.Chem as Chem
@@ -51,9 +52,6 @@ import json
 import gzip
 import six
 
-from types import MethodType
-
-
 import os
 project_root = os.path.dirname(os.path.dirname(__file__))
 
@@ -67,40 +65,36 @@ def sigmoid(x):
   return 1 / (1 + np.exp(-x))
 
 class SCScorer():
-    self.mol_to_fp = MethodType(mol_to_fp, self)
-
     def __init__(self, score_scale=score_scale):
         self.vars = []
         self.score_scale = score_scale
         self._restored = False
 
     def restore(self, weight_path=os.path.join('model.ckpt-10654.as_numpy.json.gz'), FP_rad=FP_rad, FP_len=FP_len):
-        self.FP_len = FP_len
-        self.FP_rad = FP_rad
+        self.FP_len = FP_len; self.FP_rad = FP_rad
         self._load_vars(weight_path)
+        # print('Restored variables from {}'.format(weight_path))
 
         if 'uint8' in weight_path or 'counts' in weight_path:
-            def mol_to_fp(mol):
+            def mol_to_fp(self, mol):
                 if mol is None:
                     return np.array((self.FP_len,), dtype=np.uint8)
-                fp = AllChem.GetMorganFingerprint(mol, self.FP_rad, useChirality=True)  # uint sparse vector
+                fp = AllChem.GetMorganFingerprint(mol, self.FP_rad, useChirality=True) # uitnsparsevect
                 fp_folded = np.zeros((self.FP_len,), dtype=np.uint8)
                 for k, v in six.iteritems(fp.GetNonzeroElements()):
                     fp_folded[k % self.FP_len] += v
                 return np.array(fp_folded)
         else:
-            def mol_to_fp(mol):
+            def mol_to_fp(self, mol):
                 if mol is None:
                     return np.zeros((self.FP_len,), dtype=np.float32)
-                return np.array(AllChem.GetMorganFingerprintAsBitVect(mol, self.FP_rad, nBits=self.FP_len, useChirality=True), dtype=np.bool_)
-
-        # Set the mol_to_fp method for the instance
-        self.mol_to_fp = MethodType(mol_to_fp, self)
+                return np.array(AllChem.GetMorganFingerprintAsBitVect(mol, self.FP_rad, nBits=self.FP_len,
+                    useChirality=True), dtype=np.bool_)
+        self.mol_to_fp = mol_to_fp
 
         self._restored = True
         return self
 
-
     def smi_to_fp(self, smi):
         if not smi:
             return np.zeros((self.FP_len,), dtype=np.float32)