Merge pull request #255 from JacksonBurns/minhashfp-add

add the minhash fingerprint

.gitignore

@@ -8,7 +8,7 @@
 *.pyc
 
 # build and distribution files
-AIMSim.egg-info/*
+aimsim.egg-info/*
 build/*
 dist/*
 
@@ -27,4 +27,4 @@ htmlcov/*
 .ipynb_checkpoints/*
 
 # Temp files from testing
-temp_*
+temp_*

README.md

@@ -73,9 +73,10 @@ Start `AIMSim` with a prepared configuration YAML file (`config.yaml`):
 
 _The following are available via command line use (config.yaml) only:_
 
-4. All fingerprints available from the [ccbmlib](https://github.com/vogt-m/ccbmlib) package (_specify 'ccbmlib:descriptorname' for command line input_).
-5. All descriptors and fingerprints available from [PaDELPy](https://github.com/ecrl/padelpy), an interface to PaDEL-Descriptor. (_specify 'padelpy:desciptorname' for command line input._).
-6. All descriptors available through the [Mordred](https://github.com/mordred-descriptor/mordred) library (_specify 'mordred:desciptorname' for command line input._). To enable this option, you must install with `pip install 'aimsim[mordred]'` (see disclaimer in the Installation section above).
+4. MinHash Fingerprint (see [MHFP](https://github.com/reymond-group/mhfp))
+5. All fingerprints available from the [ccbmlib](https://github.com/vogt-m/ccbmlib) package (_specify 'ccbmlib:descriptorname' for command line input_).
+6. All descriptors and fingerprints available from [PaDELPy](https://github.com/ecrl/padelpy), an interface to PaDEL-Descriptor. (_specify 'padelpy:desciptorname' for command line input._).
+7. All descriptors available through the [Mordred](https://github.com/mordred-descriptor/mordred) library (_specify 'mordred:desciptorname' for command line input._). To enable this option, you must install with `pip install 'aimsim[mordred]'` (see disclaimer in the Installation section above).
 
 ### Currently Implemented Similarity Scores
 

aimsim/ops/descriptor.py

@@ -11,6 +11,7 @@
 from aimsim.utils.ccbmlib_fingerprints import generate_fingerprints
 from padelpy import from_smiles
 from aimsim.utils.extras import requires_mordred
+from mhfp.encoder import MHFPEncoder
 
 try:
     from mordred import Calculator, descriptors
@@ -264,6 +265,27 @@ def _set_ccbmlib_fingerprint(self, molecule_graph, descriptor, **kwargs):
         self.label_ = descriptor
         self.params_ = {}
 
+    def _set_minhash_fingerprint(self, molecule_graph, **kwargs):
+        """Set the descriptor to the minhash fingerprint.
+
+        Args:
+        molecule_graph (RDKIT object): Graph of molecule to be fingerprinted.
+
+        """
+        mhfp_encoder = MHFPEncoder(
+            n_permutations=kwargs["n_permutations"],
+            seed=kwargs["seed"],
+        )
+        fp = mhfp_encoder.encode_mol(
+            molecule_graph,
+            radius=kwargs["radius"],
+            rings=kwargs["rings"],
+            kekulize=kwargs["kekulize"],
+        )
+        self.numpy_ = fp
+        self.label_ = "minhash_fingerprint"
+        self.params_ = kwargs
+
     def make_fingerprint(
         self, molecule_graph, fingerprint_type, fingerprint_params=None
     ):
@@ -285,6 +307,18 @@ def make_fingerprint(
             morgan_params = {"radius": 3, "n_bits": 1024}
             morgan_params.update(fingerprint_params)
             self._set_morgan_fingerprint(molecule_graph=molecule_graph, **morgan_params)
+        elif fingerprint_type == "minhash_fingerprint":
+            minhash_params = {
+                "n_permutations": 2048,
+                "seed": 42,
+                "radius": 3,
+                "rings": True,
+                "kekulize": True,
+            }
+            minhash_params.update(fingerprint_params)
+            self._set_minhash_fingerprint(
+                molecule_graph=molecule_graph, **minhash_params
+            )
         elif fingerprint_type == "topological_fingerprint":
             topological_params = {"min_path": 1, "max_path": 7}
             topological_params.update(fingerprint_params)
@@ -480,6 +514,7 @@ def get_all_supported_descriptors():
             "morgan_fingerprint",
             "topological_fingerprint",
             "daylight_fingerprint",
+            "minhash_fingerprint",
             "maccs_keys",
             "atom-pair_fingerprint",
             "torsion_fingerprint",

requirements.txt

@@ -14,3 +14,4 @@ padelpy
 plotly
 customtkinter<5.0.0
 tkinter-tooltip
+mhfp

tests/test_Descriptor.py

@@ -273,6 +273,31 @@ def test_padelpy_descriptors(self):
             with self.assertRaises(ValueError):
                 descriptor.to_rdkit()
 
+    def test_minhash_fingerprint(self):
+        """Test creation of minhash fingerprint"""
+        mol_graph = MolFromSmiles("CCOCC")
+        descriptor = Descriptor()
+        descriptor.make_fingerprint(
+            molecule_graph=mol_graph, fingerprint_type="minhash_fingerprint"
+        )
+        self.assertTrue(
+            descriptor.check_init(),
+            "Expected Descriptor object to be initialized",
+        )
+        self.assertEqual(
+            descriptor.label_,
+            "minhash_fingerprint",
+            "Expected label of descriptor initialized with "
+            "{} to match the fingerprint".format("minhash_fingerprint"),
+        )
+        self.assertIsInstance(
+            descriptor.to_numpy(),
+            np.ndarray,
+            "Expected numpy.ndarray from to_numpy()",
+        )
+        with self.assertRaises(ValueError):
+            descriptor.to_rdkit()
+
     def test_ccbmlib_descriptors(self):
         """Test ability to passthrough descriptors to ccbmlib."""
         mol_graph = MolFromSmiles("CCOCC")
@@ -310,6 +335,7 @@ def test_exptl_descriptors(self):
             "maccs_keys",
             "atom-pair_fingerprint",
             "torsion_fingerprint",
+            "minhash_fingerprint",
         ]
         for desc in fprint_list:
             descriptor = Descriptor()