Merge pull request #8 from valence-labs/evaluate

Sampling with SAFE

docs/api/safe.md

@@ -5,7 +5,7 @@
         members:
             - encode
             - decode
-            - SafeConverter
+            - SAFEConverter
         show_root_heading: false
 
 
@@ -18,4 +18,12 @@
         members:
             - SAFESplitter
             - SAFETokenizer
-        show_root_heading: false
+        show_root_heading: false
+
+
+
+---
+
+## Utils
+
+::: safe.trainer.safe_utils

docs/api/safe.models.md

@@ -53,5 +53,4 @@ The input config file for training a `SAFE` model is very similar to the GPT2 co
 ## Data Utils
 ::: safe.trainer.data_utils
 
-::: safe.trainer.safe_utils
 

expts/notebook/1.5-play-with-tokenizer.ipynb

@@ -89,7 +89,7 @@
     "from functools import partial\n",
     "ALLOWED_DESCRIPTORS = [\"mw\", \"fsp3\", \"n_lipinski_hba\", \"n_lipinski_hbd\", \"n_rings\", \"n_heavy_atoms\", \"n_hetero_atoms\", \"n_rotatable_bonds\", \"tpsa\"]\n",
     "def apply_converter(row):\n",
-    "    row[\"inputs\"] = sf.trainer.safe_utils.convert_to_safe(row[\"smiles\"], canonical=False, randomize=True, fraction_hs=0.4)\n",
+    "    row[\"inputs\"] = sf.utils.convert_to_safe(row[\"smiles\"], canonical=False, randomize=True, fraction_hs=0.4)\n",
     "    descriptors_dict = dm.descriptors.compute_many_descriptors(dm.to_mol(row[\"smiles\"]))\n",
     "    row[\"descriptors\"] = [descriptors_dict[x] for x in ALLOWED_DESCRIPTORS]\n",
     "    return row"

expts/notebook/1.6-training-toy.ipynb

@@ -166,7 +166,7 @@
     "from functools import partial\n",
     "ALLOWED_DESCRIPTORS = [\"mw\", \"fsp3\", \"n_lipinski_hba\", \"n_lipinski_hbd\", \"n_rings\", \"n_heavy_atoms\", \"n_hetero_atoms\", \"n_rotatable_bonds\", \"tpsa\"]\n",
     "def apply_converter(row):\n",
-    "    row[\"inputs\"] = sf.trainer.safe_utils.convert_to_safe(row[\"smiles\"], canonical=False, randomize=True, fraction_hs=0.4)\n",
+    "    row[\"inputs\"] = sf.utils.convert_to_safe(row[\"smiles\"], canonical=False, randomize=True, fraction_hs=0.4)\n",
     "    descriptors_dict = dm.descriptors.compute_many_descriptors(dm.to_mol(row[\"smiles\"]))\n",
     "    row[\"descriptors\"] = [descriptors_dict[x] for x in ALLOWED_DESCRIPTORS]\n",
     "    return row"

expts/notebook/tmp_data/proc_data/dataset_dict.json

@@ -0,0 +1 @@
+{"splits": ["train", "test", "validation"]}

expts/notebook/tmp_data/proc_data/test/dataset_info.json

@@ -0,0 +1,19 @@
+{
+  "citation": "",
+  "description": "",
+  "features": {
+    "inputs": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "descriptors": {
+      "feature": {
+        "dtype": "float64",
+        "_type": "Value"
+      },
+      "_type": "Sequence"
+    }
+  },
+  "homepage": "",
+  "license": ""
+}

expts/notebook/tmp_data/proc_data/test/state.json

@@ -0,0 +1,13 @@
+{
+  "_data_files": [
+    {
+      "filename": "data-00000-of-00001.arrow"
+    }
+  ],
+  "_fingerprint": "739e961c83556f0c",
+  "_format_columns": null,
+  "_format_kwargs": {},
+  "_format_type": null,
+  "_output_all_columns": false,
+  "_split": null
+}

expts/notebook/tmp_data/proc_data/train/dataset_info.json

@@ -0,0 +1,19 @@
+{
+  "citation": "",
+  "description": "",
+  "features": {
+    "inputs": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "descriptors": {
+      "feature": {
+        "dtype": "float64",
+        "_type": "Value"
+      },
+      "_type": "Sequence"
+    }
+  },
+  "homepage": "",
+  "license": ""
+}

expts/notebook/tmp_data/proc_data/train/state.json

@@ -0,0 +1,13 @@
+{
+  "_data_files": [
+    {
+      "filename": "data-00000-of-00001.arrow"
+    }
+  ],
+  "_fingerprint": "8232d33a79926636",
+  "_format_columns": null,
+  "_format_kwargs": {},
+  "_format_type": null,
+  "_output_all_columns": false,
+  "_split": null
+}

expts/notebook/tmp_data/proc_data/validation/dataset_info.json

@@ -0,0 +1,19 @@
+{
+  "citation": "",
+  "description": "",
+  "features": {
+    "inputs": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "descriptors": {
+      "feature": {
+        "dtype": "float64",
+        "_type": "Value"
+      },
+      "_type": "Sequence"
+    }
+  },
+  "homepage": "",
+  "license": ""
+}

expts/notebook/tmp_data/proc_data/validation/state.json

@@ -0,0 +1,13 @@
+{
+  "_data_files": [
+    {
+      "filename": "data-00000-of-00001.arrow"
+    }
+  ],
+  "_fingerprint": "739e961c83556f0c",
+  "_format_columns": null,
+  "_format_kwargs": {},
+  "_format_type": null,
+  "_output_all_columns": false,
+  "_split": null
+}

safe/__init__.py

@@ -2,8 +2,12 @@
 
 from .converter import encode
 from .converter import decode
+from .converter import SAFEConverter
 from .viz import to_image
-from ._exception import SafeDecodeError
-from ._exception import SafeEncodeError
-from ._exception import SafeFragmentationError
+from .tokenizer import SAFETokenizer
+from .sample import SAFEDesign
+from ._exception import SAFEDecodeError
+from ._exception import SAFEEncodeError
+from ._exception import SAFEFragmentationError
 from . import trainer
+from . import utils

safe/_exception.py

@@ -1,16 +1,16 @@
-class SafeDecodeError(Exception):
+class SAFEDecodeError(Exception):
     """Raised when a string cannot be decoded with the given encoding."""
 
     pass
 
 
-class SafeEncodeError(Exception):
+class SAFEEncodeError(Exception):
     """Raised when a molecule cannot be encoded using SAFE."""
 
     pass
 
 
-class SafeFragmentationError(Exception):
+class SAFEFragmentationError(Exception):
     """Raised when a the slicing algorithm return empty bonds."""
 
     pass