Update README.md

README.md

@@ -81,18 +81,69 @@ and load one of the `AttentiveSkin` datasets, e.g.,
 
     >>> AttentiveSkin = datasets.load_dataset("maomlab/AttentiveSkin", name = "AttentiveSkin")
   Downloading readme:
-  4%|▍         | 4.02/100 [00:00<00:00, 6.39kkB/s]
+  4%|▍         | 4.02/100 [00:00<00:00, 39.5kkB/s]
   Downloading data:
-  100%|██████████| 468/468 [00:00<00:00, 2.97MkB/s]
+100%|██████████| 468/468 [00:00<00:00, 4.04MkB/s]
   Downloading data:
-  100%|██████████| 1.41k/1.41k [00:00<00:00, 9.94MkB/s]
+100%|██████████| 1.41k/1.41k [00:00<00:00, 10.8MkB/s]
   Generating test split:
-  100%|██████████| 803/803 [00:00<00:00, 3.75Mexamples/s]
+100%|██████████| 803/803 [00:00<00:00, 7.17Mexamples/s]
   Generating train split:
-  100%|██████████| 2.42k/2.42k [00:00<00:00, 7.37Mexamples/s]
-
-
-
+100%|██████████| 2.42k/2.42k [00:00<00:00, 6.23Mexamples/s]
+
+and inspecting the loaded dataset
+
+    >>> AttentiveSkin
+    AttentiveSkin
+
+DatasetDict({
+    test: Dataset({
+        features: ['Name', 'Synonym', 'CAS RN', 'GHS', 'Detailed Page', 'Evidence', 'OECD TG 404', 'Data Source', 'Frequency', 'SMILES', 'SMILES URL', 'SMILES Source', 'Canonical SMILES', 'Split', 'ClusterNo', 'MolCount', 'group'],
+        num_rows: 803
+    })
+    train: Dataset({
+        features: ['Name', 'Synonym', 'CAS RN', 'GHS', 'Detailed Page', 'Evidence', 'OECD TG 404', 'Data Source', 'Frequency', 'SMILES', 'SMILES URL', 'SMILES Source', 'Canonical SMILES', 'Split', 'ClusterNo', 'MolCount', 'group'],
+        num_rows: 2416
+    })
+})
+
+### Use a dataset to train a model
+One way to use the dataset is through the [MolFlux](https://exscientia.github.io/molflux/) package developed by Exscientia.
+First, from the command line, install `MolFlux` library with `catboost` and `rdkit` support
+
+    pip install 'molflux[catboost,rdkit]'
+
+then load, featurize, split, fit, and evaluate the catboost model
+
+    import json
+    from datasets import load_dataset
+    from molflux.datasets import featurise_dataset
+    from molflux.features import load_from_dicts as load_representations_from_dicts
+    from molflux.splits import load_from_dict as load_split_from_dict
+    from molflux.modelzoo import load_from_dict as load_model_from_dict
+    from molflux.metrics import load_suite
+    
+    split_dataset = load_dataset('maomlab/AttentiveSkin')
+    
+    split_featurised_dataset = featurise_dataset(
+      split_dataset,
+      column = "SMILES",
+      representations = load_representations_from_dicts([{"name": "morgan"}, {"name": "maccs_rdkit"}]))
+
+    model = load_model_from_dict({
+        "name": "cat_boost_classifier",
+        "config": {
+            "x_features": ['SMILES::morgan', 'SMILES::maccs_rdkit'],
+            "y_features": ['Solubility']}})
+    
+    model.train(split_featurised_dataset["train"])
+    preds = model.predict(split_featurised_dataset["test"])
+    
+    classification_suite = load_suite("classification")
+    
+    scores = classification_suite.compute(
+        references=split_featurised_dataset["test"]['Solubility'],
+        predictions=preds["cat_boost_classifier::Solubility"])
 
 ## AttentiveSkin
 To Predict Skin Corrosion/Irritation Potentials of Chemicals via Explainable Machine Learning Methods