Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -72,4 +72,82 @@ We utilized the raw data uploaded on [Github](https://github.com/LumosBio/MolDat
|
|
72 |
3. Split the dataset (train, test, validation)
|
73 |
|
74 |
If you would like to try these processes with the original dataset,
|
75 |
-
please follow the instructions in the [Preprocessing Script.py](address) file located in our MolData repository.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
72 |
3. Split the dataset (train, test, validation)
|
73 |
|
74 |
If you would like to try these processes with the original dataset,
|
75 |
+
please follow the instructions in the [Preprocessing Script.py](address) file located in our MolData repository.
|
76 |
+
|
77 |
+
|
78 |
+
|
79 |
+
## Quickstart Usage
|
80 |
+
|
81 |
+
### Load a dataset in python
|
82 |
+
Each subset can be loaded into python using the Huggingface [datasets](https://huggingface.co/docs/datasets/index) library.
|
83 |
+
First, from the command line install the `datasets` library
|
84 |
+
|
85 |
+
$ pip install datasets
|
86 |
+
|
87 |
+
then, from within python load the datasets library
|
88 |
+
|
89 |
+
>>> import datasets
|
90 |
+
|
91 |
+
and load the `MolData` datasets, e.g.,
|
92 |
+
|
93 |
+
>>> MolData = datasets.load_dataset("maomlab/MolData")
|
94 |
+
Downloading readme: 100%|ββββββββββ| 5.23k/5.23k [00:00<00:00, 35.1kkB/s]
|
95 |
+
Downloading data: 100%|ββββββββββ|β34.5k//34.5k/ [00:00<00:00,β155kB/s]
|
96 |
+
Downloading data: 100%|ββββββββββ| 97.1k/97.1k [00:00<00:00,β587kB/s]
|
97 |
+
Generating test split: 100%|ββββββββββ| 594/594 [00:00<00:00, 12705.92βexamples/s]
|
98 |
+
Generating train split: 100%|ββββββββββ| 1788/1788 [00:00<00:00, 43895.91βexamples/s]
|
99 |
+
|
100 |
+
and inspecting the loaded dataset
|
101 |
+
|
102 |
+
>>> MolData
|
103 |
+
MolData
|
104 |
+
DatasetDict({
|
105 |
+
test: Dataset({
|
106 |
+
features: ['SMILES', 'Y'],
|
107 |
+
num_rows: 594
|
108 |
+
})
|
109 |
+
train: Dataset({
|
110 |
+
features: ['SMILES', 'Y'],
|
111 |
+
num_rows: 1788
|
112 |
+
})
|
113 |
+
})
|
114 |
+
|
115 |
+
### Use a dataset to train a model
|
116 |
+
One way to use the dataset is through the [MolFlux](https://exscientia.github.io/molflux/) package developed by Exscientia.
|
117 |
+
First, from the command line, install `MolFlux` library with `catboost` and `rdkit` support
|
118 |
+
|
119 |
+
pip install 'molflux[catboost,rdkit]'
|
120 |
+
|
121 |
+
then load, featurize, split, fit, and evaluate the catboost model
|
122 |
+
|
123 |
+
import json
|
124 |
+
from datasets import load_dataset
|
125 |
+
from molflux.datasets import featurise_dataset
|
126 |
+
from molflux.features import load_from_dicts as load_representations_from_dicts
|
127 |
+
from molflux.splits import load_from_dict as load_split_from_dict
|
128 |
+
from molflux.modelzoo import load_from_dict as load_model_from_dict
|
129 |
+
from molflux.metrics import load_suite
|
130 |
+
|
131 |
+
Split and evaluate the catboost model
|
132 |
+
|
133 |
+
split_dataset = load_dataset('maomlab/MolData', name = 'MolData')
|
134 |
+
|
135 |
+
split_featurised_dataset = featurise_dataset(
|
136 |
+
split_dataset,
|
137 |
+
column = "SMILES",
|
138 |
+
representations = load_representations_from_dicts([{"name": "morgan"}, {"name": "maccs_rdkit"}]))
|
139 |
+
|
140 |
+
model = load_model_from_dict({
|
141 |
+
"name": "cat_boost_regressor",
|
142 |
+
"config": {
|
143 |
+
"x_features": ['SMILES::morgan', 'SMILES::maccs_rdkit'],
|
144 |
+
"y_features": ['Y']}})
|
145 |
+
|
146 |
+
model.train(split_featurised_dataset["train"])
|
147 |
+
preds = model.predict(split_featurised_dataset["test"])
|
148 |
+
|
149 |
+
regression_suite = load_suite("regression")
|
150 |
+
|
151 |
+
scores = regression_suite.compute(
|
152 |
+
references=split_featurised_dataset["test"]['Y'],
|
153 |
+
predictions=preds["cat_boost_regressor::Y"])
|