grid-featurizer updates (bf06e813) · Commits · 钟慕尧 / deepchem

examples/pdbbind/pdbbind_datasets.py

+19 −9

Original line number	Diff line number	Diff line
		@@ -10,6 +10,7 @@ import os
		import numpy as np
		import pandas as pd
		import shutil
		import time
		from rdkit import Chem
		import deepchem as dc

		@@ -37,13 +38,13 @@ def compute_pdbbind_features(grid_featurizer, pdb_subdir, pdb_code):
		features = np.squeeze(features)
		return features

		def load_pdbbind_grid(split="index", feat="grid", subset="core"):
		"""Load PDBBind datasets. Does not do train/test split"""
		# Set some global variables up top
		regen = False

		# Create some directories for analysis
		def featurize_pdbbind(data_dir=None, feat="grid", subset="core"):
		"""Featurizes pdbbind according to provided featurization"""
		tasks = ["-logKd/Ki"]
		current_dir = os.path.dirname(os.path.realpath(__file__))
		data_dir = os.path.join(current_dir, "%s_%s" % (subset, feat))
		if os.path.exists(data_dir):
		return dc.data.DiskDataset(data_dir), tasks
		pdbbind_dir = os.path.join(current_dir, "v2015")

		# Load PDBBind dataset
		@@ -55,8 +56,9 @@ def load_pdbbind_grid(split="index", feat="grid", subset="core"):
		labels_file = os.path.join(pdbbind_dir, "INDEX_general_PL_data.2015")
		else:
		raise ValueError("Only core, refined, and full subsets supported.")
		tasks = ["-logKd/Ki"]
		print("About to load contents.")
		if not os.path.exists(labels_file):
		raise ValueError("Run get_pdbbind.sh to download dataset.")
		contents_df = load_pdbbind_labels(labels_file)
		ids = contents_df["PDB code"].values
		y = np.array([float(val) for val in contents_df["-logKd/Ki"].values])
		@@ -84,6 +86,7 @@ def load_pdbbind_grid(split="index", feat="grid", subset="core"):
		features = []
		feature_len = None
		y_inds = []
		time1 = time.time()
		for ind, pdb_code in enumerate(ids):
		print("Processing complex %d, %s" % (ind, str(pdb_code)))
		pdb_subdir = os.path.join(pdbbind_dir, pdb_code)
		@@ -96,11 +99,18 @@ def load_pdbbind_grid(split="index", feat="grid", subset="core"):
		continue
		y_inds.append(ind)
		features.append(computed_feature)
		time2 = time.time()
		print("TIMING: PDBBind Featurization took %0.3f s" % (time2-time1))
		y = y[y_inds]
		X = np.vstack(features)
		w = np.ones_like(y)

		dataset = dc.data.DiskDataset.from_numpy(X, y, w, ids)
		dataset = dc.data.DiskDataset.from_numpy(X, y, w, ids, data_dir=data_dir)
		return dataset, tasks

		def load_pdbbind_grid(split="index", feat="grid", subset="core"):
		"""Load PDBBind datasets. Does not do train/test split"""
		dataset, tasks = featurize_pdbbind(feat=feat, subset=subset)
		transformers = []

		splitters = {'index': dc.splits.IndexSplitter(),

examples/pdbbind/pdbbind_rf.py

0 → 100644

+42 −0

Original line number	Diff line number	Diff line
		"""
		Script that trains Sklearn RF models on PDBbind dataset.
		"""
		from __future__ import print_function
		from __future__ import division
		from __future__ import unicode_literals

		__author__ = "Bharath Ramsundar"
		__copyright__ = "Copyright 2016, Stanford University"
		__license__ = "GPL"

		import deepchem as dc
		import numpy as np
		from sklearn.ensemble import RandomForestRegressor
		from pdbbind_datasets import load_pdbbind_grid

		# For stable runs
		np.random.seed(123)

		pdbbind_tasks, pdbbind_datasets, transformers = load_pdbbind_grid(
		subset="core")
		train_dataset, valid_dataset, test_dataset = pdbbind_datasets

		metric = dc.metrics.Metric(dc.metrics.pearson_r2_score)

		n_features = train_dataset.X.shape[1]
		sklearn_model = RandomForestRegressor(n_estimators=500)
		model = dc.models.SklearnModel(sklearn_model)

		# Fit trained model
		model.fit(train_dataset, nb_epoch=20)
		model.save()

		print("Evaluating model")
		train_scores = model.evaluate(train_dataset, [metric], transformers)
		valid_scores = model.evaluate(valid_dataset, [metric], transformers)

		print("Train scores")
		print(train_scores)

		print("Validation scores")
		print(valid_scores)

examples/pdbbind/pdbbind_tf.py

+2 −2

Original line number	Diff line number	Diff line
		"""
		Script that trains Sklearn models on PDBbind dataset.
		Script that trains Tensorflow models on PDBbind dataset.
		"""
		from __future__ import print_function
		from __future__ import division
		@@ -16,7 +16,7 @@ from pdbbind_datasets import load_pdbbind_grid
		# For stable runs
		np.random.seed(123)

		pdbbind_tasks, pdbbind_datasets, transformers = load_pdbbind_grid()
		pdbbind_tasks, pdbbind_datasets, transformers = load_pdbbind_grid(subset="core")
		train_dataset, valid_dataset, test_dataset = pdbbind_datasets

		metric = dc.metrics.Metric(dc.metrics.pearson_r2_score)

Admin message