Merge branch 'pddbind_load' of https://github.com/rbharath/deepchem into pddbind_load

from deepchem.molnet.load_function.muv_datasets import load_muv

from deepchem.molnet.load_function.nci_datasets import load_nci

from deepchem.molnet.load_function.pcba_datasets import load_pcba, load_pcba_146, load_pcba_2475

from deepchem.molnet.load_function.pdbbind_datasets import load_pdbbind_grid

from deepchem.molnet.load_function.pdbbind_datasets import load_pdbbind_grid, extract_pdbbind

from deepchem.molnet.load_function.ppb_datasets import load_ppb

from deepchem.molnet.load_function.qm7_datasets import load_qm7

from deepchem.molnet.load_function.qm7_datasets import load_qm7_from_mat, load_qm7b_from_mat

import deepchem

import numpy as np

import pandas as pd

import logging

import tarfile

logger = logging.getLogger(__name__)

def featurize_pdbbind(data_dir=None, feat="grid", subset="core"):

                      featurizer="grid",

                      subset="core",

                      reload=True):

  """Load PDBBind datasets. Does not do train/test split"""

  """Load PDBBind datasets."""

  if featurizer == 'grid':

    dataset, tasks = featurize_pdbbind(feat=featurizer, subset=subset)

                                               transformers)

    return tasks, (train, valid, test), transformers

def extract_pdbbind(featurizer, split="random", subset="core", reload=True):

  """Load and featurize raw PDBBind dataset."""

  # TODO(rbharath): This should contain a cluster split so structures of the

  # same protein aren't in both train/test

  tasks = ["-logKd/Ki"]

  data_dir = deepchem.utils.get_data_dir()

  dataset_file = os.path.join(data_dir, "pdbbind_v2015.tar.gz")

  if subset == "core":

    data_folder = os.path.join(data_dir, "pdbbind_core")

  elif subset == "refined":

    data_folder = os.path.join(data_dir, "pdbbind_refined")

  else:

    raise ValueError("Unsupported subset %s." % subset)

  if os.path.exists(data_folder):

    logger.info("Data directory for %s already exists" % subset)

  print("dataset_file")

  print(dataset_file)

  print("os.path.exists(dataset_file)")

  print(os.path.exists(dataset_file))

  if not os.path.exists(dataset_file):

    logger.warning("About to download PDBBind full dataset. Large file, 2GB")

    deepchem.utils.download_url(

        'http://deepchem.io.s3-website-us-west-1.amazonaws.com/datasets/' +

        "pdbbind_v2015.tar.gz")

  tar = tarfile.open(dataset_file, "r:gz")

  pdbs = []

  if subset == "core":

    f = tar.extractfile("v2015/INDEX_core_name.2013")

    contentlines = f.readlines()

    for line in contentlines:

      line = line.decode("utf-8")

      line = line.split(" ")

      pdb = line[0]

      print("pdb")

      print(pdb)

      print("len(pdb)")

      print(len(pdb))

      # TODO(rbharath): Why 6 instead of 4?

      if len(pdb) == 4:

        pdbs.append(pdb)

  elif subset == "refined":

    f = tar.extractfile("v2015/INDEX_refined_name.2015")

    contentlines = f.readlines()

    for line in contentlines:

      line = str(line)

      line = line.split(" ")

      pdb = line[0]

      if len(pdb) == 4:

        pdbs.append(pdb)

  else:

    raise ValueError("Other subsets not supported.")

  print("pdbs")

  print(pdbs)

  # Make dir

  if not os.path.exists(data_folder):

    os.makedirs(data_folder)

  for ind, pdb in enumerate(pdbs):

    protein_filename = "v2015/" + pdb + "/" + pdb + "_protein.pdb"

    ligand_filename = "v2015/" + pdb + "/" + pdb + "_ligand.sdf"

    print("ind")

    print(ind)

    print("protein_filename, ligand_filename")

    print(protein_filename, ligand_filename)

    protein_f = tar.extractfile(protein_filename)

    protein_lines = protein_f.readlines()

    ligand_f = tar.extractfile(ligand_filename)

    ligand_lines = ligand_f.readlines()

    print("read lines")

    protein_out = os.path.join(data_folder, pdb + "_protein.pdb")

    with open(protein_out, "w") as f:

      print("type(protein_lines)")

      print(type(protein_lines))

      f.writelines(protein_lines)

    ligand_out = os.path.join(data_folder, pdb + "_ligand.sdf")

    with open(ligand_out, "w") as f:

      f.writelines(ligand_lines)

    return pdbs

#  for member in tar.getmembers():

#    print("member.name")

#    print(member.name)

#    if member.name == "v2015/INDEX_core_name.2013":

#      f = tar.extractfile(member)

#      contentlines = f.read()

#      print("content")

#      print(content)

#      break