✅ add pyg tests and update dependecies

You install the nightly build version via pip. The nightly version is built by the HEAD of DeepChem.

```bash

pip install tensorflow==2.2

pip install tensorflow==2.2.0

pip install --pre deepchem

```

      from rdkit.Chem import rdMolDescriptors

    except ModuleNotFoundError:

      raise ValueError("This class requires RDKit to be installed.")

    self.radius = radius

    self.size = size

    self.chiral = chiral

    assert graph.num_edges == num_edges

    assert graph.num_edge_features == num_edge_features

    # check to_pyg_data function

    target = np.array([1], dtype=np.float)

    pyg_graph = graph.to_pyg_data(target)

    from torch_geometric.data import Data

    assert isinstance(pyg_graph, Data)

  def test_invalid_graph_data(self):

    with pytest.raises(ValueError):

      invalid_node_features_type = list(np.random.random_sample((5, 5)))

    assert batch.num_edges == sum(num_edge_list)

    assert batch.num_edge_features == num_edge_features

    assert batch.graph_index.shape == (sum(num_nodes_list),)

    # check to_pyg_data function

    targets = np.array([1, 2, 3], dtype=np.float)

    batch = BatchGraphData.to_pyg_data(graph_list=graphs, targets=targets)

    from torch_geometric.data import Batch

    assert isinstance(pyg_graph, Batch)

"""

Test featurizers for inorganic crystals.

"""

import numpy as np

import unittest

import numpy as np

from deepchem.feat.material_featurizers \

  import ElementPropertyFingerprint, SineCoulombMatrix, CGCNNFeaturizer

from deepchem.feat import ElementPropertyFingerprint, SineCoulombMatrix, CGCNNFeaturizer

class TestMaterialFeaturizers(unittest.TestCase):

  `GridHyperparamOpt`. `param_dict[hp]` must be an int/float and is

  used as the center of a search range.

  Example

  -------

  Examples

  --------

  This example shows the type of constructor function expected.

  >>> import sklearn