No module named 'deepchem.models.tensorgraph.models.graph_models'

Description:

Hi, I have installed deepchem by source. To start with, I want to run Graph Convolutional Networks tutorial. However, I fail to import deepchem.models.tensorgraph.models.graph_models:

$ cat GraphConv_Test.py 
from __future__ import division
from __future__ import print_function
from __future__ import unicode_literals
import numpy as np
import tensorflow as tf
import deepchem as dc
from deepchem.models.tensorgraph.models.graph_models import GraphConvModel
# Load Tox21 dataset
tox21_tasks, tox21_datasets, transformers = dc.molnet.load_tox21(featurizer='GraphConv')

$ python GraphConv_Test.py 
Traceback (most recent call last):
  File "GraphConv_Test.py", line 8, in <module>
    from deepchem.models.tensorgraph.models.graph_models import GraphConvModel
ImportError: No module named 'deepchem.models.tensorgraph.models.graph_models'

• Commit: ‘4f4a67491176f7bdfde6229391dee3223fcbf8d2’
• DeepChem Version: ‘2.1.1’
• Platform: MacOS (version 10.14.6)

[peastman]

The class has moved in the latest source code. Change the import to

from deepchem.models import GraphConvModel

I’m in the middle of updating the examples right now. I should have a PR with the changes in the next few days.

[vsomnath]

I was not able to reproduce the above error. The error is because a metadata.joblib file is missing.

I suspect this error is happening because the DiskDataset creation process was stopped midway and while directories were created, the files were not. You could try clearing out the tox21-featurized directory and retry. If it still fails, I’ll also have a look at it.