core dump when runs graphsage with ppi data

euler version: euler_gl-0.1.2-py2.7
I generate ppi example with `ppi_data.py`. And then run the command
```
python -m tf_euler \
> --data_dir ppi \
> --max_id 56944 --feature_idx 1 --feature_dim 50 --label_idx 0 --label_dim 121 \
> --model graphsage --mode train
```

The output is below:
```
I0917 03:12:03.831502 24604 graph_builder.cc:81] Load Done: ppi/ppi_data.dat
I0917 03:12:03.831904 24554 graph_builder.cc:127] Each Thread Load Finish! Node Count:56944 Edge Count:1612348
I0917 03:12:03.831966 24554 graph_builder.cc:135] Graph Loading Finish!
I0917 03:12:04.269271 24554 graph_builder.cc:147] Graph Load Finish! Node Count:56944 Edge Count:1612348
I0917 03:12:04.511008 24554 graph_builder.cc:159] Done: build all sampler
I0917 03:12:04.511050 24554 graph_builder.cc:162] Graph build finish
WARNING:tensorflow:From /usr/local/lib/python2.7/dist-packages/euler_gl-0.1.2-py2.7.egg/tf_euler/python/encoders.py:354: The name tf.logging.warning is deprecated. Please use tf.compat.v1.logging.warning instead.

W0917 03:12:04.512913 140283727066944 deprecation_wrapper.py:119] From /usr/local/lib/python2.7/dist-packages/euler_gl-0.1.2-py2.7.egg/tf_euler/python/encoders.py:354: The name tf.logging.warning is deprecated. Please use tf.compat.v1.logging.warning instead.

WARNING:tensorflow:use_feature is deprecated and would not have any effect.
W0917 03:12:04.513195 140283727066944 encoders.py:354] use_feature is deprecated and would not have any effect.
Segmentation fault (core dumped)
```

The py-bt information of core file is
```
(gdb) py-bt
Traceback (most recent call first):
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 1864, in _create_c_op
c_op = c_api.TF_FinishOperation(op_desc)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 2027, in __init__
control_input_ops)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 3616, in create_op
op_def=op_def)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/util/deprecation.py", line 507, in new_func
return func(*args, **kwargs)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/op_def_library.py", line 788, in _apply_op_helper
op_def=op_def)
File "", line 74, in sample_node
File "/usr/local/lib/python2.7/dist-packages/euler_gl-0.1.2-py2.7.egg/tf_euler/python/run_loop.py", line 104, in run_train
count=batch_size, node_type=flags_obj.train_node_type)
File "/usr/local/lib/python2.7/dist-packages/euler_gl-0.1.2-py2.7.egg/tf_euler/python/run_loop.py", line 357, in run_network_embedding
run_train(model, flags_obj, master, is_chief)
File "/usr/local/lib/python2.7/dist-packages/euler_gl-0.1.2-py2.7.egg/tf_euler/python/run_loop.py", line 368, in run_local
run(flags_obj, master='', is_chief=True)
File "/usr/local/lib/python2.7/dist-packages/euler_gl-0.1.2-py2.7.egg/tf_euler/python/run_loop.py", line 405, in main
run_local(flags_obj, run_network_embedding)
File "/usr/local/lib/python2.7/dist-packages/absl/app.py", line 250, in _run_main
sys.exit(main(argv))
File "/usr/local/lib/python2.7/dist-packages/absl/app.py", line 299, in run
_run_main(main, args)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/platform/app.py", line 40, in run
_run(main=main, argv=argv, flags_parser=_parse_flags_tolerate_undef)
File "/usr/local/lib/python2.7/dist-packages/euler_gl-0.1.2-py2.7.egg/tf_euler/__main__.py", line 28, in
tf.app.run(run_loop.main)
File "/usr/lib/python2.7/runpy.py", line 72, in _run_code
exec code in run_globals
File "/usr/lib/python2.7/runpy.py", line 174, in _run_module_as_main
"__main__", fname, loader, pkg_name)
```

Do you have any suggestions about the core-dump problem?

Thank you!

Reproduce the PPI GraphSAGE command and start with the sample_node call in run_loop.py, then inspect TensorFlow's _create_c_op stack and the graph_builder.cc loading path shown in the logs. Done means the same training command no longer ends in a segmentation fault and proceeds past graph construction.

Written by the indexing model from the issue text.