Documentation bug: dgl_csr_neighbor_non_uniform_sample doesn't have parameter seed_arrays

[DNXie]

Description

In the documentation of these 8 APIs listed below, there is a parameter seed_arrays listed in Parameter section. But it is not in the signature and not accepted by the functions, and passing them gives abort, which crashes the program.

• mxnet.contrib.ndarray.dgl_csr_neighbor_non_uniform_sample: https://mxnet.apache.org/versions/1.6/api/python/docs/api/contrib/ndarray/index.html#mxnet.contrib.ndarray.dgl_csr_neighbor_non_uniform_sample
• mxnet.contrib.ndarray.dgl_csr_neighbor_uniform_sample: https://mxnet.apache.org/versions/1.6/api/python/docs/api/contrib/ndarray/index.html#mxnet.contrib.ndarray.dgl_csr_neighbor_uniform_sample
• mxnet.contrib.symbol.dgl_csr_neighbor_non_uniform_sample: https://mxnet.apache.org/versions/1.6/api/python/docs/api/contrib/symbol/index.html#mxnet.contrib.symbol.dgl_csr_neighbor_non_uniform_sample
• mxnet.contrib.symbol.dgl_csr_neighbor_uniform_sample: https://mxnet.apache.org/versions/1.6/api/python/docs/api/contrib/symbol/index.html#mxnet.contrib.symbol.dgl_csr_neighbor_uniform_sample
• mxnet.ndarray.contrib.dgl_csr_neighbor_non_uniform_sample: https://mxnet.apache.org/versions/1.6/api/python/docs/api/ndarray/contrib/index.html#mxnet.ndarray.contrib.dgl_csr_neighbor_non_uniform_sample
• mxnet.ndarray.contrib.dgl_csr_neighbor_uniform_sample: https://mxnet.apache.org/versions/1.6/api/python/docs/api/ndarray/contrib/index.html#mxnet.ndarray.contrib.dgl_csr_neighbor_uniform_sample
• mxnet.symbol.contrib.dgl_csr_neighbor_non_uniform_sample: https://mxnet.apache.org/versions/1.6/api/python/docs/api/symbol/contrib/index.html#mxnet.symbol.contrib.dgl_csr_neighbor_non_uniform_sample
• mxnet.symbol.contrib.dgl_csr_neighbor_uniform_sample: https://mxnet.apache.org/versions/1.6/api/python/docs/api/symbol/contrib/index.html#mxnet.symbol.contrib.dgl_csr_neighbor_uniform_sample

Error Message

(Paste the complete error message. Please also include stack trace by setting environment variable DMLC_LOG_STACK_TRACE_DEPTH=10 before running your script.)

terminate called after throwing an instance of 'dmlc::ParamError'
  what():  Cannot find argument 'seed_arrays', Possible Arguments:
----------------
num_args : int, required
    Number of input NDArray.
num_hops : long, optional, default=1
    Number of hops.
num_neighbor : long, optional, default=2
    Number of neighbor.
max_num_vertices : long, optional, default=100
    Max number of vertices.
, in operator _contrib_dgl_csr_neighbor_non_uniform_sample(name="", max_num_vertices="5", num_hops="1", num_args="3", num_neighbor="2", seed_arrays="
[0 1 2 3 4]
<NDArray 5 @cpu(0)>", probability="
[0.9 0.8 0.2 0.4 0.1]
<NDArray 5 @cpu(0)>", csr_matrix="
<CSRNDArray 5x5 @cpu(0)>")
Aborted (core dumped)

To Reproduce

(If you developed your own code, please provide a short script that reproduces the error. For existing examples, please provide link.)

This is the exact example provided in the documentation.

import mxnet as mx
import numpy as np
shape = (5, 5)
prob = mx.nd.array([0.9, 0.8, 0.2, 0.4, 0.1], dtype=np.float32)
data_np = np.array([1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20], dtype=np.int64)
indices_np = np.array([1,2,3,4,0,2,3,4,0,1,3,4,0,1,2,4,0,1,2,3], dtype=np.int64)
indptr_np = np.array([0,4,8,12,16,20], dtype=np.int64)
a = mx.nd.sparse.csr_matrix((data_np, indices_np, indptr_np), shape=shape)
seed = mx.nd.array([0,1,2,3,4], dtype=np.int64)
out = mx.nd.contrib.dgl_csr_neighbor_non_uniform_sample(csr_matrix=a, probability=prob, seed_arrays=seed, num_args=3, num_hops=1, num_neighbor=2, max_num_vertices=5)

The same error when calling the rest 7 APIs.

Environment

We recommend using our script for collecting the diagnositc information. Run the following command and paste the outputs below:

curl --retry 10 -s https://raw.githubusercontent.com/dmlc/gluon-nlp/master/tools/diagnose.py | python

# paste outputs here

OS: ubuntu 18.04
Python: 3.7.6
pip: 20.0.2
numpy: 1.18.5
mxnet: 1.6.0

[szha]

cc @lingfanyu

[jermainewang]

cc'ed @zheng-da here. All these APIs should have been supported in DGL. We could have a discussion on whether to continue support them in MXNet or deprecate them.

[szha]

@jermainewang it's ok to deprecate them in 2.0. for 1.x since it's already in the code base, we will need to fix the doc.

[hadevin]

I found this issue on ovio.org and would love to contribute! I've never contributed before.

[szha]

@hadevin thanks for offering to help. The contribution guide for our project can be found at: https://mxnet.apache.org/versions/master/community#contribution-guides

For this issue, I think we need to add a note on the documentation that the expected number of arrays for seed_arrays is the num_args minus two (csr matrix and probability).

The operators with documentation problems are in https://github.com/apache/incubator-mxnet/blob/v1.x/src/operator/contrib/dgl_graph.cc. For example:
https://github.com/apache/incubator-mxnet/blob/787416b23a1d5730fce995a6b662a51cee45e20f/src/operator/contrib/dgl_graph.cc#L866-L935

Feel free to ask here if you have any question.