Describe the bug

When trying to run cartesian from sklearn.utils.extmath on more 32 arrays, the module crashes due to np.indices functioning rejecting the input shape in the source code.

Steps/Code to Reproduce

cartesian(([(1, 2), (3, 4)], [(5, 6), (7, 8)], [(9, 10), (7, 11)], [(3, 12), (13, 14)], [(9, 15), (3, 16)], [(17, 18), (3, 19)], [(1, 20), (21, 22)], [(21, 23), (3, 24)], [(25, 26), (27, 28)], [(21, 29), (3, 30)], [(9, 31), (27, 32)], [(9, 33), (1, 34)], [(21, 35), (17, 36)], [(13, 37), (7, 38)], [(25, 39), (7, 40)], [(1, 41), (27, 42)], [(21, 43), (17, 44)], [(17, 45), (3, 46)], [(25, 47), (17, 48)], [(21, 49), (17, 50)], [(5, 51), (13, 52)], [(1, 53), (7, 54)], [(25, 55), (13, 56)], [(5, 57), (7, 58)], [(9, 59), (1, 60)], [(25, 61), (5, 62)], [(3, 63), (27, 64)], [(25, 65), (7, 66)], [(1, 67), (27, 68)], [(27, 69), (7, 70)], [(21, 71), (17, 72)], [(17, 73), (5, 74)], [(9, 75), (21, 76)], [(21, 77), (13, 78)], [(25, 79), (3, 80)], [(9, 81), (25, 82)], [(9, 83), (7, 84)], [(17, 85), (27, 86)], [(5, 87), (27, 88)], [(25, 89), (13, 90)]))

Example:

from sklearn.feature_extraction.text import CountVectorizer
from sklearn.decomposition import LatentDirichletAllocation
docs = ["Help I have a bug" for i in range(1000)]
vectorizer = CountVectorizer(input=docs, analyzer='word')
lda_features = vectorizer.fit_transform(docs)
lda_model = LatentDirichletAllocation(
    n_topics=10,
    learning_method='online',
    evaluate_every=10,
    n_jobs=4,
)
model = lda_model.fit(lda_features)

If the code is too long, feel free to put it in a public gist and link
it in the issue: https://gist.github.com
-->

Sample code to reproduce the problem

Expected Results

The cartesian product of the lists.

Actual Results

Traceback (most recent call last):
File "/home/hlc5v/.conda/envs/graph_env/lib/python3.6/runpy.py", line 193, in _run_module_as_main
"main", mod_spec)
File "/home/hlc5v/.conda/envs/graph_env/lib/python3.6/runpy.py", line 85, in _run_code
exec(code, run_globals)
File "/home/hlc5v/.conda/envs/graph_env/lib/python3.6/cProfile.py", line 160, in
main()
File "/home/hlc5v/.conda/envs/graph_env/lib/python3.6/cProfile.py", line 153, in main
runctx(code, globs, None, options.outfile, options.sort)
File "/home/hlc5v/.conda/envs/graph_env/lib/python3.6/cProfile.py", line 20, in runctx
filename, sort)
File "/home/hlc5v/.conda/envs/graph_env/lib/python3.6/profile.py", line 64, in runctx
prof.runctx(statement, globals, locals)
File "/home/hlc5v/.conda/envs/graph_env/lib/python3.6/cProfile.py", line 100, in runctx
exec(cmd, globals, locals)
File "../../DNF_Approx/GraphConvert/graphtodnf.py", line 191, in
main()
File "../../DNF_Approx/GraphConvert/graphtodnf.py", line 176, in main
pi, vars, weights = createPathSet(curr, infected, args.susceptible[j], t=int(args.time_steps))
File "../../DNF_Approx/GraphConvert/graphtodnf.py", line 103, in createPathSet
PI = cartesian_product(*pathSet)
File "../../DNF_Approx/GraphConvert/graphtodnf.py", line 113, in cartesian_product
arr = np.empty([len(a) for a in arrays] + [la], dtype=tuple)
ValueError: sequence too large; cannot be greater than 32

Versions

System:
python: 3.7.1 (v3.7.1:260ec2c36a, Oct 20 2018, 14:57:15) [MSC v.1915 64 bit (AMD64)]
executable: C:\Users\hcars\PycharmProjects\DNF_git\GraphForecastApprox\venv\Scripts\python.exe
machine: Windows-10-10.0.17763-SP0
Python dependencies:
pip: 19.0.3
setuptools: 40.8.0
sklearn: 0.22.1
numpy: 1.17.2
scipy: 1.4.1
Cython: None
pandas: 0.25.1
matplotlib: 3.1.1
joblib: 0.14.1
Built with OpenMP: True