mvn: Importing the multiarray numpy extension module failed about jpmml-sparkml HOT 5 CLOSED

It is now possible to build the JPMML-SparkML-Package EGG file even if there is no local copy of Apache Spark available.

Original error was: cannot import name multiarray

Your Numpy (or Python) installation is broken. This message has got nothing to do with JPMML-SparkML and/or PySpark.

from jpmml-sparkml.

Other packages work correctly - this applies also to numpy.
However, when I am trying to build the JPMML-SparkML-Package I am getting:

Traceback (most recent call last):
  File "setup.py", line 3, in <module>
    from jpmml_sparkml import __license__, __version__
  File "/home/bluedata/jpmml-sparkml-package/target/egg-sources/jpmml_sparkml/__init__.py", line 4, in <module>
    from pyspark.ml.common import _py2java
  File "/usr/lib/spark/spark-2.1.1-bin-hadoop2.7/python/pyspark/ml/__init__.py", line 22, in <module>
    from pyspark.ml.base import Estimator, Model, Transformer
  File "/usr/lib/spark/spark-2.1.1-bin-hadoop2.7/python/pyspark/ml/base.py", line 21, in <module>
    from pyspark.ml.param import Params
  File "/usr/lib/spark/spark-2.1.1-bin-hadoop2.7/python/pyspark/ml/param/__init__.py", line 26, in <module>
    import numpy as np
ImportError: No module named numpy
[ERROR] Command execution failed.

So... looks like this might be a compatibility issue. Numpy is installed.
Which numpy and Python versions are supported?
On my side:
numpy==1.13.0
python --version == Python 2.6.6
python3 --version == Python 3.4.5

from jpmml-sparkml.

The JPMML-SparkML package does not interact with the NumPy package in any way. The communication, if any, is mediated by the PySpark package.

Your ImportError: No module named numpy is raised by the PySpark package. So, you should open another issue in the PySpark (appears to be a submodule of Apache Spark) issue tracker.

from jpmml-sparkml.

Also, please update your local copy of the JPMML-SparkML-Package codebase to the HEAD version before filing any complaints. At the moment you're "troubleshooting" a legacy version (eg. on line 3 of setup.py there is no from jpmml_sparkml import __license__, __version__ anymore).

from jpmml-sparkml.

I downloaded the jpmml-sparkml-package just couple of days ago so it looks like a new updates were released just recently. Whatever was the cause I see that the bug is now fixed - I removed the package and cloned the git repo again and... I could successfully build it.

from jpmml-sparkml.