Python: undefined symbol: PyUnicodeUCS4_Decode

[Louis Luangkesorn]

This is actually a solved problem, but I'm posting this here because I have to do this again and I want to be able to find the answer, and it may be a sign that there is a problem in the setup.py script. (once I figure out where to send it, I'll submit it as a bug report)

When I was installing Gurobi python using Canopy I get

=======

import gurobipy

---------------------------------------------------------------------------

ImportError                               Traceback (most recent call last)

<ipython-input-3-1a30be95fe9a> in <module>()

----> 1 import gurobipy

/Canopy/Canopy_64bit/User/lib/python2.7/site-packages/gurobipy/__init__.py in <module>()

----> 1 from .gurobipy import *

ImportError: /Canopy/Canopy_64bit/User/lib/python2.7/site-packages/gurobipy/gurobipy.so: undefined symbol: PyUnicodeUCS4_Decode

=======

What I think is the underlying issue is that Canopy Python is 16bit unicode support (I think)

>>> sys.maxunicode

65535

But Gurobi's setup.py script thinks that it is 32 bit

I solved this by copying the following into my 

userdir/Canopy/Canopy_64bit/User/lib/python2.7/site-packages/gurobipy folder

sudo cp /opt/gurobi563/linux64/lib/python2.7_utf16/gurobipy.so  .

thus replacing the gurobipy.so file that was there.  And now it works.

>>> execfile('/opt/gurobi563/linux64/examples/python/mip1.py')

Optimize a model with 2 rows, 3 columns and 5 nonzeros

Found heuristic solution: objective 2

Presolve removed 2 rows and 3 columns

Presolve time: 0.02s

Explored 0 nodes (0 simplex iterations) in 0.05 seconds

Thread count was 1 (of 4 available processors)

Optimal solution found (tolerance 1.00e-04)

Best objective 3.000000000000e+00, best bound 3.000000000000e+00, gap 0.0%

x 1

y 0

z 1

Obj: 3

===========

I get the opposite problem in Continuum Anaconda Python

>>> sys.maxunicode

1114111

>>> import gurobipy

Traceback (most recent call last):

  File "<stdin>", line 1, in <module>

  File "userdir/anaconda/lib/python2.7/site-packages/gurobipy/__init__.py", line 1, in <module>

    from .gurobipy import *

ImportError: userdir/anaconda/lib/python2.7/site-packages/gurobipy/gurobipy.so: undefined symbol: PyUnicodeUCS2_Decode

I solved this by copying the following into  /anaconda/lib/python2.7/site-packages/gurobipy/

sudo cp /opt/gurobi563/linux64/lib/python2.7_utf32/gurobipy.so .

with the same (successful) result

Louis

========

Ubuntu 14.04

Gurobi 5.6.3

Enthought Canopy Python 2.7.6 | 64-bit | (default, Jun  4 2014, 16:32:15) 

[GCC 4.1.2 20080704 (Red Hat 4.1.2-52)] on linux2

Python 2.7.7 |Anaconda 2.0.1 (64-bit)| (default, Jun  2 2014, 12:34:02) 

[GCC 4.1.2 20080704 (Red Hat 4.1.2-54)] on linux2

[Jakob Sch.]

Hi Louis,

The two python distributions that you mention are not officially supported by gurobi. It's nice that you found a workaround for your problem and shared it with the community.

Best regards,
Jakob

[ALHUSSIN ABUDIYAH]

Hi

Can you clarify more about this method to work gurobi with Anaconda because I have the same problem

Best,

Alhussin

[Louis Luangkesorn]

Oops, sorry I did not see this.  In case this is still a live issue (I have not updated my Gurobi yet), the problem was the shared library /site-packages/gurobipy/gurobipy.so, which triggered the error: undefined symbol: PyUnicodeUCS2_Decode

The issue is that Anaconda is 32 bit unicode encoding,which you can verify within Python with 

>>> sys.maxunicode

1114111

So somehow the install thinks that it is 16 bit unicode thus uses the shared library that is built requiring PyUnicodeUCS2_Decode.  So the solution is to go into the Python installation and copy the correct gurobipy.so shared library from /python2.7_utf32/gurobipy.so and replace the /site-packages/gurobipy/gurobipy.so with the version compiled with 32 bit unicode support.

Hope this helps.

Louis