[SciPy-dev] Umfpack deprecation warning

[Pauli Virtanen]

Mon, 01 Dec 2008 10:03:34 -0500, Nathan Bell wrote:
> On Mon, Dec 1, 2008 at 8:06 AM, Pauli Virtanen <pav at iki.fi> wrote:
>>
>> This is a bit annoying as it comes up basically for all scripts that
>> import Scipy.
>>
>>
> I don't have umfpack installed, so I can't test it myself.  Please try
> the following:

Excellent, this seems to work. (However, finding the Umfpack scikit is 
not so easy, as the code is only in SVN.)

Btw, would it make sense to raise the warning even though the scikit is 
installed (and remove the umfpack support altogether in the future), and 
modify scikits.umfpack so that the user could call scikits.umfpack.solve?

Can we get this to 0.7.0?

Like so:

>>> import scipy
>>> import scipy.sparse
>>> import scipy.sparse.linalg
>>> a = scipy.sparse.coo_matrix([[1,2],[3,4]])
>>> b = scipy.array([5,6])
>>> scipy.sparse.linalg.spsolve(a, b)
/usr/local/lib/python2.5/site-packages/scipy-0.7.0.dev5203-py2.5-linux-
x86_64.egg/scipy/sparse/linalg/dsolve/linsolve.py:62: 
SparseEfficiencyWarning: spsolve requires CSC or CSR matrix format
  warn('spsolve requires CSC or CSR matrix format', 
SparseEfficiencyWarning)
/usr/local/lib/python2.5/site-packages/scipy-0.7.0.dev5203-py2.5-linux-
x86_64.egg/scipy/sparse/linalg/dsolve/linsolve.py:78: DeprecationWarning: 
scipy.sparse.linalg.dsolve.umfpack will be removed, install 
scikits.umfpack instead
  ' install scikits.umfpack instead', DeprecationWarning )
array([-4. ,  4.5])
>>> scipy.sparse.linalg.spsolve(a, b)
array([-4. ,  4.5])
>>> scipy.sparse.linalg.factorized(a)
/usr/local/lib/python2.5/site-packages/scipy-0.7.0.dev5203-py2.5-linux-
x86_64.egg/scipy/sparse/linalg/dsolve/linsolve.py:143: 
DeprecationWarning: scipy.sparse.linalg.dsolve.umfpack will be removed, 
install scikits.umfpack instead
  ' install scikits.umfpack instead', DeprecationWarning )
/usr/local/lib/python2.5/site-packages/scipy-0.7.0.dev5203-py2.5-linux-
x86_64.egg/scipy/sparse/linalg/dsolve/linsolve.py:147: 
SparseEfficiencyWarning: splu requires CSC matrix format
  warn('splu requires CSC matrix format', SparseEfficiencyWarning)
<function solve at 0x2ba7eadfbc08>
>>> scipy.sparse.linalg.factorized(a)
<function solve at 0x2ba7eadff578>

and after installing the scikit:

>>> import scipy
>>> import scipy.sparse
>>> import scipy.sparse.linalg
>>> a = scipy.sparse.coo_matrix([[1,2],[3,4]])
>>> b = scipy.array([5,6])
>>> scipy.sparse.linalg.spsolve(a, b)
/usr/local/lib/python2.5/site-packages/scipy-0.7.0.dev5203-py2.5-linux-
x86_64.egg/scipy/sparse/linalg/dsolve/linsolve.py:62: 
SparseEfficiencyWarning: spsolve requires CSC or CSR matrix format
  warn('spsolve requires CSC or CSR matrix format', 
SparseEfficiencyWarning)
array([-4. ,  4.5])
>>> scipy.sparse.linalg.factorized(a)
/usr/local/lib/python2.5/site-packages/scipy-0.7.0.dev5203-py2.5-linux-
x86_64.egg/scipy/sparse/linalg/dsolve/linsolve.py:147: 
SparseEfficiencyWarning: splu requires CSC matrix format
  warn('splu requires CSC matrix format', SparseEfficiencyWarning)
<function solve at 0x2ac1391a1488>

-- 
Pauli Virtanen