Re: [Numpy-discussion] Matrix Class
On 14-Feb-15 11:35 AM, josef.p...@gmail.com wrote:
On Wed, Feb 11, 2015 at 4:18 PM, Ryan Nelson rnelsonc...@gmail.com wrote:
Colin,
I currently use Py3.4 and Numpy 1.9.1. However, I built a quick test conda
environment with Python2.7 and Numpy 1.7.0, and I get the same:
Python 2.7.9 |Continuum Analytics, Inc.| (default, Dec 18 2014, 16:57:52)
[MSC v
.1500 64 bit (AMD64)]
Type copyright, credits or license for more information.
IPython 2.3.1 -- An enhanced Interactive Python.
Anaconda is brought to you by Continuum Analytics.
Please check out: http://continuum.io/thanks and https://binstar.org
? - Introduction and overview of IPython's features.
%quickref - Quick reference.
help - Python's own help system.
object? - Details about 'object', use 'object??' for extra details.
In [1]: import numpy as np
In [2]: np.__version__
Out[2]: '1.7.0'
In [3]: np.mat([4,'5',6])
Out[3]:
matrix([['4', '5', '6']],
dtype='|S1')
In [4]: np.mat([4,'5',6], dtype=int)
Out[4]: matrix([[4, 5, 6]])
###
As to your comment about coordinating with Statsmodels, you should see the
links in the thread that Alan posted:
http://permalink.gmane.org/gmane.comp.python.numeric.general/56516
http://permalink.gmane.org/gmane.comp.python.numeric.general/56517
Josef's comments at the time seem to echo the issues the devs (and others)
have with the matrix class. Maybe things have changed with Statsmodels.
Not changed, we have a strict policy against using np.matrix.
generic efficient versions for linear operators, kronecker or sparse
block matrix styly operations would be useful, but I would use array
semantics, similar to using dot or linalg functions on ndarrays.
Josef
(long reply canceled because I'm writing too much that might only be
of tangential interest or has been in some of the matrix discussion
before.)
Josef,
Many thanks. I have gained the impression that there is some antipathy
to np.matrix, perhaps this is because, as others have suggested, the
array doesn't provide an appropriate framework.
Where are such policy decisions documented? Numpy doesn't appear to
have a BDFL.
I had read Alan's links back in February and now have note of them.
Colin W.
I know I mentioned Sage and SageMathCloud before. I'll just point out that
there are folks that use this for real research problems, not just as a
pedagogical tool. They have a Matrix/vector/column_matrix class that do what
you were expecting from your problems posted above. Indeed below is a
(truncated) cut and past from a Sage Worksheet. (See
http://www.sagemath.org/doc/tutorial/tour_linalg.html)
##
In : Matrix([1,'2',3])
Error in lines 1-1
Traceback (most recent call last):
TypeError: unable to find a common ring for all elements
In : Matrix([[1,2,3],[4,5]])
ValueError: List of rows is not valid (rows are wrong types or lengths)
In : vector([1,2,3])
(1, 2, 3)
In : column_matrix([1,2,3])
[1]
[2]
[3]
##
Large portions of the custom code and wrappers in Sage are written in
Python. I don't think their Matrix object is a subclass of ndarray, so
perhaps you could strip out the Matrix stuff from here to make a separate
project with just the Matrix stuff, if you don't want to go through the Sage
interface.
On Wed, Feb 11, 2015 at 11:54 AM, cjw c...@ncf.ca wrote:
On 11-Feb-15 10:21 AM, Ryan Nelson wrote:
So:
In [2]: np.mat([4,'5',6])
Out[2]:
matrix([['4', '5', '6']], dtype='U11')
In [3]: np.mat([4,'5',6], dtype=int)
Out[3]: matrix([[4, 5, 6]])
Thanks Ryan,
We are not singing from the same hymn book.
Using PyScripter, I get:
*** Python 2.7.9 (default, Dec 10 2014, 12:28:03) [MSC v.1500 64 bit
(AMD64)] on win32. ***
import numpy as np
print('Numpy version: ', np.__version__)
('Numpy version: ', '1.9.0')
Could you say which version you are using please?
Colin W
On Tue, Feb 10, 2015 at 5:07 PM, cjw c...@ncf.ca wrote:
It seems to be agreed that there are weaknesses in the existing Numpy
Matrix
Class.
Some problems are illustrated below.
I'll try to put some suggestions over the coming weeks and would
appreciate
comments.
Colin W.
Test Script:
if __name__ == '__main__':
a= mat([4, 5, 6]) # Good
print('a: ', a)
b= mat([4, '5', 6]) # Not the expected result
print('b: ', b)
c= mat([[4, 5, 6], [7, 8]]) # Wrongly accepted as rectangular
print('c: ', c)
d= mat([[1, 2, 3]])
try:
d[0, 1]= 'b'# Correctly flagged, not numeric
except ValueError:
print(d[0, 1]= 'b' # Correctly flagged, not numeric,
'
ValueError')
print('d: ', d)
Result:
*** Python 2.7.9 (default, Dec 10 2014, 12:28:03) [MSC v.1500 64 bit
(AMD64)] on win32. ***
a: [[4 5 6]]
b: [['4' '5' '6']]
c: [[[4, 5, 6] [7, 8]]]
d[0, 1]= 'b' # Correctly flagged, not numeric ValueError
Re: [Numpy-discussion] Matrix Class
On Sat, Feb 14, 2015 at 12:05 PM, cjw c...@ncf.ca wrote: On 14-Feb-15 11:35 AM, josef.p...@gmail.com wrote: On Wed, Feb 11, 2015 at 4:18 PM, Ryan Nelson rnelsonc...@gmail.com wrote: Colin, I currently use Py3.4 and Numpy 1.9.1. However, I built a quick test conda environment with Python2.7 and Numpy 1.7.0, and I get the same: Python 2.7.9 |Continuum Analytics, Inc.| (default, Dec 18 2014, 16:57:52) [MSC v .1500 64 bit (AMD64)] Type copyright, credits or license for more information. IPython 2.3.1 -- An enhanced Interactive Python. Anaconda is brought to you by Continuum Analytics. Please check out: http://continuum.io/thanks and https://binstar.org ? - Introduction and overview of IPython's features. %quickref - Quick reference. help - Python's own help system. object? - Details about 'object', use 'object??' for extra details. In [1]: import numpy as np In [2]: np.__version__ Out[2]: '1.7.0' In [3]: np.mat([4,'5',6]) Out[3]: matrix([['4', '5', '6']], dtype='|S1') In [4]: np.mat([4,'5',6], dtype=int) Out[4]: matrix([[4, 5, 6]]) ### As to your comment about coordinating with Statsmodels, you should see the links in the thread that Alan posted: http://permalink.gmane.org/gmane.comp.python.numeric.general/56516 http://permalink.gmane.org/gmane.comp.python.numeric.general/56517 Josef's comments at the time seem to echo the issues the devs (and others) have with the matrix class. Maybe things have changed with Statsmodels. Not changed, we have a strict policy against using np.matrix. generic efficient versions for linear operators, kronecker or sparse block matrix styly operations would be useful, but I would use array semantics, similar to using dot or linalg functions on ndarrays. Josef (long reply canceled because I'm writing too much that might only be of tangential interest or has been in some of the matrix discussion before.) Josef, Many thanks. I have gained the impression that there is some antipathy to np.matrix, perhaps this is because, as others have suggested, the array doesn't provide an appropriate framework. It's not directly antipathy, it's cost-benefit analysis. np.matrix has few advantages, but makes reading and maintaining code much more difficult. Having to watch out for multiplication `*` is a lot of extra work. Checking shapes and fixing bugs with unexpected dtypes is also a lot of work, but we have large benefits. For a long time the policy in statsmodels was to keep pandas out of the core of functions (i.e. out of the actual calculations) and restrict it to inputs and returns. However, pandas is becoming more popular and can do some things much better than plain numpy, so it is slowly moving inside some of our core calculations. It's still an easy source of bugs, but we do gain something. Benefits like these don't exist for np.matrix. Where are such policy decisions documented? Numpy doesn't appear to have a BDFL. In general it's a mix of mailing list discussions and discussion in issues and PRs. I'm not directly involved in numpy and don't subscribe to the numpy's github notifications. For scipy (and partially for statsmodels): I think large parts of policies for code and workflow are not explicitly specified, but are more an understanding of maintainers and developers that can slowly change over time, build up through spread out discussion as temporary consensus (or without strong objections). scipy has a hacking text file to describe some of it, but I haven't read it in ages. (long term changes compared to 6 years ago: required code review and required test coverage.) Josef I had read Alan's links back in February and now have note of them. Colin W. I know I mentioned Sage and SageMathCloud before. I'll just point out that there are folks that use this for real research problems, not just as a pedagogical tool. They have a Matrix/vector/column_matrix class that do what you were expecting from your problems posted above. Indeed below is a (truncated) cut and past from a Sage Worksheet. (See http://www.sagemath.org/doc/tutorial/tour_linalg.html) ## In : Matrix([1,'2',3]) Error in lines 1-1 Traceback (most recent call last): TypeError: unable to find a common ring for all elements In : Matrix([[1,2,3],[4,5]]) ValueError: List of rows is not valid (rows are wrong types or lengths) In : vector([1,2,3]) (1, 2, 3) In : column_matrix([1,2,3]) [1] [2] [3] ## Large portions of the custom code and wrappers in Sage are written in Python. I don't think their Matrix object is a subclass of ndarray, so perhaps you could strip out the Matrix stuff from here to make a separate project with just the Matrix stuff, if you don't want to go through the Sage interface. On Wed, Feb 11, 2015 at 11:54 AM, cjw c...@ncf.ca wrote: On 11-Feb-15 10:21 AM, Ryan Nelson wrote: So: In [2]: np.mat([4,'5',6]) Out[2]:
Re: [Numpy-discussion] Matrix Class
On Sat, Feb 14, 2015 at 12:36 PM, josef.p...@gmail.com wrote: On Sat, Feb 14, 2015 at 12:05 PM, cjw c...@ncf.ca wrote: On 14-Feb-15 11:35 AM, josef.p...@gmail.com wrote: On Wed, Feb 11, 2015 at 4:18 PM, Ryan Nelson rnelsonc...@gmail.com wrote: Colin, I currently use Py3.4 and Numpy 1.9.1. However, I built a quick test conda environment with Python2.7 and Numpy 1.7.0, and I get the same: Python 2.7.9 |Continuum Analytics, Inc.| (default, Dec 18 2014, 16:57:52) [MSC v .1500 64 bit (AMD64)] Type copyright, credits or license for more information. IPython 2.3.1 -- An enhanced Interactive Python. Anaconda is brought to you by Continuum Analytics. Please check out: http://continuum.io/thanks and https://binstar.org ? - Introduction and overview of IPython's features. %quickref - Quick reference. help - Python's own help system. object? - Details about 'object', use 'object??' for extra details. In [1]: import numpy as np In [2]: np.__version__ Out[2]: '1.7.0' In [3]: np.mat([4,'5',6]) Out[3]: matrix([['4', '5', '6']], dtype='|S1') In [4]: np.mat([4,'5',6], dtype=int) Out[4]: matrix([[4, 5, 6]]) ### As to your comment about coordinating with Statsmodels, you should see the links in the thread that Alan posted: http://permalink.gmane.org/gmane.comp.python.numeric.general/56516 http://permalink.gmane.org/gmane.comp.python.numeric.general/56517 Josef's comments at the time seem to echo the issues the devs (and others) have with the matrix class. Maybe things have changed with Statsmodels. Not changed, we have a strict policy against using np.matrix. generic efficient versions for linear operators, kronecker or sparse block matrix styly operations would be useful, but I would use array semantics, similar to using dot or linalg functions on ndarrays. Josef (long reply canceled because I'm writing too much that might only be of tangential interest or has been in some of the matrix discussion before.) Josef, Many thanks. I have gained the impression that there is some antipathy to np.matrix, perhaps this is because, as others have suggested, the array doesn't provide an appropriate framework. It's not directly antipathy, it's cost-benefit analysis. np.matrix has few advantages, but makes reading and maintaining code much more difficult. Having to watch out for multiplication `*` is a lot of extra work. Checking shapes and fixing bugs with unexpected dtypes is also a lot of work, but we have large benefits. For a long time the policy in statsmodels was to keep pandas out of the core of functions (i.e. out of the actual calculations) and restrict it to inputs and returns. However, pandas is becoming more popular and can do some things much better than plain numpy, so it is slowly moving inside some of our core calculations. It's still an easy source of bugs, but we do gain something. Any bits of Pandas that might be good for numpy/scipy to steal? snip Chuck ___ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] Matrix Class
On Sat, Feb 14, 2015 at 4:27 PM, Charles R Harris charlesr.har...@gmail.com wrote: On Sat, Feb 14, 2015 at 12:36 PM, josef.p...@gmail.com wrote: On Sat, Feb 14, 2015 at 12:05 PM, cjw c...@ncf.ca wrote: On 14-Feb-15 11:35 AM, josef.p...@gmail.com wrote: On Wed, Feb 11, 2015 at 4:18 PM, Ryan Nelson rnelsonc...@gmail.com wrote: Colin, I currently use Py3.4 and Numpy 1.9.1. However, I built a quick test conda environment with Python2.7 and Numpy 1.7.0, and I get the same: Python 2.7.9 |Continuum Analytics, Inc.| (default, Dec 18 2014, 16:57:52) [MSC v .1500 64 bit (AMD64)] Type copyright, credits or license for more information. IPython 2.3.1 -- An enhanced Interactive Python. Anaconda is brought to you by Continuum Analytics. Please check out: http://continuum.io/thanks and https://binstar.org ? - Introduction and overview of IPython's features. %quickref - Quick reference. help - Python's own help system. object? - Details about 'object', use 'object??' for extra details. In [1]: import numpy as np In [2]: np.__version__ Out[2]: '1.7.0' In [3]: np.mat([4,'5',6]) Out[3]: matrix([['4', '5', '6']], dtype='|S1') In [4]: np.mat([4,'5',6], dtype=int) Out[4]: matrix([[4, 5, 6]]) ### As to your comment about coordinating with Statsmodels, you should see the links in the thread that Alan posted: http://permalink.gmane.org/gmane.comp.python.numeric.general/56516 http://permalink.gmane.org/gmane.comp.python.numeric.general/56517 Josef's comments at the time seem to echo the issues the devs (and others) have with the matrix class. Maybe things have changed with Statsmodels. Not changed, we have a strict policy against using np.matrix. generic efficient versions for linear operators, kronecker or sparse block matrix styly operations would be useful, but I would use array semantics, similar to using dot or linalg functions on ndarrays. Josef (long reply canceled because I'm writing too much that might only be of tangential interest or has been in some of the matrix discussion before.) Josef, Many thanks. I have gained the impression that there is some antipathy to np.matrix, perhaps this is because, as others have suggested, the array doesn't provide an appropriate framework. It's not directly antipathy, it's cost-benefit analysis. np.matrix has few advantages, but makes reading and maintaining code much more difficult. Having to watch out for multiplication `*` is a lot of extra work. Checking shapes and fixing bugs with unexpected dtypes is also a lot of work, but we have large benefits. For a long time the policy in statsmodels was to keep pandas out of the core of functions (i.e. out of the actual calculations) and restrict it to inputs and returns. However, pandas is becoming more popular and can do some things much better than plain numpy, so it is slowly moving inside some of our core calculations. It's still an easy source of bugs, but we do gain something. Any bits of Pandas that might be good for numpy/scipy to steal? I'm not a Pandas expert. Some of it comes into statsmodels because we need the data handling also inside a function, e.g. keeping track of labels, indices, and so on. Another reason is that contributors are more familiar with pandas's way of solving a problems, even if I suspect numpy would be more efficient. However, a recent change, replaces where I would have used np.unique with pandas.factorize which is supposed to be faster. https://github.com/statsmodels/statsmodels/pull/2213 Two or three years ago my numpy way of group handling (using np.unique, bincount and similar) was still faster than the pandas `apply` version, I'm not sure that's still true. And to emphasize: all our heavy stuff especially the big models still only have numpy and scipy inside (with the exception of one model waiting in a PR). Josef snip Chuck ___ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion ___ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion
Re: [Numpy-discussion] Matrix Class
On Sat, Feb 14, 2015 at 5:21 PM, josef.p...@gmail.com wrote: On Sat, Feb 14, 2015 at 4:27 PM, Charles R Harris charlesr.har...@gmail.com wrote: On Sat, Feb 14, 2015 at 12:36 PM, josef.p...@gmail.com wrote: On Sat, Feb 14, 2015 at 12:05 PM, cjw c...@ncf.ca wrote: On 14-Feb-15 11:35 AM, josef.p...@gmail.com wrote: On Wed, Feb 11, 2015 at 4:18 PM, Ryan Nelson rnelsonc...@gmail.com wrote: Colin, I currently use Py3.4 and Numpy 1.9.1. However, I built a quick test conda environment with Python2.7 and Numpy 1.7.0, and I get the same: Python 2.7.9 |Continuum Analytics, Inc.| (default, Dec 18 2014, 16:57:52) [MSC v .1500 64 bit (AMD64)] Type copyright, credits or license for more information. IPython 2.3.1 -- An enhanced Interactive Python. Anaconda is brought to you by Continuum Analytics. Please check out: http://continuum.io/thanks and https://binstar.org ? - Introduction and overview of IPython's features. %quickref - Quick reference. help - Python's own help system. object? - Details about 'object', use 'object??' for extra details. In [1]: import numpy as np In [2]: np.__version__ Out[2]: '1.7.0' In [3]: np.mat([4,'5',6]) Out[3]: matrix([['4', '5', '6']], dtype='|S1') In [4]: np.mat([4,'5',6], dtype=int) Out[4]: matrix([[4, 5, 6]]) ### As to your comment about coordinating with Statsmodels, you should see the links in the thread that Alan posted: http://permalink.gmane.org/gmane.comp.python.numeric.general/56516 http://permalink.gmane.org/gmane.comp.python.numeric.general/56517 Josef's comments at the time seem to echo the issues the devs (and others) have with the matrix class. Maybe things have changed with Statsmodels. Not changed, we have a strict policy against using np.matrix. generic efficient versions for linear operators, kronecker or sparse block matrix styly operations would be useful, but I would use array semantics, similar to using dot or linalg functions on ndarrays. Josef (long reply canceled because I'm writing too much that might only be of tangential interest or has been in some of the matrix discussion before.) Josef, Many thanks. I have gained the impression that there is some antipathy to np.matrix, perhaps this is because, as others have suggested, the array doesn't provide an appropriate framework. It's not directly antipathy, it's cost-benefit analysis. np.matrix has few advantages, but makes reading and maintaining code much more difficult. Having to watch out for multiplication `*` is a lot of extra work. Checking shapes and fixing bugs with unexpected dtypes is also a lot of work, but we have large benefits. For a long time the policy in statsmodels was to keep pandas out of the core of functions (i.e. out of the actual calculations) and restrict it to inputs and returns. However, pandas is becoming more popular and can do some things much better than plain numpy, so it is slowly moving inside some of our core calculations. It's still an easy source of bugs, but we do gain something. Any bits of Pandas that might be good for numpy/scipy to steal? I'm not a Pandas expert. Some of it comes into statsmodels because we need the data handling also inside a function, e.g. keeping track of labels, indices, and so on. Another reason is that contributors are more familiar with pandas's way of solving a problems, even if I suspect numpy would be more efficient. However, a recent change, replaces where I would have used np.unique with pandas.factorize which is supposed to be faster. https://github.com/statsmodels/statsmodels/pull/2213 Numpy could use some form of hash table for its arraysetops, which is where pandas is getting its advantage from. It is a tricky thing though, see e.g. these timings: a = np.ranomdom.randint(10, size=1000) srs = pd.Series(a) %timeit np.unique(a) 10 loops, best of 3: 13.2 µs per loop %timeit srs.unique() 10 loops, best of 3: 15.6 µs per loop %timeit pd.factorize(a) 1 loops, best of 3: 25.6 µs per loop %timeit np.unique(a, return_inverse=True) 1 loops, best of 3: 82.5 µs per loop This last timings are with 1.9.0 an 0.14.0, so numpy doesn't have https://github.com/numpy/numpy/pull/5012 yet, which makes the operation in which numpy is slower about 2x faster. And if you need your unique values sorted, then things are more even, especially if numpy runs 2x faster: %timeit pd.factorize(a, sort=True) 1 loops, best of 3: 36.4 µs per loop The algorithms scale differently though, so for sufficiently large data Pandas is going to win almost certainly. Not sure if they support all dtypes, nor how efficient their use of memory is. I did a toy implementation of a hash table, mimicking Python's dictionary, for numpy some time
