merge sage's numpy.pxd with the cython numpy.pxd #4571

jasongrout · 2008-11-20T21:56:56Z

[15:53] <robertwb> yep, we should merge them

CC: @robertwb @sagetrac-dagss

Component: numerical

Author: Robert Bradshaw, Dag Sverre Seljebotn

Reviewer: Robert Bradshaw, Dag Sverre Seljebotn, Minh Van Nguyen

Merged: Sage 4.1.1.alpha1

Issue created by migration from https://trac.sagemath.org/ticket/4571

The text was updated successfully, but these errors were encountered:

robertwb · 2009-06-28T09:00:20Z

comment:1

See http://trac.cython.org/cython_trac/ticket/339

robertwb · 2009-06-28T09:02:02Z

Attachment: 4571-numpy-pxd.patch.gz

robertwb · 2009-06-28T09:02:28Z

comment:3

Preliminary patch, need a new Cython spkg for it to work all the way.

sagetrac-dagss · 2009-07-07T15:54:38Z

comment:4

Attachment: 4571-more-fixes.patch.gz

Here are some more fixes. With this, and the latest state of the

http://hg.cython.org/cython

repo, all the relevant tests seem to pass. I do get two failures (due to Cython upgrade? something else?), see test.log in

/home/dagss/sage-4.0.2-sage.math.washington.edu-x86_64-Linux

Once this works let's tag http://hg.cython.org/cython as 0.11.2.1 for #6438?

sagetrac-dagss · 2009-07-07T15:57:04Z

comment:5

Note (re: discussion on Cython ML) that shape in numpy.pxd is still a field, long story (well, it is performance critical), and I shouldn't change it.

sagetrac-dagss · 2009-07-07T16:09:27Z

comment:6

Another quick comment: With this, full Cython/NumPy functionality is available from the notebook and works well.

robertwb · 2009-07-09T08:50:35Z

comment:7

Those errors seem completely unrelated, I'll see if I get them too. Other than that, it looks really good.

I've used Cython + NumPy from the notebook before, as long as you weren't in sage.ext then "cimport numpy" worked as expected (though there was a point where it didn't).

robertwb · 2009-07-21T20:52:44Z

comment:8

Hmm... I also got

The following tests failed:

        sage -t -long devel/sage-main/sage/rings/polynomial/polynomial_template.pxi # 1 doctests failed
        sage -t -long devel/sage-main/sage/interfaces/sage0.py # 1 doctests failed

robertwb · 2009-07-22T14:19:05Z

comment:9

Looks good. The single doctest failure was minor, fixed in #6438.

sagetrac-mvngu · 2009-07-22T16:17:32Z

Reviewer: Robert Bradshaw, Dag Sverre Seljebotn

sagetrac-mvngu · 2009-07-22T16:17:32Z

Author: Robert Bradshaw, Dag Sverre Seljebotn

sagetrac-mvngu · 2009-07-22T16:21:52Z

comment:11

Replying to @sagetrac-dagss:

Here are some more fixes. With this, and the latest state of the

The patch 4571-more-fixes.patch doesn't contain a commit message and your username is not on the patch. Your username should follow the format:

Full Name <email@somewhere.com>

The username makes it easy to identify the patch, when it is committed, as your contribution.

sagetrac-mvngu · 2009-07-23T02:23:03Z

comment:12

With both patches applied in the following order:

4571-numpy-pxd.patch
4571-more-fixes.patch
I see the following build failure:

[mvngu@sage sage-4.1.1.alpha0]$ ./sage -br main

----------------------------------------------------------
sage: Building and installing modified Sage library files.


Installing c_lib
scons: `install' is up to date.
Updating Cython code....
Building modified file sage/finance/time_series.pyx.
Building modified file sage/matrix/change_ring.pyx.
Building modified file sage/matrix/matrix_complex_double_dense.pyx.
Building modified file sage/matrix/matrix_double_dense.pyx.
Building modified file sage/matrix/matrix_real_double_dense.pyx.
Building modified file sage/modules/vector_complex_double_dense.pyx.
Building modified file sage/modules/vector_double_dense.pyx.
Building modified file sage/modules/vector_real_double_dense.pyx.
Building sage/rings/polynomial/real_roots.pyx because it depends on sage/modules/vector_double_dense.pxd.
Building sage/stats/hmm/chmm.pyx because it depends on sage/matrix/matrix_double_dense.pxd.
Building sage/stats/hmm/hmm.pyx because it depends on sage/matrix/matrix_double_dense.pxd.
python `which cython` --embed-positions --incref-local-binop -I/scratch/mvngu/release/sage-4.1.1.alpha0/devel/sage-main -o sage/finance/time_series.c sage/finance/time_series.pyx
warning: /scratch/mvngu/release/sage-4.1.1.alpha0/devel/sage-main/sage/finance/time_series.pyx:1722:24: cdef variable 'j' declared after it is used

Error converting Pyrex file to C:
------------------------------------------------------------
...
            [20.0000, -3.0000, 4.5000, -2.0000]
        """
        cnumpy.import_array() #This must be called before using the numpy C/api or you will get segfault
        cdef cnumpy.npy_intp dims[1]
        dims[0] = self._length
        cdef cnumpy.ndarray n = cnumpy.PyArray_SimpleNewFromData(1, dims, cnumpy.NPY_DOUBLE, self._values)
                                                                       ^
------------------------------------------------------------

/scratch/mvngu/release/sage-4.1.1.alpha0/devel/sage-main/sage/finance/time_series.pyx:1862:72: Cannot convert 'numpy.npy_intp [1]' to Python object

Error converting Pyrex file to C:
------------------------------------------------------------
...
            [20.0000, -3.0000, 4.5000, -2.0000]
        """
        cnumpy.import_array() #This must be called before using the numpy C/api or you will get segfault
        cdef cnumpy.npy_intp dims[1]
        dims[0] = self._length
        cdef cnumpy.ndarray n = cnumpy.PyArray_SimpleNewFromData(1, dims, cnumpy.NPY_DOUBLE, self._values)
                                                                                                ^
------------------------------------------------------------

/scratch/mvngu/release/sage-4.1.1.alpha0/devel/sage-main/sage/finance/time_series.pyx:1862:97: Cannot convert 'double *' to Python object
Error running command, failed with status 256.
sage: There was an error installing modified sage library code.

robertwb · 2009-07-23T08:50:21Z

comment:13

Looks like it'll have to be rebased, worked fine on vanilla 4.1. Where's the .tar on sage.math?

sagetrac-mvngu · 2009-07-23T08:53:59Z

comment:14

Replying to @robertwb:

Looks like it'll have to be rebased, worked fine on vanilla 4.1. Where's the .tar on sage.math?

Source and binary versions are up at

http://sage.math.washington.edu/home/mvngu/release/sage-4.1.1.alpha0.tar

http://sage.math.washington.edu/home/mvngu/release/sage-4.1.1.alpha0-sage.math.washignton.edu-x86_64-Linux.tar.gz

robertwb · 2009-07-25T12:07:43Z

comment:15

Works fine form me. sage-4.1.1.alpha0-sage.math.washignton.edu-x86_64-Linux.tar.gz + #6438 + #4571. Are you sure applied the latest Cython spkg first? 4.1.1.alpha0 doesn't seem to have it.

sagetrac-mvngu · 2009-07-25T15:45:02Z

Changed reviewer from Robert Bradshaw, Dag Sverre Seljebotn to Robert Bradshaw, Dag Sverre Seljebotn, Minh Van Nguyen

sagetrac-mvngu · 2009-07-25T15:45:02Z

comment:17

dagss: The patch 4571-more-fixes.patch doesn't contain your username. I'm committing it in your name. Merged in this order

the SPKG at Upgrade to Cython 0.11.2 #6438
4571-numpy-pxd.patch
4571-more-fixes.patch

sagetrac-mvngu · 2009-07-25T15:45:02Z

Merged: Sage 4.1.1.alpha1

jasongrout added this to the sage-4.1.1 milestone Nov 20, 2008

jasongrout added c: build labels Nov 20, 2008

jasongrout assigned sagetrac-mabshoff Nov 20, 2008

robertwb added the s: needs work label Jun 28, 2009

robertwb added c: numerical and removed c: build labels Jul 22, 2009

sagetrac-mvngu mannequin added s: needs work and removed s: positive review labels Jul 23, 2009

robertwb added s: needs review and removed s: needs work labels Jul 25, 2009

sagetrac-mvngu mannequin added s: positive review and removed s: needs review labels Jul 25, 2009

sagetrac-mvngu mannequin closed this as completed Jul 25, 2009

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

merge sage's numpy.pxd with the cython numpy.pxd #4571

merge sage's numpy.pxd with the cython numpy.pxd #4571

jasongrout commented Nov 20, 2008

robertwb commented Jun 28, 2009

robertwb commented Jun 28, 2009

robertwb commented Jun 28, 2009

sagetrac-dagss mannequin commented Jul 7, 2009

sagetrac-dagss mannequin commented Jul 7, 2009

sagetrac-dagss mannequin commented Jul 7, 2009

robertwb commented Jul 9, 2009

robertwb commented Jul 21, 2009

robertwb commented Jul 22, 2009

sagetrac-mvngu mannequin commented Jul 22, 2009

sagetrac-mvngu mannequin commented Jul 22, 2009

sagetrac-mvngu mannequin commented Jul 22, 2009

sagetrac-mvngu mannequin commented Jul 23, 2009

robertwb commented Jul 23, 2009

sagetrac-mvngu mannequin commented Jul 23, 2009

robertwb commented Jul 25, 2009

sagetrac-mvngu mannequin commented Jul 25, 2009

sagetrac-mvngu mannequin commented Jul 25, 2009

sagetrac-mvngu mannequin commented Jul 25, 2009

merge sage's numpy.pxd with the cython numpy.pxd #4571

merge sage's numpy.pxd with the cython numpy.pxd #4571

Comments

jasongrout commented Nov 20, 2008

robertwb commented Jun 28, 2009

robertwb commented Jun 28, 2009

robertwb commented Jun 28, 2009

sagetrac-dagss mannequin commented Jul 7, 2009

sagetrac-dagss mannequin commented Jul 7, 2009

sagetrac-dagss mannequin commented Jul 7, 2009

robertwb commented Jul 9, 2009

robertwb commented Jul 21, 2009

robertwb commented Jul 22, 2009

sagetrac-mvngu mannequin commented Jul 22, 2009

sagetrac-mvngu mannequin commented Jul 22, 2009

sagetrac-mvngu mannequin commented Jul 22, 2009

sagetrac-mvngu mannequin commented Jul 23, 2009

robertwb commented Jul 23, 2009

sagetrac-mvngu mannequin commented Jul 23, 2009

robertwb commented Jul 25, 2009

sagetrac-mvngu mannequin commented Jul 25, 2009

sagetrac-mvngu mannequin commented Jul 25, 2009

sagetrac-mvngu mannequin commented Jul 25, 2009