Ticket #3103 (closed defect: fixed)

Opened 5 years ago

Last modified 4 months ago

Coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2

Reported by: ncalexan Owned by: was
Priority: major Milestone: sage-5.7
Component: linear algebra Keywords: vector CDF coerce
Cc: ncalexan, rbradshaw, jason Work issues:
Report Upstream: N/A Reviewers: Robert Bradshaw
Authors: Charles Bouillaguet Merged in: sage-5.7.beta1
Dependencies: Stopgaps:

Description

With

sage: version()
'SAGE Version 3.0.1.alpha0, Release Date: 2008-04-26'

I get the following coercion errors:

sage: vector(CDF, [2, 2]) * vector(ZZ, [1, 3])
---------------------------------------------------------------------------
<type 'exceptions.TypeError'>             Traceback (most recent call last)

/Users/ncalexan/sage-3.0.1.alpha0/devel/sage-nca/sage/rings/<ipython console> in <module>()

/Users/ncalexan/sage-3.0.1.alpha0/devel/sage-nca/sage/rings/element.pyx in sage.structure.element.Vector.__mul__ (sage/structure/element.c:10413)()

/Users/ncalexan/sage-3.0.1.alpha0/devel/sage-nca/sage/rings/coerce.pyx in sage.structure.coerce.CoercionModel_cache_maps.bin_op_c (sage/structure/coerce.c:5292)()

<type 'exceptions.TypeError'>: unsupported operand parent(s) for '*': 'Vector space of dimension 2 over Complex Double Field' and 'Ambient free module of rank 2 over the principal ideal domain Integer Ring'
sage: vector(CDF, [2, 2]) * vector(QQ, [1, 3])
---------------------------------------------------------------------------
<type 'exceptions.TypeError'>             Traceback (most recent call last)

/Users/ncalexan/sage-3.0.1.alpha0/devel/sage-nca/sage/rings/<ipython console> in <module>()

/Users/ncalexan/sage-3.0.1.alpha0/devel/sage-nca/sage/rings/element.pyx in sage.structure.element.Vector.__mul__ (sage/structure/element.c:10413)()

/Users/ncalexan/sage-3.0.1.alpha0/devel/sage-nca/sage/rings/coerce.pyx in sage.structure.coerce.CoercionModel_cache_maps.bin_op_c (sage/structure/coerce.c:5292)()

<type 'exceptions.TypeError'>: unsupported operand parent(s) for '*': 'Vector space of dimension 2 over Complex Double Field' and 'Vector space of dimension 2 over Rational Field'

Attachments

trac_3103_add_doctest.patch Download (911 bytes) - added by Bouillaguet 4 months ago.

Change History

comment:1 Changed 5 years ago by mabshoff

  • Milestone set to sage-3.0.2

comment:2 Changed 5 years ago by robertwb

This is not due to coercion, but rather that dot product is not implemented for CDFn.

sage: (CDF^2)([1,2]) * (CDF^2)([2,3])
------------------------------------------------------------
Traceback (most recent call last):
  File "<ipython console>", line 1, in <module>
  File "element.pyx", line 1884, in sage.structure.element.Vector.__mul__ (sage/structure/element.c:10398)
  File "element.pyx", line 1898, in sage.structure.element.Vector._dot_product_c (sage/structure/element.c:10564)
  File "element.pyx", line 1904, in sage.structure.element.Vector._dot_product_c_impl (sage/structure/element.c:10649)
<type 'exceptions.TypeError'>: unsupported operand parent(s) for '*': 'Vector space of dimension 2 over Complex Double Field' and 'Vector space of dimension 2 over Complex Double Field'

Which, looking at element.pyx line 1904 shows that _dot_product_c_impl was never implemented for CDF. This should almost certainly be a NotImplementedError?, or something like that, not a TypeError?. In fact, there should just be a generic implementation.

comment:3 Changed 5 years ago by ncalexan

  • Cc rbradshaw added
  • Summary changed from coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2 to [with patch, needs review] coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2

The attached patch fixes the errors (they were in fact dot_product errors and not coercion) and adds doctests and functionality for preserving inner product matrices.

comment:4 Changed 5 years ago by robertwb

  • Summary changed from [with patch, needs review] coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2 to [with patch, positive review] coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2

The patch solves the problem and cleans up a lot of old code.

comment:5 Changed 5 years ago by mabshoff

  • Summary changed from [with patch, positive review] coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2 to [with patch, positive review, needs works] coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2

This patch causes a number of doctest failures:

        sage -t -long devel/sage/sage/modular/modsym/space.py # 5 doctests failed
        sage -t -long devel/sage/sage/modular/modform/constructor.py # 3 doctests failed
        sage -t -long devel/sage/sage/modular/hecke/module.py # 12 doctests failed
        sage -t -long devel/sage/sage/modular/abvar/abvar_newform.py # 19 doctests failed
        sage -t -long devel/sage/sage/modular/abvar/abvar_ambient_jacobian.py # 2 doctests failed
        sage -t -long devel/sage/sage/modular/abvar/abvar.py # 3 doctests failed
        sage -t -long devel/sage/sage/modular/abvar/homspace.py # 2 doctests failed

Cheers,

Michael

comment:6 Changed 5 years ago by mabshoff

  • Summary changed from [with patch, positive review, needs works] coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2 to [with patch, needs works] coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2

comment:7 Changed 5 years ago by jason

  • Cc jason added

comment:8 Changed 4 years ago by AlexGhitza

  • Summary changed from [with patch, needs works] coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2 to [with patch, needs work] coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2

comment:9 Changed 4 years ago by jason

In Sage 3.4.1.rc2, this works:

----------------------------------------------------------------------
| Sage Version 3.4.1.rc2, Release Date: 2009-04-10                   |
| Type notebook() for the GUI, and license() for information.        |
----------------------------------------------------------------------
sage: vector(CDF, [2, 2]) * vector(ZZ, [1, 3])
8.0

So this sounds like it is fixed. However, are there nice doctests from the above patch that could be added to the current code? Nick?

comment:10 Changed 4 years ago by mabshoff

Yes, I agree with Jason. Once we add a patch that adds a doctest this ticket can be merged and closed. By maybe someone ought to check out if Nick's patch contains anything else.

Cheers,

Michael

comment:11 Changed 4 months ago by Bouillaguet

  • Status changed from needs_work to needs_review
  • Report Upstream set to N/A
  • Authors set to Charles Bouillaguet

Since the problem has been fixed, here is a very simple doctest that enforces it won't happen again.

Patchbot !

apply trac_3103_add_doctest.patch

Changed 4 months ago by Bouillaguet

comment:12 Changed 4 months ago by Bouillaguet

took care of stupid TAB in the patch

comment:13 Changed 4 months ago by robertwb

Apply only trac_3103_add_doctest.patch

comment:14 Changed 4 months ago by robertwb

  • Status changed from needs_review to positive_review

comment:15 Changed 4 months ago by jdemeyer

  • Reviewers set to Robert Bradshaw
  • Summary changed from [with patch, needs work] coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2 to Coercion errors of vectors from ZZ^2 and QQ^2 into CDF^2
  • Milestone changed from sage-5.6 to sage-5.7

comment:16 Changed 4 months ago by jdemeyer

  • Status changed from positive_review to closed
  • Resolution set to fixed
  • Merged in set to sage-5.7.beta1
Note: See TracTickets for help on using tickets.