Do we even need bound_symbols?
I think it can be useful for checking what can be done with an
expression, for example if it is desired to manipulate an expression
like this:
n*Sum(x**n, (n, 0, oo)) = n*Sum(x**m, (m, 0, oo)) = Sum(n*x**m,
(m, 0, oo))
If a symbol is present in both free and
Do we even need bound_symbols?
I think it can be useful for checking what can be done with an
expression, for example if it is desired to manipulate an expression
like this:
n*Sum(x**n, (n, 0, oo)) = n*Sum(x**m, (m, 0, oo)) = Sum(n*x**m,
(m, 0, oo))
If a symbol is present in both free and
I am really happy to see the fruit of my gsoc landing in the master
branch. Thanks to all who have been involved!
Ondrej, could you rebuild the documentation in
http://certik.github.com/sympy ?
Øyvind
on., 29.09.2010 kl. 11.19 -0700, skrev Ondrej Certik:
Thanks Andy!
Ondrej
On Wed,
ti., 17.08.2010 kl. 01.14 -0600, skrev Aaron S. Meurer:
Well, here's the problem with the spaceless option:
var('ab cd ef')
(ab, cd, ef)
var('ab cd')
(ab, cd)
var('ab')
(a, b)
This one has bitten me many times. It would be great to make it more
consistent.
Øyvind
So definitely
/second_review_round
Øyvind
On 28 Jul, 23:52, Øyvind Jensen jensen.oyv...@gmail.com wrote:
On 27 Jul, 19:07, Brian Granger elliso...@gmail.com wrote:
Oyvind,
I don't know if you saw the new operator priority stuff that I did this
weekend. This will make it possible to more reliably have Expr
Thanks for your comments everybody! When you are finished reviewing,
please remember to push the button in smartbear so that we move on to
the next phase.
Øyvind
On 28 Jul, 08:19, Ondrej Certik ond...@certik.cz wrote:
On Tue, Jul 27, 2010 at 9:11 AM, Andy Ray Terrel andy.ter...@gmail.com
just create an issue and do
it later when things have stabilized.
Øyvind
Cheers,
Brian
On Mon, Jul 26, 2010 at 1:36 PM, Øyvind Jensen jensen.oyv...@gmail.comwrote:
I have uploaded my work on the C code printer, C code generator and the
tensor module to smartbear for review:
http
I have uploaded my work on the C code printer, C code generator and the
tensor module to smartbear for review:
http://hosted.smartbear.com/sympy/go?page=ReviewDisplayreviewid=6
The ability to create code that loops over array arguments is much
improved, mainly because the contraction logic is
Thanks, I'll fix it very soon.
Øyvind
On 9 Jul, 23:51, certik1 cert...@llnl.gov wrote:
Hi,
the new patches about codegen are awesome, as well as Tuple and so on.
Great job!
Ondrej
P.S.
I tried to compile on my rhel5 box and got some easy to fix things for
python2.4.
Hi Toon,
Thanks for your patches/suggestions and for looking over my work!
lø., 10.07.2010 kl. 09.09 +0200, skrev Toon Verstraelen:
Hi,
I've been looking into the fcode and codegen implementation. Sorry for
the delay. First of all: I'm really happy with the work being done. I
wish I
ma., 05.07.2010 kl. 10.08 -0700, skrev Ondrej Certik:
On Mon, Jul 5, 2010 at 5:16 AM, Øyvind Jensen jensen.oyv...@gmail.com wrote:
On 4 Jul, 19:54, Ondrej Certik ond...@certik.cz wrote:
On Sun, Jul 4, 2010 at 10:45 AM, Aaron S. Meurer asmeu...@gmail.com
wrote:
Oops, you are right
to., 01.07.2010 kl. 12.28 -0500, skrev Brian Granger:
On Tue, Jun 29, 2010 at 5:10 PM, Øyvind Jensen jensen.oyv...@gmail.com
wrote:
ti., 29.06.2010 kl. 11.26 -0700, skrev Matt Curry:
Hello,
I've been working on implementing Hilbert spaces in SymPy. I have
created/improved two files
I get some failing tests:
__ /home/oy/git_workingdir/sympy/sympy/physics/tests/
test_hilbert.py ___
File /home/oy/git_workingdir/sympy/sympy/physics/tests/
test_hilbert.py, line 1, in module
from
File /home/oy/git_workingdir/sympy/sympy/physics/hilbert.py, line
267
obj = Expr.__new__(cls, *args, commutative=False)
^
SyntaxError: invalid syntax
The syntax error is because I have python 2.5 installed. The following
line is compatible
on., 30.06.2010 kl. 12.05 -0700, skrev Matt Curry:
Okay, I've fixed up a few more things you've mentioned and pushed it
to github.
Cool. I just finished commenting on github
http://github.com/mattcurry/sympy/commit/8e63d9d62bca475bcd6e6ee711e166cb6fbe
In general it looks very good: It's
ti., 29.06.2010 kl. 11.26 -0700, skrev Matt Curry:
Hello,
I've been working on implementing Hilbert spaces in SymPy. I have
created/improved two files located on my github account:
hilbert.py:
http://github.com/mattcurry/sympy/blob/hilbert/sympy/physics/hilbert.py
test_hilbert.py:
ti., 15.06.2010 kl. 00.04 +0200, skrev Vinzent Steinberg:
What about updating the docstring?
It is updated:
@@ -309,6 +346,8 @@ def fcode(expr, **settings):
some parameter statements for the number symbols. If
False, the same information is
On 9 Jun, 23:03, Andy Ray Terrel andy.ter...@gmail.com wrote:
Instead of creating new branches every time you update your repo I
suggest using stgit [0]. You should be careful though because you can
mess up and overwrite things, but as long as you are the person
pulling and pushing to your
I think I could make good use of SymTuples for code generation with
array arguments, and I believe we did reach a conclusion above, so I
rebased the patch to current master and created another patch moving
SymTuple into the core. They are both in the branch
This will let the line wrapping function respect settings of the fortran
printer.
---
sympy/printing/fcode.py| 107 ++--
sympy/printing/tests/test_fcode.py |5 +-
2 files changed, 56 insertions(+), 56 deletions(-)
diff --git
Implemented tests of the generated strings, and a script in test_external/ to
test resulting binaries.
The existing CCodeGen and related tests was used as templates for the new
functionality.
---
sympy/test_external/test_codegen_f95_ifort.py | 240
sympy/utilities/codegen.py
fcode takes a new keyword: source_format='fixed'|'free'
---
sympy/printing/fcode.py| 71
sympy/printing/tests/test_fcode.py |6 ++--
2 files changed, 58 insertions(+), 19 deletions(-)
diff --git a/sympy/printing/fcode.py
ti., 11.05.2010 kl. 02.36 -0700, skrev smichr:
As I already said, it would be much easier if you could create another
branch for review (1766-for-review) which contains only stable changes.
This would help to get it in faster.
The commits before 1766-- are stable. I
I pushed it to the branch fix_latex
ti., 04.05.2010 kl. 15.09 -0700, skrev Ondrej Certik:
+1, is it in your branch somewhere?
Ondrej
OnMon, May 3, 2010 at 3:32 AM, Øyvind Jensen jensen.oyv...@gmail.com wrote:
Before: -1^{x}
Now: \left(-1\right)^{x}
Added a test
Well, Jeger means hunter and it is more of a joke, really... From the
good old days ;-)
on., 05.05.2010 kl. 13.02 -0700, skrev Ondrej Certik:
On Wed, May 5, 2010 at 12:40 PM, Øyvind Jensen jensen.oyv...@gmail.com
wrote:
I cannot figure out how to fork that repo. Thought there should
Before: -1^{x}
Now: \left(-1\right)^{x}
Added a test
---
sympy/printing/latex.py|3 ++-
sympy/printing/tests/test_latex.py |3 +++
2 files changed, 5 insertions(+), 1 deletions(-)
diff --git a/sympy/printing/latex.py b/sympy/printing/latex.py
index
SymTuple did not fulfill foo.func(*foo.args) == foo. This patch fixes that
by storing the wrapped tuple as .args[0] instead of directly as .args
Updated all SymTuple methods, added a test and fixed some whitespace.
---
sympy/physics/secondquant.py| 35
Added a test and fixed some related whitespace errors.
---
sympy/physics/secondquant.py| 20 +---
sympy/physics/tests/test_secondquant.py |9 +
2 files changed, 10 insertions(+), 19 deletions(-)
diff --git a/sympy/physics/secondquant.py
., 05.04.2010 kl. 00.06 +0200, skrev Vinzent Steinberg:
2010/3/31 Øyvind Jensen jensen.oyv...@gmail.com:
In situations where it is natural to generate dummy symbols with
Symbol.as_dummy(), the reverse process may also be useful.
Uncommented two tests for this.
---
sympy/core/symbol.py
In situations where it is natural to generate dummy symbols with
Symbol.as_dummy(), the reverse process may also be useful.
Uncommented two tests for this.
---
sympy/core/symbol.py| 11 +++
sympy/core/tests/test_symbol.py |4 ++--
2 files changed, 13 insertions(+), 2
I think he meant to put commutative=False, not noncommutative=True.
Exactly, I put up another branch fix__lt__2, where this is fixed.
It's weird that Symbol will let you create nonexistent assumptions
on it:
But that can be quite useful:
In [4]: r = Symbol('r',something=True)
In [5]:
All comparisons of for non-commutative symbols used to return False. Now
an Inequality or StrictInequality is returned instead.
Added a test.
---
sympy/core/basic.py| 12 +---
sympy/core/tests/test_basic.py |8
2 files changed, 13 insertions(+), 7 deletions(-)
Great suggestions! I'll create a new patch.
Øyvind
ma., 22.02.2010 kl. 14.13 -0700, skrev Aaron S. Meurer:
On Feb 22, 2010, at 12:43 PM, Øyvind Jensen wrote:
Before, 'refine_Pow' would simplify (-1)**exp only if 'exp' as a whole
couold
be determined as even or odd. A doctest like
Before, 'refine_Pow' would simplify (-1)**exp only if 'exp' as a whole could
be determined as even or odd. A doctest like,
refine_Pow((-1)**(x+y), Assume(x, Q.even))
(-1)**y
would fail. This patch fixes that by treating powers of S.NegativeOne as a
special case. Every term in
An exception could be raised as a false positive. We should only
demand consistency between two resolutors if they have both reached a
conclusion. Before this patch, an exception would be raised in the event of an
inconclusive resolutor following a conclusive.
Added a test for the issue
---
Sorry, I sent the wrong patch. Please consider the next email instead.
Øyvind
ti., 23.02.2010 kl. 20.23 +0100, skrev Øyvind Jensen:
An exception could be raised as a false positive. We should only
demand consistency between two resolutors if they have both reached a
conclusion. Before
An exception could be raised as a false positive. We should only
demand consistency between two resolutors if they have both reached a
conclusion. Before this patch, an exception would be raised in the event of an
inconclusive resolutor following a conclusive.
Added a test for the issue
---
Before, 'refine_Pow' would simplify (-1)**exp only if 'exp' as a whole couold
be determined as even or odd. A doctest like,
refine_Pow((-1)**(x+y), Assume(x, Q.even))
(-1)**y
would fail. This patch fixes that by treating powers of S.NegativeOne as
a special case. Every term
sø., 21.02.2010 kl. 13.25 +0100, skrev Vinzent Steinberg:
2010/2/21 Øyvind Jensen jensen.oyv...@gmail.com:
The keyword pretty_indices used to be boolean, but is now a dict of lists
containing letters to label the new dummy indices. This allows the user
much
better control over
---
sympy/physics/secondquant.py | 25 +
1 files changed, 13 insertions(+), 12 deletions(-)
diff --git a/sympy/physics/secondquant.py b/sympy/physics/secondquant.py
index edc9987..687bf9d 100644
--- a/sympy/physics/secondquant.py
+++ b/sympy/physics/secondquant.py
@@
---
sympy/physics/secondquant.py |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/sympy/physics/secondquant.py b/sympy/physics/secondquant.py
index 687bf9d..b7ae493 100644
--- a/sympy/physics/secondquant.py
+++ b/sympy/physics/secondquant.py
@@ -2269,8 +2269,8 @@ def
---
sympy/physics/secondquant.py |6 --
1 files changed, 4 insertions(+), 2 deletions(-)
diff --git a/sympy/physics/secondquant.py b/sympy/physics/secondquant.py
index 858c222..6a5fa60 100644
--- a/sympy/physics/secondquant.py
+++ b/sympy/physics/secondquant.py
@@ -2666,8 +2666,10 @@ def
The keyword pretty_indices used to be boolean, but is now a dict of lists
containing letters to label the new dummy indices. This allows the user much
better control over the resulting expression.
Docstring and doctests updated accordingly.
---
sympy/physics/secondquant.py | 104
Good, thanks.
Øyvind
sø., 06.12.2009 kl. 14.38 +0100, skrev Vinzent Steinberg:
2009/12/5 Øyvind Jensen jensen.oyv...@gmail.com:
Another fix to .jacobian() docstring. Improved the layout a bit. No
change in wording.
Øyvind
I felt free to change your patch a bit and to push
Another fix to .jacobian() docstring. Improved the layout a bit. No
change in wording.
Øyvind
--
You received this message because you are subscribed to the Google Groups
sympy-patches group.
To post to this group, send email to sympy-patc...@googlegroups.com.
To unsubscribe from this group,
, Øyvind Jensen jensen.oyv...@gmail.com
wrote:
The failed assertion XYZ should simply be removed. It fails because
of the platform-dependent hash, and is not relevant any more.
It used to be relevant when I tried to use the relation
[[A,B],C] + [[C,A],B] + [[B,C],A] = 0
The failed assertion XYZ should simply be removed. It fails because
of the platform-dependent hash, and is not relevant any more.
It used to be relevant when I tried to use the relation
[[A,B],C] + [[C,A],B] + [[B,C],A] = 0
to enforce a canonical nesting of commutators. I couldn't
to github, please pull.
Ondrej, I fixed the 'str' variable as well.
Øyvind
On 15 Okt, 18:54, Brian Granger ellisonbg@gmail.com wrote:
On Thu, Oct 15, 2009 at 2:26 AM, Øyvind Jensen
jensen.oyv...@gmail.comwrote:
Brian, thanks for your input!
I will make
Brian, thanks for your input!
I will make that Exception baseclass, and raise those
NotImplementedErrors. I'll do it the coming weekend.
I really like your idea of implementing the particle/hole stuff on a
higher level, and creating algorithms independent of boson/fermi
formalism. That is a
Nice!
Here are patches for the quick and easy stuff. They can be applied to
the branch you posted. I also noticed some other small issues, and they
are corrected in the patches.
4) please use explicit imports in the example. Then it would be clear
that the import:
from
Yes, I am aware of that. Thanks for your patches, I'll try it and see
where the bottleneck is. Btw in your first patch, you seem to have a
bogus line (with kwargs), which is not used anywhere, unless I am
missing something.
Ondrej
That was indeed an error. It is fixed in this new patch.
Thanks for your swift responses, I've got some comments:
* In various places you use the variable str. This is dangerous as
str is a builtin type (a string).
Absolutely, It should be substituted immediately.
* It is a little awkward to have the occupation #'s stored differently
for bosons
52 matches
Mail list logo