[sage-support] Re: numeric approximations for symbolic expressions
On Fri, 10 Oct 2008 17:48:57 -0500 Jason Grout [EMAIL PROTECTED] wrote: Burcin Erocal wrote: On Fri, 10 Oct 2008 10:35:16 -0500 Jason Grout [EMAIL PROTECTED] wrote: The general request still stands, though: is there a way to numerically approximate all the constants in a symbolic expression, but keep the variables as variables? The pynac based symbolics code does the following: sage: x,y,z = var(x y z, ns=1) sage: t = 1 + sqrt(2)*x + sin(x) sage: t.n() sin(x) + (1.4142135623730951)*x + 1.0 This is with the bundle at #3872 and patches at #4244 applied, and the package at #4243. Is there any chance this could be merged into 3.1.3 to get wider testing? That is, if it's a purely optional framework. Barring that, is there a chance we could get a single huge patch that consolidates all of this, or even just a single patch that contains the changes in the bundle? Since 3.1.3 is almost out the door, and these tickets have not been reviewed yet, they won't be in 3.1.3. I've been told they are high priority for 3.2, so hopefully they will be in the first alpha. I have some fixes I haven't pushed to trac yet. I will try to set up a canonical location where you can get the latest version of my tree. This might be tomorrow afternoon though. Cheers, Burcin --~--~-~--~~~---~--~~ To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-support URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-support] Re: numeric approximations for symbolic expressions
On Fri, Oct 10, 2008 at 7:11 AM, Jason Grout [EMAIL PROTECTED] wrote: How do I get a numeric approximation for symbolic expressions that have variables? I want to leave the variables alone, but get numeric approximations for all constants. For example, here's how it works in mathematica: In[1]:= a:=1+Sqrt[2]*x In[2]:= a Out[2]= 1 + Sqrt[2] x In[3]:= N[a] Out[3]= 1. + 1.41421 x However, the corresponding thing does not work in Sage: sage: a=1+sqrt(2)*x sage: a sqrt(2)*x + 1 sage: n(a) --- TypeError Traceback (most recent call last) For polynomials, do this: sage: f = 1 + sqrt(2)*x sage: f.polynomial(RDF) 1.41421356237*x + 1.0 --~--~-~--~~~---~--~~ To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-support URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-support] Re: numeric approximations for symbolic expressions
William Stein wrote: On Fri, Oct 10, 2008 at 7:11 AM, Jason Grout [EMAIL PROTECTED] wrote: How do I get a numeric approximation for symbolic expressions that have variables? I want to leave the variables alone, but get numeric approximations for all constants. For example, here's how it works in mathematica: In[1]:= a:=1+Sqrt[2]*x In[2]:= a Out[2]= 1 + Sqrt[2] x In[3]:= N[a] Out[3]= 1. + 1.41421 x However, the corresponding thing does not work in Sage: sage: a=1+sqrt(2)*x sage: a sqrt(2)*x + 1 sage: n(a) --- TypeError Traceback (most recent call last) For polynomials, do this: sage: f = 1 + sqrt(2)*x sage: f.polynomial(RDF) 1.41421356237*x + 1.0 Aah, thanks. I guess I also now see the f.series command, which does what I originally wanted as well (lets me specify the ring for the coefficients to the series approximation). The general request still stands, though: is there a way to numerically approximate all the constants in a symbolic expression, but keep the variables as variables? Thanks, Jason --~--~-~--~~~---~--~~ To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-support URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-support] Re: numeric approximations for symbolic expressions
On Fri, 10 Oct 2008 10:35:16 -0500 Jason Grout [EMAIL PROTECTED] wrote: The general request still stands, though: is there a way to numerically approximate all the constants in a symbolic expression, but keep the variables as variables? The pynac based symbolics code does the following: sage: x,y,z = var(x y z, ns=1) sage: t = 1 + sqrt(2)*x + sin(x) sage: t.n() sin(x) + (1.4142135623730951)*x + 1.0 This is with the bundle at #3872 and patches at #4244 applied, and the package at #4243. Cheers, Burcin --~--~-~--~~~---~--~~ To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-support URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-support] Re: numeric approximations for symbolic expressions
Burcin Erocal wrote: On Fri, 10 Oct 2008 10:35:16 -0500 Jason Grout [EMAIL PROTECTED] wrote: The general request still stands, though: is there a way to numerically approximate all the constants in a symbolic expression, but keep the variables as variables? The pynac based symbolics code does the following: sage: x,y,z = var(x y z, ns=1) sage: t = 1 + sqrt(2)*x + sin(x) sage: t.n() sin(x) + (1.4142135623730951)*x + 1.0 This is with the bundle at #3872 and patches at #4244 applied, and the package at #4243. Nice. That's exactly what I was wanting. Chalk me up as one more person eager to get this in shape and into Sage. Thanks, Jason --~--~-~--~~~---~--~~ To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-support URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-support] Re: numeric approximations for symbolic expressions
On Oct 10, 10:02 am, Jason Grout [EMAIL PROTECTED] wrote: Burcin Erocal wrote: On Fri, 10 Oct 2008 10:35:16 -0500 Jason Grout [EMAIL PROTECTED] wrote: The general request still stands, though: is there a way to numerically approximate all the constants in a symbolic expression, but keep the variables as variables? The pynac based symbolics code does the following: sage: x,y,z = var(x y z, ns=1) sage: t = 1 + sqrt(2)*x + sin(x) sage: t.n() sin(x) + (1.4142135623730951)*x + 1.0 This is with the bundle at #3872 and patches at #4244 applied, and the package at #4243. Nice. That's exactly what I was wanting. Chalk me up as one more person eager to get this in shape and into Sage. Yep, it is certainly the highest priority ticket for 3.2 and 3.1.3 has unfortunately delayed longer than planned. I plan to do 3.1.3.rc0 tonight and unless something major happens it should be very close to final. Then we can open 3.2 during the coding sprint at SD 10 here at Loria. Thanks, Jason Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-support URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-support] Re: numeric approximations for symbolic expressions
Burcin Erocal wrote: On Fri, 10 Oct 2008 10:35:16 -0500 Jason Grout [EMAIL PROTECTED] wrote: The general request still stands, though: is there a way to numerically approximate all the constants in a symbolic expression, but keep the variables as variables? The pynac based symbolics code does the following: sage: x,y,z = var(x y z, ns=1) sage: t = 1 + sqrt(2)*x + sin(x) sage: t.n() sin(x) + (1.4142135623730951)*x + 1.0 This is with the bundle at #3872 and patches at #4244 applied, and the package at #4243. Is there any chance this could be merged into 3.1.3 to get wider testing? That is, if it's a purely optional framework. Barring that, is there a chance we could get a single huge patch that consolidates all of this, or even just a single patch that contains the changes in the bundle? Thanks, Jason --~--~-~--~~~---~--~~ To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-support URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-support] Re: numeric approximations for symbolic expressions
On Oct 10, 3:48 pm, Jason Grout [EMAIL PROTECTED] wrote: SNIP Is there any chance this could be merged into 3.1.3 to get wider testing? That is, if it's a purely optional framework. Barring that, is there a chance we could get a single huge patch that consolidates all of this, or even just a single patch that contains the changes in the bundle? Nope, I don't see this happening until I a) valgrind the new doctests b) build it on Cygwin, Solaris, Linux/Itanium and G4/G5 on OSX 10.4 and 10.5 and have all doctests pass and the doctests pass. ghmm has taught me that just because it mostly works and there is only a couple allegedly easy bugs to fix a patch like that can still hold up a release for an extended amount of time (2 weeks in 3.1.2 in my guesstimation). And those bugs in ghmm on Itanium are still not fixed, but we did disable doctests globally in those two files in order to make the doctests pass. 3.1.3 is supposed to be out soon, i.e. before the SD 10 coding sprint, and I am not willing to take the risk of pynac's integration bundle. Thanks, Jason Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-support URLs: http://www.sagemath.org -~--~~~~--~~--~--~---