In fact, substituting x and y directly into the equation of the curve
to plot, and clearing denominators,
produces something pretty good,IMHO:

implicit_plot(v^2*3*sqrt(1-u^2-v^2)-u^3*9+u*(1-u^2-v^2),(u,-1,1),(v,-1,1))


On Thu, Mar 5, 2020 at 4:51 PM Dima Pasechnik <[email protected]> wrote:
>
> On Thu, Mar 5, 2020 at 2:32 PM Fernando Gouvea <[email protected]> wrote:
> >
> > This works, in the sense that there's no error. One does get a bunch of 
> > extraneous points near the boundary of the disk. It's as if plot_points 
> > were trying to connect the point at (0,1) and the point at (0,-1) along the 
> > circle, even though f_uv is 1 on the circle.
> >
> > Strangely, they occur only on the right hand side (i.e., positive u, not 
> > negative u). I tried setting plot_points to be 500, but the bad points 
> > don't go away. Changing the curve to y^2-x^3+x-1=0 doesn't make them go 
> > away either.
> >
>
> the reason is that implicit_plot attempts to approximate the function
> it assumes continuous, so if it's negative inside, but near, the
> boundary, and positive nearby, but outside, then a fake zero is being
> drawn very close to the boundary.
>
> That's why it should be better to create a plot in polar coordinates
> and then transform it.
>
>
>
> > Fernando
> >
> > On 3/5/2020 8:22 AM, Dima Pasechnik wrote:
> >
> > The easiest way is to use Python functions rather than symbolic ones;
> > define a function that is 1 outside the unit disk, and implicitly plot it.
> >
> > sage: def f_uv(u,v):
> > ....:     if u^2+v^2>=1:
> > ....:         return 1
> > ....:     else:
> > ....:         x=u*sqrt(9/(1-u^2-v^2))
> > ....:         y=v*sqrt(9/(1-u^2-v^2))
> > ....:         return y^2-x^3+x
> > ....: implicit_plot(f_uv,(u,-1,1),(v,-1,1))
> > >
> >>
> >> > On Tue, Mar 3, 2020 at 8:20 PM Fernando Gouvea <[email protected]> 
> >> > wrote:
> >> >
> >> > Here's what I ended up trying, with r=3:
> >> >
> >> > var('x y u v')
> >> > x=u*sqrt(9/(1-u^2-v^2))
> >> > y=v*sqrt(9/(1-u^2-v^2))
> >> > implicit_plot(y^2-x^3+x==0,(u,-1,1),(v,-1,1))
> >> >
> >> > That gives an error:
> >> >
> >> > /opt/sagemath-8.9/local/lib/python2.7/site-packages/sage/ext/interpreters/wrapper_rdf.pyx
> >> >  in sage.ext.interpreters.wrapper_rdf.Wrapper_rdf.__call__ 
> >> > (build/cythonized/sage/ext/interpreters/wrapper_rdf.c:2237)()
> >> >      74         for i from 0 <= i < len(args):
> >> >      75             self._args[i] = args[i]
> >> > ---> 76         return self._domain(interp_rdf(c_args
> >> >      77             , self._constants
> >> >      78             , self._py_constants
> >> >
> >> > ValueError: negative number to a fractional power not real
> >> >
> >> > Is there some way to tell implicit_plot to stay inside u^2+v^2\leq 1? Or 
> >> > to ignore complex values?
> >> >
> >> > I'd just change the limits of u and v to make the rectangle of the
> >> > values you plot in, anyway,
> >> > to well stay inside the unit circle.
> >> >
> >> > The equivalent code seems to give the correct graph in Mathematica.
> >> >
> >> > Fernando
> >> >
> >> > On 2/29/2020 5:29 PM, Fernando Gouvea wrote:
> >> >
> >> > Some years ago in a book review, David Roberts had the idea of plotting 
> >> > an algebraic curve using the transformation  (u,v) = (x,y)/(r2 + x2 + 
> >> > y2)1/2, which transforms the plane into a circle and makes it easy to 
> >> > visualize the projective completion of the curve. You can see some of 
> >> > his plots at 
> >> > https://www.maa.org/press/maa-reviews/rational-algebraic-curves-a-computer-algebra-approach
> >> >
> >> > I’d love to do this kind of plot for my students. Can anyone offer help 
> >> > on how to do it with Sage? (Of course the dream scenario would be to add 
> >> > this option to the plot method for curves...)
> >> >
> >> > I’ve been using implicit_plot for most of my examples, which seems to be 
> >> > equivalent of using C.plot() when C is a curve.
> >> >
> >> > Thanks,
> >> >
> >> > Fernando
> >
> > --
> > =============================================================
> > Fernando Q. Gouvea         http://www.colby.edu/~fqgouvea
> > Carter Professor of Mathematics
> > Dept. of Mathematics and Statistics
> > Colby College
> > 5836 Mayflower Hill
> > Waterville, ME 04901
> >
> > We now face a choice between Christ and nothing, because Christ has
> > claimed everything so that renouncing him can only be nihilism.
> >   -- Peter Leithart
> >
> > --
> > You received this message because you are subscribed to the Google Groups 
> > "sage-support" group.
> > To unsubscribe from this group and stop receiving emails from it, send an 
> > email to [email protected].
> > To view this discussion on the web visit 
> > https://groups.google.com/d/msgid/sage-support/bc81b0c4-62ff-65b9-a7f2-e995f26f9234%40colby.edu.

-- 
You received this message because you are subscribed to the Google Groups 
"sage-support" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-support/CAAWYfq1hhty2-zM5mnPcznkjN2HUVyE6biN6kxjYP88w11Tgkg%40mail.gmail.com.

Reply via email to