Dear Jeroen,
Le lundi 13 août 2018 21:01:37 UTC+2, Jeroen Demeyer a écrit :
>
> On 2018-08-13 19:07, Emmanuel Charpentier wrote:
> > So the question is : how can one *replace* the Python REPL with Sage's
> ?
>
> This seems like an instance of the XY problem: I don't think that
> "replacing" REPLs is the correct way to deal with this. Can't you
> configure/patch reticulate to use Sage instead instead of Python?
>
That's exactly what I'm trying to do.
Before delving in reticulate surgery, I'm trying to use what it offers. I
notice :
- use_python : path of the python binary. I set that to the path of my
main sage *script* which sets correctly the various environment variables...
- repl_python : invokes the repl. Accepts an optional argument which is
the name of a module imported before the REPL is launched.
So my tentative plan is either :
- write a module that imports IPython and launches
IPython.start_ipython(argv=['-c' '%load ext sage'])
This can be emulated by running this interactively with py_run_string, and
gives *partal* results, as illustrated below in a session of Sage's R
interpreter (with comments) :
> library(reticulate)
> use_python("/usr/local/sage/sage", required=TRUE)
# No error : this has been accepted as a Python interpreter
> py_run_string("import IPython ; IPython.start_ipython(argv=['-c',
'%load_ext sage'])")
# Again no error. And, by the way :
> py_run_string("print arctan(x).diff(x)")
1/(x^2 + 1)
# Sage is present, and gives sensible results to sensible inputs.
# Now, can we communicate between Sage and R ?
> py_run_string("foo=arctan(x).integrate(x)")
> py$foo
x*arctan(x) - 1/2*log(x^2 + 1)
We can fetch in R the value of a Sage variable
> paste("*** ",py_to_r(py$latex(py$foo))," ***")
[1] "*** x \\arctan\\left(x\\right) - \\frac{1}{2} \\, \\log\\left(x^{2} +
1\\right) ***"
This result can be "sensibly" converted to R and used here (as a string. It
would be interesting to try to create an R algebraic expression... But this
supposes that we clear the Sage/reticulate nameclash. Later...
But there's a (large) fly in that ointment :
> py_run_string("print 2^3")
1
> py_run_string("print 2**3")
8
The preparser is out to lunch.. And this has consequences...
> py_run_string("print (x^2).diff(x)")
Error in py_run_string_impl(code, local, convert) :
RuntimeError: Use ** for exponentiation, not '^', which means xor
in Python, and has the wrong precedence.
Detailed traceback:
File "<string>", line 1, in <module>
File "sage/structure/element.pyx", line 955, in
sage.structure.element.Element.__xor__
(build/cythonized/sage/structure/element.c:9013)
raise RuntimeError("Use ** for exponentiation, not '^', which means
xor\n"+\
# Sage gives us expressions that it won't read.
# But the problem appears to be limited to the preparser :
> py_run_string("print (x**2).diff(x)")
2*x
- Write a module that wil import * from sage.all and somehow starts a
REPL using the preparser.
This I can't even emulate well In a new session of Sage's R :
> library(reticulate)
> use_python("/usr/local/sage/sage")
> py_run_string("from sage.all import *")
> py_run_string("preparser(True)")
> py_run_string("print 2^3")
1
# Preparser out to lunch again...
> py_run_string("print sin(x).diff(x)")
Error in py_run_string_impl(code, local, convert) :
NameError: name 'x' is not defined
Detailed traceback:
File "<string>", line 1, in <module>
# x has not been defined as the default symbolic variable. This we can fix :
> py_run_string("x=var('x')")
> py_run_string("print sin(x).diff(x)")
cos(x)
Again, reasonable results of reasonable inputs.
Let's try the REPL :
> repl_python()
Python 2.7.15 (/usr/local/sage/sage)
Reticulate 1.10 REPL -- A Python interpreter in R.
This banner can be suppressed if necessary.
>>> 2^3
1
No preparser, again...
>>> tan(x).integrate(x).trig_simplify()
-log(cos(x))
Reasonable. Let's check that we can access that from R :
>>> bar=tan(x).integrate(x).trig_simplify()
>>> exit
> py$bar
-log(cos(x))
Seems so...
So that second solution, seemingly lighter, seems promising.
But I have two obstacles
1. What I know I don't know : how to insert the preparser ?
2. What I don't know I don't know : what am I forgetting ?
Ideas ? Pointers ?
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
