I noticed that, when I use non-ASCII characters in my code or comments, the execution of sage blocks from org-babel was interrupted, emacs complaining of a mispatch in encoding and asking for an encoding. Answering utf-8 solves the problem.

Since this happens every time a non-ASCII-containing block is executed, this becomes a chore (to remain polite...). Could we have this configurable by a custom variable or, failing that, by having utf-8 encoding ?

Thanks for the sage integration with org-babel. It works well for text, code, and tables output, but not for some graphics output types. More specifically I am not able to specify pdf or eps as file output types. I need at least one vector format, even svg would be fine. Alas but to no avail.

Your png example works well. But change that same src block png to pdf or eps, the output generated is not valid postscript file. (see attached pdf for an sample file generated by sage). Is this something that can be fixed with ob-sagemath or a problem with org-babel?

A related question: how can I use the src_block vars inside sage code in that block? For example, :file below?

Try this src_block in any org file to recreate the problem PDF file:

#+begin_src sage  :file image.pdf
g=Graph({1:[2,3], 2:[1,3], 3:[1,2]});g
#+end_src

ai-graph-theory-1.pdf

This issue was observed on Sage 9.0 (packaged for Ubuntu 20.04), using the latest git master branch of ob-sagemath. If a sage
plot command block is written as in

#+begin_src sage :file images/sin.png :exports both
  P = plot(sin(x), (0, 2*pi), figsize=[5, 4]); P
#+end_src

then after C-c C-c event though images/sin.png is generated it is not loaded into the org-file but instead the result shown are blank as in

#+RESULTS

Also the mini buffer shows a message stating no output was produced. However if the same plot block is written as

#+begin_src sage :results file
  P = plot(sin(x), (0, 2*pi), figsize=[5, 4]); P
#+end_src

Then the plot is indeed shown in the org file.

This problem does not occur using an older version of Sage 8.6.6 on Debian/Buster using the same git version of ob-sagemath

Hello, thank you for this package! I installed ob-sagemath and sage-shell-mode, and thankfully it works.

However, the performance is a bit slow on my setting, and also the session cannot be re-used. I wonder if I have missed something.

1

#+BEGIN_SRC sage :session foo
1+1
#+END_SRC

This code block takes 6 seconds to evaluate. Is this normal?

2

#+BEGIN_SRC sage :session foo
1+1+1
#+END_SRC

In the same session, lets say I change my code to 1+1+1. This time it tries to evaluate, but doesn't give an output for 60 seconds. Changing foo to fooo does solve this question. I suspect that in my setting, a session can only be used properly one time.

I would like to be able to open a sage session using C-c C-v C-z in a sage source block.
When I use this shortcut, it currently shows "Error: no org-babel-initiate-session function for sage".

Hi,

it'd be great if ob-sage could support something like :results latex which would insert LaTeX into the org document, something like this:

#+BEGIN_SRC sage :results latex
sage: K.<zeta> = CyclotomicField(8)
sage: OK = K.ring_of_integers()
sage: h = -36*zeta^3 + 44*zeta^2 + 14*zeta + 28
sage: h
#+END_SRC

#+RESULTS:
\(-36 \zeta_{8}^{3} + 44 \zeta_{8}^{2} + 14 \zeta_{8} + 28\)

I know that :results raw almost accomplishes this, but it leaves out the enclosing \(…\)

The file should rename to ob-sage.el to match org-babel load convenstion.

I believe sagemath has completely moved to python3.*, which support print("...") but not print "..." by default. This raises a conflict in the example usage

#+BEGIN_SRC  sage :session foo
  print 'This code will be evaluated in the process buffer *Sage<foo>*.'
#+END_SRC

I think it's better to change it to

#+BEGIN_SRC sage foo.sage
  1+1
#+END_SRC

for the favor of those who are new.

I'm having trouble writing the output of execution to a local file. For example

#+header: :results output file silent
#+header: :session nil
#+header: :file analog-type-I-phase-detector.dat :output-dir data/
#+begin_src sage
import numpy as np
from sage.symbolic.integration.integral import definite_integral

x = var("x")
phi = var("phi")
fref = sin(x)
fsig = sin(x + phi)
for phase in np.linspace(0, 2 * np.pi, num=10):
    avg = definite_integral(fref * fsig(phi=phase), x, 0, 2 * pi)
    print("{:.3f}\t{:.3f}".format(phase, float(avg)))
#+end_src

Does not write the data/analog-type-I-phase-detector.dat file. It simply echos the file link in the minibuffer. A similar python example works fine:

#+begin_src python :results output silent file :file test.txt
print("hello")
#+end_src

Is there a way to write stdout to a local file with sage? thanks.

EDIT:
Something like this works, but it would be nice to be able to use the file header directly instead of delegating to python logic.

#+begin_src sage :results output silent :var cwd=(file-name-directory buffer-file-name)
import numpy as np
from os.path import abspath
from sage.symbolic.integration.integral import definite_integral

fname = cwd + "data/analog-type-I-phase-detector.dat"

x = var("x")
phi = var("phi")
fref = sin(x)
fsig = sin(x - phi)
with open(fname, "w") as f:
    f.write("phi vout\n")
    for phase in np.linspace(0, 2 * np.pi, num=1000):
        avg = definite_integral(fref * fsig(phi=phase), x, 0, 2 * pi)
        f.write("{:.3f}\t{:.3f}\n".format(phase, float(avg)))
#+end_src