Commit c865641f by Ned Batchelder

Merge pull request #23 from edx/ned/move-loncapa-etc-into-sandbox

Ned/move loncapa etc into sandbox
parents c0859cb0 88d3f253
(Originally written by Ike.)
At a high level, the main challenges of checking symbolic math expressions are (1) making sure the expression is mathematically legal, and (2) simplifying the expression for comparison with what is expected.
(1) Generation (and testing) of legal input is done by using MathJax to provide input math in an XML format known as Presentation MathML (PMathML). Such expressions typeset correctly, but may not be mathematically legal, like "5 / (1 = 2)". The PMathML is converted into "Content MathML" (CMathML), which is by definition mathematically legal, using an XSLT 2.0 stylesheet, via a module in SnuggleTeX. CMathML is then converted into a sympy expression. This work is all done in `lms/lib/symmath/formula.py`
(2) Simplifying the expression and checking against what is expected is done by using sympy, and a set of heuristics based on options flags provided by the problem author. For example, the problem author may specify that the expected expression is a matrix, in which case the dimensionality of the input expression is checked. Other options include specifying that the comparison be checked numerically in addition to symbolically. The checking is done in stages, first with no simplification, then with increasing levels of testing; if a match is found at any stage, then an "ok" is returned. Helpful messages are also returned, eg if the input expression is of a different type than the expected. This work is all done in `lms/lib/symmath/symmath_check.py`
Links:
SnuggleTex: http://www2.ph.ed.ac.uk/snuggletex/documentation/overview-and-features.html
MathML: http://www.w3.org/TR/MathML2/overview.html
SymPy: http://sympy.org/en/index.html
......@@ -4,6 +4,7 @@ setup(
name="sandbox-packages",
version="0.1.1",
packages=[
"loncapa",
"verifiers",
],
py_modules=[
......
from setuptools import setup
setup(
name="symmath",
version="0.1",
packages=["symmath"],
install_requires=[
"sympy",
],
)
(Originally written by Ike.)
At a high level, the main challenges of checking symbolic math expressions are
(1) making sure the expression is mathematically legal, and (2) simplifying the
expression for comparison with what is expected.
(1) Generation (and testing) of legal input is done by using MathJax to provide
input math in an XML format known as Presentation MathML (PMathML). Such
expressions typeset correctly, but may not be mathematically legal, like "5 /
(1 = 2)". The PMathML is converted into "Content MathML" (CMathML), which is
by definition mathematically legal, using an XSLT 2.0 stylesheet, via a module
in SnuggleTeX. CMathML is then converted into a sympy expression. This work is
all done in `symmath/formula.py`.
(2) Simplifying the expression and checking against what is expected is done by
using sympy, and a set of heuristics based on options flags provided by the
problem author. For example, the problem author may specify that the expected
expression is a matrix, in which case the dimensionality of the input
expression is checked. Other options include specifying that the comparison be
checked numerically in addition to symbolically. The checking is done in
stages, first with no simplification, then with increasing levels of testing;
if a match is found at any stage, then an "ok" is returned. Helpful messages
are also returned, eg if the input expression is of a different type than the
expected. This work is all done in `symmath/symmath_check.py`.
Links:
SnuggleTex: http://www2.ph.ed.ac.uk/snuggletex/documentation/overview-and-features.html
MathML: http://www.w3.org/TR/MathML2/overview.html
SymPy: http://sympy.org/en/index.html
......@@ -4,3 +4,4 @@
common/lib/calc
common/lib/chem
common/lib/sandbox-packages
common/lib/symmath
......@@ -2,5 +2,6 @@
-e common/lib/calc
-e common/lib/capa
-e common/lib/chem
-e common/lib/symmath
-e common/lib/xmodule
-e .
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment