This should be an easy-to-do task. And also you'll learn a lot while doing it!

(No comment was entered for this change.)

Here's a start.

	test_wester.py
	11.2 KB Download

(No comment was entered for this change.)

Here is what I get:

$ bin/test test_wester.py 
============================= test process starts ==============================
executable:   /usr/bin/python  (2.5.2-final-0)

test_wester.py[84] ...............ffffffffffffffffffffffffffffffXfffffffffffffff
ffffff.................                                                     [OK]

________________________________ xpassed tests _________________________________
test_wester.py:'XPASS: test_H4'

====== tests finished: 32 passed, 51 xfailed, 1 xpassed in 34.91 seconds =======


so test_H4 actually passes. Thanks for the work, great job!

I am thinking where to put the test, any ideas?

I'm developing it more here:

http://github.com/rkern/sympy-rkern/blob/HEAD/sympy/core/tests/test_wester.py

test_H4 technically passes, but per the comment there, it's purely an accident.

Also note that adding this file to the test suite appears to be causing problems with
other tests. See issue #1264.

K4, K5: expand(complex=True) should work. x and y need to be declared real.

Also D4 works (ceil should be ceiling)

@fredrik.johansson: Fixed, thank you.

I rebased your commits, available here:

http://github.com/certik/sympy/tree/wester

Currently this is the result of the tests for me:

$ bin/test sympy/core/tests/test_wester.py 
============================= test process starts ==============================
executable:   /usr/bin/python  (2.5.2-final-0)

sympy/core/tests/test_wester.py[115] .F............ffffffffffffffsffffffffffffff
fffffffffsffffffffffffffffffffffffffffff............s.s......E.s.....F.F  [FAIL]

________________________________________________________________________________
___________________ sympy/core/tests/test_wester.py:test_J6 ____________________
  File "/home/ondra/repos/sympy/sympy/core/tests/test_wester.py", line 466, in test_J6
    assert mpmath.jn(2, 1+1j).ae(mpc('0.04157988694396212', '0.24739764151330632'))
AttributeError: 'module' object has no attribute 'jn'

________________________________________________________________________________
___________________ sympy/core/tests/test_wester.py:test_C2 ____________________
  File "/home/ondra/repos/sympy/sympy/core/tests/test_wester.py", line 39, in test_C2
    (41, 1), (43, 1), (47, 1)])
AssertionError
________________________________________________________________________________
___________________ sympy/core/tests/test_wester.py:test_K8 ____________________
  File "/home/ondra/repos/sympy/sympy/core/tests/test_wester.py", line 555, in test_K8
    assert simplify(sqrt(1/z) - 1/sqrt(z)) != 0
AssertionError
________________________________________________________________________________
___________________ sympy/core/tests/test_wester.py:test_K10 ___________________
  File "/home/ondra/repos/sympy/sympy/core/tests/test_wester.py", line 565, in test_K10
    assert simplify(sqrt(1/z) + 1/sqrt(z)) == 0
AssertionError

 tests finished: 39 passed, 3 failed, 5 skipped, 67 xfailed, 1 exceptions in 5.11
seconds 
DO *NOT* COMMIT!


Plus there are two more fixes in your repository, that should be polished & merged.

(No comment was entered for this change.)

Ondrej, I've started adding tests to the file in your git hub, but I couldn't figure out how to make git download 
the wester branch, so I downloaded it manually (still quite new to git here).  How can I push my changes?  I went 
ahead and created my own git hub account if that helps.  

Thanks! To push your changes, just do:

git push repository branch_name

where branch_name is the name of your branch and repository is the name of your
repository, if you create a new repository at github, it tells you everything.

I have added the L tests here: http://github.com/asmeurer/sympy/blob/4fe5c2cbffe43abafa1238ceb7a0bb597b59d201/sympy/core/tests/te
st_wester.py.  I will probably be adding more over the weekend.  Sorry about the delay.  It took me longer to 
figure out how to use git than it did to write the tests :)

Thanks, looks good. Yes, but you will not regret learning git. :)

I'm having trouble with test M3: solve(x**4+x**3+x**2+x+1,x)
Plugging this expression in to SymPy generates a humongous expression.  With my terminal fullscreen (238x58), the 
solution list fills my whole screen with pretty printing turned off, and several screens with it turned on (obviously, I refrain 
from pasting it here).  Running
l = []
s = solve(x**4+x**3+x**2+x+1,x)
for i in s:
    l.append(simplify(i))

reduces the size of the output to about half of my screen, but it is still huge.  I ran the test command in Maple, and got 
f:=solve(x**4+x**3+x**2+x+1,x);
f:=-1/4+(1/4)*sqrt(5)+(1/4*I)*sqrt(2)*sqrt(5+sqrt(5)), -1/4-(1/4)*sqrt(5)+(1/4*I)*sqrt(2)*sqrt(5-sqrt(5)), -1/4-
(1/4)*sqrt(5)-(1/4*I)*sqrt(2)*sqrt(5-sqrt(5)), -1/4+(1/4)*sqrt(5)-(1/4*I)*sqrt(2)*sqrt(5+sqrt(5))
I had no luck getting even Maple to simplify the SymPy expression down to this, though that could just be because I am 
relatively new to Maple.  I was able to determine that they are the same solutions by numerically evaluating them both.  

Test M4 is verify a solution of M3, but 
for i in s:
    print simplify((x**4+x**3+x**2+x+1).subs({x:i}))

raises a maximum recursion error.  Removing the simplify just makes each print fill the whole screen (they should all be 0).  

Any idea how to make this return a more simplified expression?  Should I report this as a bug?  The solution SymPy gives is 
technically correct (I think), but I hardly feel that this should write an assert line in the test_wester.py that is several hundred 
characters long when it could be much simpler.  

This is definitely a bug.  I would separate it out from this issue and work on it
there.  Simplify needs help.

I won't judge if

 solve(x**4+x**3+x**2+x+1,x)

gives correct result or not, because it's only useful for testing pretty-printing
engine. solve(), in fact roots(), should recognize that the input is a cyclotomic
polynomial and return simple result (currently quartic formula is used and is
responsible for this mess). I will sent a patch fixing this issue.

(No comment was entered for this change.)