Crazy chemical equations

Question

You should get string of chemical equation (no spaces, only letters (upper-case and lower-case), numbers, brackets and math signs) from user and print the answer if equation is balanced or not (any pair of positive/negative answers: Yes/No, true/false, 1/0). To make code shorter you can assume that input strings can contain only these elements: Al, Ar, B, Be, C, Cl, Cr, Cu, Fe, H, He, K, N, O, S. And one more thing: there could be - signs. It's all about math: + means addition, - means subtraction.

Examples:

Input:

C6H5COOH-O2=7CO2+3H2O

Output:

No

---

Input:

2Fe(CN)6+2SO2+202=Fe2(SO4)2+6C2N2

Output:

Yes

---

Input:

2SO2=2SO4-2O2

Output:

Yes

The shortest code wins.

Answer 1

Mathematica 152

f=Times@@@Tr@CoefficientRules@ToExpression@(r=StringReplace)[r[
#<>")",{"="->"-(",x:_?LetterQ|")"~~y_?DigitQ:>x<>"^"<>y}],x_?UpperCaseQ:>" "<>x]⋃{}=={0}&

Result:

f@"C6H5COOH-O2=7CO2+3H2O"
f@"2Fe(CN)6+2SO2+2O2=Fe2(SO4)2+6C2N2"
f@"2SO2=2SO4-2O2"

False

True

True

I treat the chemical formula as a polynomial, e.g.

Then I just count the coefficients.

Answer 2

Python 2.7, 316 276 chars

import re
r=re.sub
x='(^|[=+-])'
y=r'\1+\2'
z='(\w)([A-Z(])'
t=r('=','==',r(z,y,r(z,y,r('([A-Za-z)])(\d)',r'\1*\2',r('([=+-]|$)',r')\1',r(x+'(\d+)',r'\1\2*(',r(x+'([A-Z])',r'\1(\2',raw_input())))))))
E=re.findall('[A-Za-z]+',t)
print all(eval(t,{f:f==e for f in E})for e in E)

It does a lot of regex-rewriting to convert the input equation into something evalable. Then it checks the equation for each element individually.

For instance, the example equations rewrite to (the t variable):

(C*6+H*5+C+O+O+H)-(O*2)==7*(C+O*2)+3*(H*2+O)
2*(Fe+(C+N)*6)+2*(S+O*2)+2*(O*2)==(Fe*2+(S+O*4)*2)+6*(C*2+N*2)
2*(S+O*2)==2*(S+O*4)-2*(O*2)

I'm sure there's more golf to be had on the regex part.

Answer 3

Haskell, 400 351 308 characters

import Data.List
import Text.Parsec
(&)=fmap
r=return
s=string
l f b=(>>=(&b).f)
x=(=<<).replicate
m=sort&chainl1(l x(concat&many1(l(flip x)n i))n)((s"+">>r(++))<|>(s"-">>r(\\)))
i=between(s"(")(s")")m<|>(:[])&(l(:)(many lower)upper)
n=option 1$read&many1 digit
main=getContents>>=parseTest(l(==)(s"=">>m)m)

This just might have all the golf squeezed out of it. I don't know if there is another 100 51 8 characters to be saved!

& echo 'C6H5COOH-O2=7CO2+3H2O' | runhaskell Chemical.hs
False

& echo '2Fe(CN)6+2SO2+2O2=Fe2(SO4)2+6C2N2' | runhaskell Chemical.hs
True

& echo '2SO2=2SO4-2O2' | runhaskell Chemical.hs
True

Here's the ungolf'd version, in case anyone want to follow along. It is a simple Parsec based parser:

import Control.Applicative ((<$>), (<*>))
import Data.List
import Text.Parsec
import Text.Parsec.String (Parser)

type Atom = String

{- golf'd as x -}
multiple :: Int -> [Atom] -> [Atom]
multiple n = concat . replicate n

{- golf'd as m -}
chemicals :: Parser [Atom]
chemicals = sort <$> chainl1 molecules op
  where
    op :: Eq a => Parser ([a] -> [a] -> [a])
    op = (char '+' >> return (++))
     <|> (char '-' >> return (\\))

    molecules :: Parser [Atom]
    molecules = multiple <$> number <*> terms

    terms :: Parser [Atom]
    terms = concat <$> many1 term

    term :: Parser [Atom]
    term = flip multiple <$> item <*> number

{- gofl'd as i -}
item :: Parser [Atom]
item = between (char '(') (char ')') chemicals
   <|> (:[]) <$> atom
  where
    atom :: Parser Atom
    atom = (:) <$> upper <*> many lower

{- gofl'd as n -}
number :: Parser Int
number = option 1 $ read <$> many1 digit

{- gofl'd as main -}
main :: IO ()
main = getContents >>= parseTest chemEquality
  where
    chemEquality :: Parser Bool
    chemEquality = (==) <$> chemicals <*> (char '=' >> chemicals)