overloading rational add

[TMS]

I have overloaded the basic addition operator, adding 2 fractions, but now I'm having trouble adding a fraction with a non fraction, ie: 3/4 + 1.

Here is part of my program:

def gcd(a, b):
    while b:
        a, b = b, a%b
    return abs(a)
 
class Rational:
    """
    function to initialize variables
    """
    def __init__(self, num, den = 1):
        assert den != 0         #make sure denominator isn't 0
        self.n = num
        self.d = den
        reduced = gcd(num, den) # reduce fraction
        self.n = num/reduced
        self.d= den/reduced
    """
    function to print answer as a string, strips trailing L on longs)
    """
    def __repr__(self):
        return '(%d/%d)' % (self.n, self.d)
    """
    function to overload addition operator to add 2 fractions
    """
    def __add__(self, terms):
        return Rational(self.n*terms.d + terms.n*self.d, self.d*terms.d)
 
 

This is the error I get:
Traceback (most recent call last):
File "C:\Python25\Rational.py", line 87, in <module>
print " a + 1:", a + 1
File "C:\Python25\Rational.py", line 38, in __add__
return Rational(self.n*terms.d + terms.n*self.d, self.d*terms.d)
AttributeError: 'int' object has no attribute 'd'

So, I figured what I need to do is give the whole number a denominator, and that didn't work. Can anyone walk me through this?

Thank you !

TMS

[bartonc]

I have overloaded the basic addition operator, adding 2 fractions, but now I'm having trouble adding a fraction with a non fraction, ie: 3/4 + 1.

This is the error I get:
Traceback (most recent call last):
File "C:\Python25\Rational.py", line 87, in <module>
print " a + 1:", a + 1
File "C:\Python25\Rational.py", line 38, in __add__
return Rational(self.n*terms.d + terms.n*self.d, self.d*terms.d)
AttributeError: 'int' object has no attribute 'd'

So, I figured what I need to do is give the whole number a denominator, and that didn't work. Can anyone walk me through this?

Thank you !

TMS

What you need to do is check the type of the argument in question and either add differently if it's not isinstance(arg, Rational) or convert it to a Rational type object before you add. The former is probably the way to go.

def gcd(a, b):
    while b:
        a, b = b, a%b
    return abs(a)
 
class Rational:
    def __init__(self, num, den = 1):
    """
    function to initialize variables
    """
        assert den != 0         #make sure denominator isn't 0
        self.n = num
        self.d = den
        reduced = gcd(num, den) # reduce fraction
        self.n = num/reduced
        self.d= den/reduced
    def __repr__(self):
    """
    function to print answer as a string, strips trailing L on longs)
    """
        return '(%d/%d)' % (self.n, self.d)
    def __add__(self, terms):
    """
    function to overload addition operator to add 2 fractions
    """
        if isinstance(terms, Rational):
            return Rational(self.n*terms.d + terms.n*self.d, self.d*terms.d)
        else:
            pass    # replace with your code

Also, put the docstrings inside the def so that python's documentation features can find them.

[TMS]

OK, cool. I will work on that. TY

[TMS]

[font=Verdana][size=2]THANK you! I was working on that and couldn't figure it out. The instance thing helped because it stopped it from going back through the __init__.... Hadn't heard of the instance thing. Thanks![/size][/font]

[bartonc]

OK, cool. I will work on that. TY

Ya know what's really cool? When you post back with working code so others can make use of the knowledge that you have gained.

[TMS]

kk, lol, here it is:

 def __add__(self, terms):
        """
        function to overload addition operator to add 2 fractions or a fraction
            and a whole number
        """
        if isinstance(terms, Rational):
            return Rational(self.n*terms.d + terms.n*self.d, self.d*terms.d)
        else:
            terms = Rational(terms, 1)
            return Rational(self.n*terms.d + terms.n*self.d, self.d*terms.d)
 

works great.

I have another question regarding floats. If I pass a float, I get an error (obviously), so I tried to fix it by defining a float function:

def __float__(self):
        return float(self.n)
 

Silly me, thought I could get away that easy. But no, I get an error that looks like this when I try to pass Rational(1.3) with a default denominator of 1:


Traceback (most recent call last):
File "C:\Python25\Rational.py", line 106, in <module>
print " c:", c
File "C:\Python25\Rational.py", line 34, in __repr__
return '(%d/%d)' % (self.n, self.d)
TypeError: int argument required

Any hints?

[bartonc]

I have another question regarding floats. If I pass a float, I get an error (obviously), so I tried to fix it by defining a float function:

def __float__(self):
        return float(self.n)
 

Silly me, thought I could get away that easy. But no, I get an error that looks like this when I try to pass Rational(1.3) with a default denominator of 1:


Traceback (most recent call last):
File "C:\Python25\Rational.py", line 106, in <module>
print " c:", c
File "C:\Python25\Rational.py", line 34, in __repr__
return '(%d/%d)' % (self.n, self.d)
TypeError: int argument required

Any hints?

The hint is in the error msg:
"return '(%d/%d)' % (self.n, self.d)
TypeError: int argument required"
To format floats you need

>>> a = 10.0
>>> floatStr = '%.3f' %a
>>> print floatStr
10.000
>>> 

[bvdet]

kk, lol, here it is:

 def __add__(self, terms):
        """
        function to overload addition operator to add 2 fractions or a fraction
            and a whole number
        """
        if isinstance(terms, Rational):
            return Rational(self.n*terms.d + terms.n*self.d, self.d*terms.d)
        else:
            terms = Rational(terms, 1)
            return Rational(self.n*terms.d + terms.n*self.d, self.d*terms.d)
 

works great.

I have another question regarding floats. If I pass a float, I get an error (obviously), so I tried to fix it by defining a float function:

def __float__(self):
        return float(self.n)
 

Silly me, thought I could get away that easy. But no, I get an error that looks like this when I try to pass Rational(1.3) with a default denominator of 1:


Traceback (most recent call last):
File "C:\Python25\Rational.py", line 106, in <module>
print " c:", c
File "C:\Python25\Rational.py", line 34, in __repr__
return '(%d/%d)' % (self.n, self.d)
TypeError: int argument required

Any hints?

If you want to enable the input of floats, this will work for floats up to 4 decimal places:

def __init__(self, num, den=1):
        if isinstance(num, float) or isinstance(den, float):
            num, den = map(lambda x: x * 1000, [num, den])
            num, den = map(int, [num, den])
 
        assert den != 0
        reduced = gcd(num, den)
        self.n = num/reduced
        self.d = den/reduced
 
>>> a = Rational(1.33, 5.13)
>>> print a
(7/27)
>>> 

[bvdet]

TMS,

Sometimes I create __str__ and __repr__ methods for classes. The __str__ method will execute when the instance is printed. The __repr__ method will recreate the object when passed to eval(). Your project:

>>> a = Rational(3,4)
>>> a
Rational(3, 4)
>>> print a
(3/4)
>>> z = repr(a)
>>> z
'Rational(3, 4)'
>>> z = eval(repr(a))
>>> z
Rational(3, 4)
>>> print z
(3/4)

[bvdet]

TMS,

You can eliminate duplicate code in your __add__ method this way:

    def __add__(self, other):
        if not isinstance(other, Rational):
            other = Rational(other)
        return Rational(self.n * other.d + other.n * self.d, self.d * other.d)

[bvdet]

TMS,

A couple more suggestions and I'm done.
1. If you are overloading multiple operators in your class definition, you may want to add a __coerce__ method:

    def __add__(self, other):
        other = self.__coerce__(other)
        return Rational(self.n * other.d + other.n * self.d, self.d * other.d)
 
    def __coerce__(self, other):
        if isinstance(other, Rational):
            return other
        try:
            return Rational(other)
        except:
            raise ValueError, 'Invalid operand'

2. The assert statement is mostly used for debugging. I would prefer to test for 0 like this:

if den != 0:
    ...... do stuff ......
else:
    raise ValueError, 'The denominator must not be 0'

There may be better ways.

HTH :)

[bartonc]

TMS,

A couple more suggestions and I'm done.
1. If you are overloading multiple operators in your class definition, you may want to add a __coerce__ method:

    def __add__(self, other):
        other = self.__coerce__(other)
        return Rational(self.n * other.d + other.n * self.d, self.d * other.d)
 
    def __coerce__(self, other):
        if isinstance(other, Rational):
            return other
        try:
            return Rational(other)
        except:
            raise ValueError, 'Invalid operand'

2. The assert statement is mostly used for debugging. I would prefer to test for 0 like this:

if den != 0:
    ...... do stuff ......
else:
    raise ValueError, 'The denominator must not be 0'

There may be better ways.

HTH :)

Great suggestions BV. You are the man!

[TMS]

Honestly!! Thank you! Those are great suggestions. We are just starting the unit on Exceptions so the "Try" ... will be great because I can start practicing with it. Thank you very much!


TMS