Path: csiph.com!usenet.pasdenom.info!bete-des-vosges.org!news.redatomik.org!newsfeed.xs4all.nl!newsfeed7.news.xs4all.nl!newsgate.cistron.nl!newsgate.news.xs4all.nl!post.news.xs4all.nl!not-for-mail
To: python-list@python.org
From: Peter Otten <__peter__@web.de>
Subject: Re: scalar vs array and program control
Date: Sat, 25 Jul 2015 15:46:03 +0200
Organization: None
References: <87vbd850qa.fsf@gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 7Bit
User-Agent: KNode/4.13.3
Precedence: list
Newsgroups: comp.lang.python
Message-ID: <mailman.979.1437831975.3674.python-list@python.org>
Lines: 46
NNTP-Posting-Host: 2001:888:2000:d::a6
Xref: csiph.com comp.lang.python:94552

Seb wrote:

> Hello,
> 
> I'm fairly new to Python, struggling to write in a more object-oriented,
> functional style.  I just wrote a function that takes two arrays
> representing sine (y) and cosine (x) angle coordinates, and returns the
> angle in degrees.  I had initially written the function to take
> array-like arguments x/y, but I'd like to generalize and take scalars as
> well.  However, the function has a subsetting operations, which don't
> work with scalars:
> 
>     vmag = np.sqrt((x ** 2) + (y ** 2))
>     ang = np.arctan2(y, x)
>     ang[ang < 0] = ang[ang < 0] + (2 * np.pi) # output range 0 - 2*pi
>     ang[vmag == 0] = 0          # when magnitude is 0 the angle is also 0
>     ang[ang == 0] = 2 * np.pi   # convention
> 
> If I want to take scalars x/y, I naively thought about implementing an
> if/else statement right before the subsetting operations.  However, my
> intuition tells me there must be a proper object-oriented solution to
> this.  Any tips appreciated.

Here's a "duck-typed" solution:

import numpy as np

def _f(x, y):
    vmag = np.sqrt((x ** 2) + (y ** 2))
    ang = np.arctan2(y, x)
    ang[ang < 0] = ang[ang < 0] + (2 * np.pi) # output range 0 - 2*pi
    ang[vmag == 0] = 0          # when magnitude is 0 the angle is also 0
    ang[ang == 0] = 2 * np.pi   # convention
    return ang

def f(x, y):
    try:
        return _f(x, y)
    except IndexError:
        return _f(np.array([x]), y)[0]

However, this is quite inefficient for scalars. Personally I'd implement two 
versions, one for scalars and one for vectors. This may be a little less 
convenient, but most of the time the client code has to deal either with 
scalars or vectors, not both.