py-nutshell.pod6

=begin pod

=TITLE Python to Perl 6 - Nutshell

=SUBTITLE Learning Perl 6 from Python, in a nutshell.

This page is an attempt to provide a way to learn Perl 6
for folks coming from a Python background.  We discuss
the equivalent syntax in Perl 6 for a number of Python
constructs and idioms.

=head1 Basic Syntax

=head2 Hello, world

Let's start with printing "Hello, world!". L<put> in Perl 6 is the
equivalent of L<print> in Python.  Like Python 2, parentheses are
optional.  A return is added to the end of the line.

Python 2

=for code :lang<python>
print "Hello, world!"

Python 3

=for code :lang<python>
print("Hello, world!")

Perl 6

    put "Hello, world!"

There is also L<say>, which behaves similarly, but will call
the L<gist> method of its argument.

Perl 6

   my $hello; ...;
   say "Hello, world!";  # also prints "Hello, world"
   say $hello;           # same as: put $hello.gist

In Python, C<'> and C<"> are interchangeable.
In Perl 6, both may be used for quoting, but double
quotes (C<">) signify that interpolation should be performed.
For instance, variables that start with a C<$>, and expressions
contained in curly braces are interpolated.

Perl 6

    my $planet = 'earth';
    say "Hello, $planet";   # Hello, earth
    say 'Hello, $planet';   # Hello, $planet
    say "Hello, planet number { 1 + 2 }"; # Hello, planet number 3

=head2 Statement Separators

In Python, a newline signifies the end of a statement.
There are a few exceptions: A backslash before a newline
continues a statement across lines.  Also if there is
an unmatched opening parentheses, bracket, or curly
brace, the statement continues across lines, until the
matching braces are closed.

In Perl 6, a semicolon signifies the end of a statement.
The semicolon may be omitted if it is the last statement
of a block.  The semicolon may also be omitted if there
is a closing curly brace followed by a newline.

Python

=for code :lang<python>
print 1 + 2 + \
    3 + 4
print ( 1 +
    2 )

Perl 6

    say 1 + 2 +
        3 + 4;
    if True { say 1 + 2 }

=head2 Blocks

In Python, indentation is used to indicate a block.  Perl 6
uses curly braces.

Python

=for code :lang<python>
if 1 == 2:
    print "Wait, what?"
else:
    print "1 is not 2."

Perl 6

    if 1 == 2 {
        say "Wait, what?"
    } else {
        say "1 is not 2."
    }

Parentheses are optional in both languages in expressions in
conditionals, as shown above.

=head2 Variables

In Python, variables are declared and initialized at the same time:

=for code :lang<python>
foo = 12
bar = 19

In Perl 6, C<my> declares a lexical variable.  A variable can be
initialized with C<=>.  i.e. these can be written as two statements or one.

    my $foo;       # declare
    $foo = 12;     # initialize
    my $bar = 19;  # both at once

Also, as you may have noticed, variables in Perl 6 usually start with
sigils -- symbols indicating the type of their container.  Variables
starting with a C<$> hold scalars.  Variables starting with an C<@>
hold arrays, and variables starting with a C<%> hold a hash (dict).
Immutable variables can be sigil-less, if they are declared with a C<\>.

Python

=begin code :lang<python>
s = 10
l = [1, 2, 3]
d = { a : 12, b : 99 }

print s
print l[2]
print d['a']
# 10, 2, 12
=end code

Perl 6

    my $s = 10;
    my @l = 1, 2, 3;
    my %d = a => 12, b => 99;
    my \x = 99;

    say $s;
    say @l[1];
    say %d<a>;  # or %d{'a'}
    say x;
    # 10, 2, 12, 99


=head2 Scope

In Python, functions and classes create a new scope, but no other
block constructor (e.g. loops, conditionals) creates a scope.  In
Python 2, list comprehensions do not create a new scope, but in Python 3, they do.

In Perl 6, every block creates a lexical scope.

Python

=for code :lang<python>
if True:
    x = 10
print x
# x is now 10

Perl 6

=for code :skip-test<compile time error>
if True {
    my $x = 10
}
say $x
# error, $x is not declared in this scope

=for code
my $x;
if True {
    $x = 10
}
say $x
# ok, x is 10

Python

=for code :lang<python>
x = 10
for x in 1, 2, 3:
   pass
print x
# x is 3

Perl 6

    my \x = 10;
    for 1, 2, 3 -> \x {
        # do nothing
        }
    say x;
    # x is 10

Lambdas in Python can be written as blocks or pointy blocks in Perl 6.

Python

=for code :lang<python>
l = lambda i: i + 12

Perl 6

=for code :preamble<my $i>
my $l = -> $i { $i + 12 }

Another Perl 6 idiom for constructing lambdas is the Whatever star, C<*>.

Perl 6

    my $l = * + 12    # same as above

A C<*> in an expression will become a placeholder for the argument,
and transform the expression into a lambda at compile time.  Each
C<*> in an expression is a separate positional parameter.

See the section below for more constructs regarding subroutines and blocks.

Another example (from the Python L<FAQ|https://docs.python.org/3/faq/programming.html#why-do-lambdas-defined-in-a-loop-with-different-values-all-return-the-same-result>):

Python

=for code :lang<python>
squares = []
for x in range(5):
    squares.append(lambda: x ** 2)
print squares[2]()
print squares[4]()
# both 16 since there is only one x

Perl 6

    my \squares = [];
    for ^5 -> \x {
        squares.append({ x² });
    }
    say squares[2]();
    say squares[4]();
    # 4, 16 since each loop iteration has a lexically scoped x,

Note that C<^N> is like C<range(N)>.  Similarly,
C<N..^M> works like C<range(N, M)> (a list from N
to M - 1).  C<N..M> is a list from N to M.  The
C<^> before or after the C<..> indicates that the
beginning or ending endpoint of the list (or both)
should be excluded.

Also, C<x²> is a cute way of writing C<$x ** 2> (which also
works fine); the unicode superscript 2 squares a number.
Many of the other unicode operators work as you would expect
(exponents, fractions, π), but every unicode operator
or symbol that can be used in Perl 6 has an ASCII equivalent.

=head2 Control Flow

Python has C<for> loops and C<while> loops:

=for code :lang<python>
for i in 1, 2:
    print i
j = 1
while j < 3:
    print j
    j += 1

# 1, 2, 1, 2

Perl 6 also has C<for> loops and C<while> loops:

    for 1, 2 -> $i {
        say $i
    }
    my $j = 1;
    while $j < 3 {
        say $j;
        $j += 1
    }

(Perl 6 also has a few more looping constructs: C<repeat, until>,
C<repeat, while>, C<until>, and C<loop>.)

C<last> leaves a loop in Perl 6, and is analogous to
C<break> in Python.  C<continue> in Python is C<next>
in Perl 6.

Python

=for code :lang<python>
for i in range(10):
    if i == 3:
        continue
    if i == 5:
        break
    print i

Perl 6

    for ^10 -> $i {
        next if $i == 3;
        last if $i == 5;
        say $i;
    }

Using C<if> as a statement modifier (as above) is acceptable
in Perl 6, even outside of a list comprehension.

The C<yield> statement in Python which produces a
C<generator> is like a C<gather>, C<take> construct
in Perl 6.  These both print 1, 2, 3.

Python

=begin code :lang<python>
def count():
    for i in 1, 2, 3:
        yield i

for c in count():
    print c
=end code

Perl 6

    sub count {
        gather {
            for 1, 2, 3 -> $i {
                take $i
            }
        }
    }

    for count() -> $c {
        say $c;
    }


=head2 Lambdas, Functions and Subroutines

Declaring a function (subroutine) with C<def> in Python is accomplished
with C<sub> in Perl 6.

=begin code :lang<python>
def add(a, b):
    return a + b

sub add(\a, \b) {
    return a + b
}
=end code

The C<return> is optional; the value of the last expression is used as
the return value:

=begin code
sub add(\a, \b) {
    a + b
}
=end code

=begin code
# using variables with sigils
sub add($a, $b) {
    $a + $b
}
=end code

Python 2 functions can be called with positional arguments
or keyword arguments.  These are determined by the caller.
In Python 3, some arguments may be "keyword only".
In Perl 6, positional and named arguments are determined
by the signature of the routine.

Python

=begin code :lang<python>
def speak(word, times):
    for i in range(times):
        print word
speak('hi', 2)
speak(word='hi', times=2)
=end code

Perl 6

Positional parameters:

    sub speak($word, $times) {
      say $word for ^$times
    }
    speak('hi', 2);

Named parameters start with a colon:

    sub speak(:$word, :$times) {
      say $word for ^$times
    }
    speak(word => 'hi', times => 2);
    speak(:word<hi>, :times<2>);      # Alternative, more idiomatic

Perl 6 supports multiple dispatch, so several signatures
could be made available by declaring a routine as a C<multi>.

    multi sub speak($word, $times) {
      say $word for ^$times
    }
    multi sub speak(:$word, :$times) {
        speak($word, $times);
    }
    speak('hi', 2);
    speak(:word<hi>, :times<2>);

Named parameters can be sent using a variety of formats:

    sub hello {...};
    # all the same
    hello(name => 'world'); # fat arrow syntax
    hello(:name('world'));  # pair constructor
    hello :name<world>;     # <> quotes words and makes a list
    my $name = 'world';
    hello(:$name);          # lexical var with the same name

Creating an anonymous function can be done with C<sub>, with a block or with
a pointy block.

Python

=for code :lang<python>
square = lambda x: x ** 2

Perl 6

    my $square = sub ($x) { $x ** 2 };  # anonymous sub
    my $square = -> $x { $x ** 2 };     # pointy block
    my $square = { $^x ** 2 };          # placeholder variable
    my $square = { $_ ** 2 };           # topic variable

Placeholder variables are lexicographically ordered to form positional parameters.
i.e. these are the same:

    my $power = { $^x ** $^y };
    my $power = -> $x, $y { $x ** $y };

=head2 List Comprehensions

Postfix statement modifiers and blocks can be combined to make list comprehensions.

Python

=for code :lang<python>
print [ i * 2 for i in 3, 9 ]                      # OUTPUT: «[6, 18]␤»

Perl 6

    say ( $_ * 2 for 3, 9 );                           # OUTPUT: «(6 18)␤»
    say ( { $^i * 2 } for 3, 9 );                      # OUTPUT: «(6 18)␤»
    say ( -> \i { i * 2 } for 3, 9 );                  # OUTPUT: «(6 18)␤»

Conditionals can be applied, but the C<if> comes first,
unlike in Python where the if comes second.

=for code :lang<python>
print [ x * 2 for x in 1, 2, 3 if x > 1 ]          # OUTPUT: «[4, 6]␤»

vs

    say ( $_ * 2 if $_ > 1 for 1, 2, 3 );              # OUTPUT: «(4 6)␤»

For nested loops, the cross product operator C<X>
will help:

=for code :lang<python>
print [ i + j for i in 3,9 for j in 2,10 ]         # OUTPUT: «[5, 13, 11, 19]␤»

becomes either of these:

    say ( { $_[0] + $_[1] } for (3,9) X (2,10) );      # OUTPUT: «(5 13 11 19)␤»
    say ( -> (\i, \j) { i + j } for (3,9) X (2,10) );  # OUTPUT: «(5 13 11 19)␤»

Using C<map> (which is just like Python's C<map>) and
C<grep> (which is like Python's C<filter>) is an alternative.

=head2 Classes and Objects

Here's an example from the Python L<docs|https://docs.python.org/3/tutorial/classes.html#class-and-instance-variables>.
First, "instance variables", aka attributes in Perl 6:

Python:

=for code :lang<python>
class Dog:
    def __init__(self, name):
        self.name = name

Perl 6:

    class Dog {
        has $.name;
    }

Constructors by default take named arguments in Perl 6,
and use the method C<new>.

Python

=for code :lang<python>
d = Dog('Fido')
e = Dog('Buddy')
print d.name
print e.name

Perl 6

    class Dog {}; ...;
    my $d = Dog.new(:name<Fido>);
    my $e = Dog.new(:name<Buddy>);
    say $d.name;
    say $e.name;

Class attributes in Perl 6 can be declared in a few ways.  One way
is to just declare a lexical variable and a method for accessing it.

Python:

=for code :lang<python>
class Dog:
    kind = 'canine'                # class attribute
    def __init__(self, name):
        self.name = name           # instance attribute
d = Dog('Fido')
e = Dog('Buddy')
print d.kind
print e.kind
print d.name
print e.name

Perl 6:

    class Dog {
        my $kind = 'canine';           # class attribute
        method kind { $kind }
        has $.name;                    # instance attribute
    }

    my $d = Dog.new(:name<Fido>);
    my $e = Dog.new(:name<Buddy>);
    say $d.kind;
    say $e.kind;
    say $d.name;
    say $e.name;

To mutate attributes, in Perl 6 you"ll want to use C<is rw>:

Python:

=for code :lang<python>
class Dog:
    def __init__(self, name):
        self.name = name
d = Dog()
d.name = 'rover'

Perl 6:

    class Dog {
        has $.name is rw;
    }
    my $d = Dog.new;
    $d.name = 'rover';

Inheritance is done using C<is>:

Python

=begin code :lang<python>
class Animal:
    def jump(self):
        print ("I am jumping")

class Dog(Animal):
    pass

d = Dog()
d.jump()
=end code

Perl 6

    class Animal {
        method jump {
            say "I am jumping"
        }
    }

    class Dog is Animal {
    }

    my $d = Dog.new;
    $d.jump;

Multiple inheritance is possible by using C<is> multiple times, or with C<also>.

Python

=for code :lang<python>
class Dog(Animal, Friend, Pet):
    pass

Perl 6

    class Animal {}; class Friend {}; class Pet {};
    ...;
    class Dog is Animal is Friend is Pet {};

or

    class Animal {}; class Friend {}; class Pet {};
    ...;
    class Dog is Animal {
        also is Friend;
        also is Pet;
        ...
    }

=head2 Decorators
X<|Decorators (Python)>
X<|@decorator (Python)>

Decorators in Python are a way of wrapping a function
in another one.  In Perl 6, this is done with C<wrap>.

Python

=begin code :lang<python>
def greeter(f):
    def new():
        print 'hello'
        f()
    return new

@greeter
def world():
    print 'world'

world();
=end code

Perl 6

    sub world {
        say 'world'
    }

    &world.wrap(sub () {
        say 'hello';
        callsame;
    });

    world;

An alternative would be to use a trait:

    # declare the trait 'greeter'
    multi sub trait_mod:<is>(Routine $r, :$greeter) {
        $r.wrap(sub {
            say 'hello';
            callsame;
        })
    }

    sub world is greeter {
        say 'world';
    }

    world;

=head2 Context Managers
X<|Context Managers (Python)>
X<|with (Python)>

Context managers in Python declare actions that happen when entering
or exiting a scope.

Here's a python context manager that prints the strings
'hello', 'world', and 'bye':

=begin code :lang<python>
class hello:
    def __exit__(self, type, value, traceback):
        print 'bye'
    def __enter__(self):
        print 'hello'

with hello():
    print 'world'
=end code

For enter and exit events, passing a block as
an argument would be one option:

    sub hello(Block $b) {
        say 'hello';
        $b();
        say 'bye';
    }

    hello {
        say 'world';
    }

A related idea is 'L<Phasers|/language/phasers>' which may be set up to
run on entering or leaving a block.

    {
        LEAVE say 'bye';
        ENTER say 'hello';
        say 'world';
    }

=head2 Input
X<|raw_input (Python)>
X<|input (Python)>

See the L<prompt> function.

=end pod

# vim: expandtab softtabstop=4 shiftwidth=4 ft=perl6