OnWorks favicon

hy - Online in the Cloud

Run hy in OnWorks free hosting provider over Ubuntu Online, Fedora Online, Windows online emulator or MAC OS online emulator

This is the command hy that can be run in the OnWorks free hosting provider using one of our multiple free online workstations such as Ubuntu Online, Fedora Online, Windows online emulator or MAC OS online emulator



hy - hy Documentation [image: Hy] [image]

Try Hy https://try-hy.appspot.com

PyPI https://pypi.python.org/pypi/hy

Source https://github.com/hylang/hy

List hylang-discuss

IRC #hy on Freenode

Build status

Hy is a wonderful dialect of Lisp that's embedded in Python.

Since Hy transforms its Lisp code into the Python Abstract Syntax Tree, you have
the whole beautiful world of Python at your fingertips, in Lisp form!



[image: Karen Rustard's Cuddles] [image]

(Thanks to Karen Rustad for Cuddles!)


1. Create a Virtual Python Environment.

2. Activate your Virtual Python Environment.

3. Install hy from PyPI with pip install hy.

4. Start a REPL with hy.

5. Type stuff in the REPL:

=> (print "Hy!")
=> (defn salutationsnm [name] (print (+ "Hy " name "!")))
=> (salutationsnm "YourName")
Hy YourName!


6. Hit CTRL-D when you're done.

OMG! That's amazing! I want to write a Hy program.

7. Open up an elite programming editor and type:

(print "I was going to code in Python syntax, but then I got Hy.")

8. Save as awesome.hy.

9. And run your first Hy program:

hy awesome.hy

Take a deep breath so as to not hyperventilate.

Smile villainously and sneak off to your hydeaway and do unspeakable things.


Welcome to the Hy tutorial!

In a nutshell, Hy is a Lisp dialect, but one that converts its structure into Python ...
literally a conversion into Python's abstract syntax tree! (Or to put it in more crude
terms, Hy is lisp-stick on a Python!)

This is pretty cool because it means Hy is several things:

· A Lisp that feels very Pythonic

· For Lispers, a great way to use Lisp's crazy powers but in the wide world of Python's
libraries (why yes, you now can write a Django application in Lisp!)

· For Pythonistas, a great way to start exploring Lisp, from the comfort of Python!

· For everyone: a pleasant language that has a lot of neat ideas!

Basic intro to Lisp for Pythonistas
Okay, maybe you've never used Lisp before, but you've used Python!

A "hello world" program in Hy is actually super simple. Let's try it:

(print "hello world")

See? Easy! As you may have guessed, this is the same as the Python version of:

print "hello world"

To add up some super simple math, we could do:

(+ 1 3)

Which would return 4 and would be the equivalent of:

1 + 3

What you'll notice is that the first item in the list is the function being called and the
rest of the arguments are the arguments being passed in. In fact, in Hy (as with most
Lisps) we can pass in multiple arguments to the plus operator:

(+ 1 3 55)

Which would return 59.

Maybe you've heard of Lisp before but don't know much about it. Lisp isn't as hard as you
might think, and Hy inherits from Python, so Hy is a great way to start learning Lisp.
The main thing that's obvious about Lisp is that there's a lot of parentheses. This might
seem confusing at first, but it isn't so hard. Let's look at some simple math that's
wrapped in a bunch of parentheses that we could enter into the Hy interpreter:

(setv result (- (/ (+ 1 3 88) 2) 8))

This would return 38. But why? Well, we could look at the equivalent expression in

result = ((1 + 3 + 88) / 2) - 8

If you were to try to figure out how the above were to work in python, you'd of course
figure out the results by solving each inner parenthesis. That's the same basic idea in
Hy. Let's try this exercise first in Python:

result = ((1 + 3 + 88) / 2) - 8
# simplified to...
result = (92 / 2) - 8
# simplified to...
result = 46 - 8
# simplified to...
result = 38

Now let's try the same thing in Hy:

(setv result (- (/ (+ 1 3 88) 2) 8))
; simplified to...
(setv result (- (/ 92 2) 8))
; simplified to...
(setv result (- 46 8))
; simplified to...
(setv result 38)

As you probably guessed, this last expression with setv means to assign the variable
"result" to 38.

See? Not too hard!

This is the basic premise of Lisp. Lisp stands for "list processing"; this means that the
structure of the program is actually lists of lists. (If you're familiar with Python
lists, imagine the entire same structure as above but with square brackets instead, any
you'll be able to see the structure above as both a program and a datastructure.) This is
easier to understand with more examples, so let's write a simple Python program, test it,
and then show the equivalent Hy program:

def simple_conversation():
print "Hello! I'd like to get to know you. Tell me about yourself!"
name = raw_input("What is your name? ")
age = raw_input("What is your age? ")
print "Hello " + name + "! I see you are " + age + " years old."


If we ran this program, it might go like:

Hello! I'd like to get to know you. Tell me about yourself!
What is your name? Gary
What is your age? 38
Hello Gary! I see you are 38 years old.

Now let's look at the equivalent Hy program:

(defn simple-conversation []
(print "Hello! I'd like to get to know you. Tell me about yourself!")
(setv name (raw-input "What is your name? "))
(setv age (raw-input "What is your age? "))
(print (+ "Hello " name "! I see you are "
age " years old.")))


If you look at the above program, as long as you remember that the first element in each
list of the program is the function (or macro... we'll get to those later) being called
and that the rest are the arguments, it's pretty easy to figure out what this all means.
(As you probably also guessed, defn is the Hy method of defining methods.)

Still, lots of people find this confusing at first because there's so many parentheses,
but there are plenty of things that can help make this easier: keep indentation nice and
use an editor with parenthesis matching (this will help you figure out what each
parenthesis pairs up with) and things will start to feel comfortable.

There are some advantages to having a code structure that's actually a very simple data
structure as the core of Lisp is based on. For one thing, it means that your programs are
easy to parse and that the entire actual structure of the program is very clearly exposed
to you. (There's an extra step in hy where the structure you see is converted to Python's
own representations ... in "purer" Lisps such as Common Lisp or Emacs Lisp, the data
structure you see in the code and the data structure that is executed is much more
literally close.)

Another implication of this is macros: if a program's structure is a simple data
structure, that means you can write code that can write code very easily, meaning that
implementing entirely new language features can be very fast. Previous to Hy, this wasn't
very possible for Python programmers ... now you too can make use of macros' incredible
power (just be careful to not aim them footward)!

Hy is a Lisp-flavored Python
Hy converts to Python's own abstract syntax tree, so you'll soon start to find that all
the familiar power of python is at your fingertips.

You have full access to Python's data types and standard library in Hy. Let's experiment
with this in the hy interpreter:

=> [1 2 3]
[1, 2, 3]
=> {"dog" "bark"
... "cat" "meow"}
{'dog': 'bark', 'cat': 'meow'}
=> (, 1 2 3)
(1, 2, 3)

If you are familiar with other Lisps, you may be interested that Hy supports the Common
Lisp method of quoting:

=> '(1 2 3)
(1L 2L 3L)

You also have access to all the built-in types' nice methods:

=> (.strip " fooooo ")

What's this? Yes indeed, this is precisely the same as:

" fooooo ".strip()

That's right---Lisp with dot notation! If we have this string assigned as a variable, we
can also do the following:

(setv this-string " fooooo ")

What about conditionals?:

(if (try-some-thing)
(print "this is if true")
(print "this is if false"))

As you can tell above, the first argument to if is a truth test, the second argument is
the body if true, and the third argument (optional!) is if false (ie. else).

If you need to do more complex conditionals, you'll find that you don't have elif
available in Hy. Instead, you should use something called cond. In Python, you might do
something like:

somevar = 33
if somevar > 50:
print "That variable is too big!"
elif somevar < 10:
print "That variable is too small!"
print "That variable is jussssst right!"

In Hy, you would do:

[(> somevar 50)
(print "That variable is too big!")]
[(< somevar 10)
(print "That variable is too small!")]
(print "That variable is jussssst right!")])

What you'll notice is that cond switches off between a some statement that is executed and
checked conditionally for true or falseness, and then a bit of code to execute if it turns
out to be true. You'll also notice that the else is implemented at the end simply by
checking for true -- that's because true will always be true, so if we get this far, we'll
always run that one!

You might notice above that if you have code like:

(if some-condition

But wait! What if you want to execute more than one statement in the body of one of

You can do the following:

(if (try-some-thing)
(print "this is if true")
(print "and why not, let's keep talking about how true it is!))
(print "this one's still simply just false"))

You can see that we used do to wrap multiple statements. If you're familiar with other
Lisps, this is the equivalent of progn elsewhere.

Comments start with semicolons:

(print "this will run")
; (print "but this will not")
(+ 1 2 3) ; we'll execute the addition, but not this comment!

Looping is not hard but has a kind of special structure. In Python, we might do:

for i in range(10):
print "'i' is now at " + str(i)

The equivalent in Hy would be:

(for [i (range 10)]
(print (+ "'i' is now at " (str i))))

You can also import and make use of various Python libraries. For example:

(import os)

(if (os.path.isdir "/tmp/somedir")
(os.mkdir "/tmp/somedir/anotherdir")
(print "Hey, that path isn't there!"))

Python's context managers (with statements) are used like this:

(with [[f (open "/tmp/data.in")]]
(print (.read f)))

which is equivalent to:

with open("/tmp/data.in") as f:
print f.read()

And yes, we do have List comprehensions! In Python you might do:

odds_squared = [
pow(num, 2)
for num in range(100)
if num % 2 == 1]

In Hy, you could do these like:

(setv odds-squared
(pow num 2)
(num (range 100))
(= (% num 2) 1)))

; And, an example stolen shamelessly from a Clojure page:
; Let's list all the blocks of a Chessboard:

(, x y)
(x (range 8)

; [(0, 'A'), (0, 'B'), (0, 'C'), (0, 'D'), (0, 'E'), (0, 'F'), (0, 'G'), (0, 'H'),
; (1, 'A'), (1, 'B'), (1, 'C'), (1, 'D'), (1, 'E'), (1, 'F'), (1, 'G'), (1, 'H'),
; (2, 'A'), (2, 'B'), (2, 'C'), (2, 'D'), (2, 'E'), (2, 'F'), (2, 'G'), (2, 'H'),
; (3, 'A'), (3, 'B'), (3, 'C'), (3, 'D'), (3, 'E'), (3, 'F'), (3, 'G'), (3, 'H'),
; (4, 'A'), (4, 'B'), (4, 'C'), (4, 'D'), (4, 'E'), (4, 'F'), (4, 'G'), (4, 'H'),
; (5, 'A'), (5, 'B'), (5, 'C'), (5, 'D'), (5, 'E'), (5, 'F'), (5, 'G'), (5, 'H'),
; (6, 'A'), (6, 'B'), (6, 'C'), (6, 'D'), (6, 'E'), (6, 'F'), (6, 'G'), (6, 'H'),
; (7, 'A'), (7, 'B'), (7, 'C'), (7, 'D'), (7, 'E'), (7, 'F'), (7, 'G'), (7, 'H')]

Python has support for various fancy argument and keyword arguments. In Python we might

>>> def optional_arg(pos1, pos2, keyword1=None, keyword2=42):
... return [pos1, pos2, keyword1, keyword2]
>>> optional_arg(1, 2)
[1, 2, None, 42]
>>> optional_arg(1, 2, 3, 4)
[1, 2, 3, 4]
>>> optional_arg(keyword1=1, pos2=2, pos1=3, keyword2=4)
[3, 2, 1, 4]

The same thing in Hy:

=> (defn optional-arg [pos1 pos2 &optional keyword1 [keyword2 42]]
... [pos1 pos2 keyword1 keyword2])
=> (optional-arg 1 2)
[1 2 None 42]
=> (optional-arg 1 2 3 4)
[1 2 3 4]

If you're running a version of Hy past 0.10.1 (eg, git master), there's also a nice new
keyword argument syntax:

=> (optional-arg :keyword1 1
... :pos2 2
... :pos1 3
... :keyword2 4)
[3, 2, 1, 4]

Otherwise, you can always use apply. But what's apply?

Are you familiar with passing in *args and **kwargs in Python?:

>>> args = [1 2]
>>> kwargs = {"keyword2": 3
... "keyword1": 4}
>>> optional_arg(*args, **kwargs)

We can reproduce this with apply:

=> (setv args [1 2])
=> (setv kwargs {"keyword2" 3
... "keyword1" 4})
=> (apply optional-arg args kwargs)
[1, 2, 4, 3]

There's also a dictionary-style keyword arguments construction that looks like:

(defn another-style [&key {"key1" "val1" "key2" "val2"}]
[key1 key2])

The difference here is that since it's a dictionary, you can't rely on any specific
ordering to the arguments.

Hy also supports *args and **kwargs. In Python:

def some_func(foo, bar, *args, **kwargs):
import pprint
pprint.pprint((foo, bar, args, kwargs))

The Hy equivalent:

(defn some-func [foo bar &rest args &kwargs kwargs]
(import pprint)
(pprint.pprint (, foo bar args kwargs)))

Finally, of course we need classes! In Python, we might have a class like:

class FooBar(object):
Yet Another Example Class
def __init__(self, x):
self.x = x

def get_x(self):
Return our copy of x
return self.x

In Hy:

(defclass FooBar [object]
"Yet Another Example Class"
(fn [self x]
(setv self.x x)
; Currently needed for --init-- because __init__ needs None
; Hopefully this will go away :)

(fn [self]
"Return our copy of x"

You can also do class-level attributes. In Python:

class Customer(models.Model):
name = models.CharField(max_length=255)
address = models.TextField()
notes = models.TextField()

In Hy:

(defclass Customer [models.Model]
[[name (models.CharField :max-length 255})]
[address (models.TextField)]
[notes (models.TextField)]])

Hy <-> Python interop
By importing Hy, you can use Hy directly from Python!

If you save the following in greetings.hy:

(defn greet [name] (print "hello from hy," name))

Then you can use it directly from python, by importing hy before importing the module. In

import hy
import greetings


You can also declare a function in python (or even a class!) and use it in Hy!

If you save the following in greetings.py in Python:

def greet(name):
print("hello, %s" % (name))

You can use it in Hy:

(import greetings)
(.greet greetings "foo")

To use keyword arguments, you can use in greetings.py:

def greet(name, title="Sir"):
print("Greetings, %s %s" % (title,name))

(import greetings)
(.greet greetings "Foo")
(.greet greetings "Foo" "Darth")
(apply (. greetings greet) ["Foo"] {"title" "Lord"})

Which would output:

Greetings, Sir Foo

Greetings, Darth Foo

Greetings, Lord Foo

Hy also features something known as the "threading macro", a really neat feature of
Clojure's. The "threading macro" (written as ->) is used to avoid deep nesting of

The threading macro inserts each expression into the next expression's first argument

Let's take the classic:

(loop (print (eval (read))))

Rather than write it like that, we can write it as follows:

(-> (read) (eval) (print) (loop))

Now, using python-sh, we can show how the threading macro (because of python-sh's setup)
can be used like a pipe:

=> (import [sh [cat grep wc]])
=> (-> (cat "/usr/share/dict/words") (grep "-E" "^hy") (wc "-l"))

Which, of course, expands out to:

(wc (grep (cat "/usr/share/dict/words") "-E" "^hy") "-l")

Much more readable, no? Use the threading macro!


“You know, Minister, I disagree with Dumbledore on many counts…but you cannot deny he’s
got style…” — Phineas Nigellus Black, Harry Potter and the Order of the Phoenix

The Hy style guide intends to be a set of ground rules for the Hyve (yes, the Hy community
prides itself in appending Hy to everything) to write idiomatic Hy code. Hy derives a lot
from Clojure & Common Lisp, while always maintaining Python interopability.

The Tao of Hy
Ummon asked the head monk, "What sutra are you lecturing on?"
"The Nirvana Sutra."
"The Nirvana Sutra has the Four Virtues, hasn't it?"
"It has."
Ummon asked, picking up a cup, "How many virtues has this?"
"None at all," said the monk.
"But ancient people said it had, didn't they?" said Ummon.
"What do you think of what they said?"
Ummon struck the cup and asked, "You understand?"
"No," said the monk.
"Then," said Ummon, "You'd better go on with your lectures on the sutra."
— the (koan) macro

The following illustrates a brief list of design decisions that went into the making of

· Look like a Lisp; DTRT with it (e.g. dashes turn to underscores, earmuffs turn to

· We're still Python. Most of the internals translate 1:1 to Python internals.

· Use Unicode everywhere.

· Fix the bad decisions in Python 2 when we can (see true_division).

· When in doubt, defer to Python.

· If you're still unsure, defer to Clojure.

· If you're even more unsure, defer to Common Lisp.

· Keep in mind we're not Clojure. We're not Common Lisp. We're Homoiconic Python, with
extra bits that make sense.

Layout & Indentation
· Avoid trailing spaces. They suck!

· Indentation shall be 2 spaces (no hard tabs), except when matching the indentation of
the previous line.

;; Good (and preferred)
(defn fib [n]
(if (<= n 2)
(+ (fib (- n 1)) (fib (- n 2)))))

;; Still okay
(defn fib [n]
(if (<= n 2) n (+ (fib (- n 1)) (fib (- n 2)))))

;; Still okay
(defn fib [n]
(if (<= n 2)
(+ (fib (- n 1)) (fib (- n 2)))))

;; Hysterically ridiculous
(defn fib [n]
(if (<= n 2)
n ;; yes, I love randomly hitting the space key
(+ (fib (- n 1)) (fib (- n 2)))))

· Parentheses must never be left alone, sad and lonesome on their own line.

;; Good (and preferred)
(defn fib [n]
(if (<= n 2)
(+ (fib (- n 1)) (fib (- n 2)))))

;; Hysterically ridiculous
(defn fib [n]
(if (<= n 2)
(+ (fib (- n 1)) (fib (- n 2)))

· Vertically align let blocks.

(let [[foo (bar)]
[qux (baz)]]
(foo qux))

· Inline comments shall be two spaces from the end of the code; they must always have a
space between the comment character and the start of the comment. Also, try to not
comment the obvious.

;; Good
(setv ind (dec x)) ; indexing starts from 0

;; Style-compliant but just states the obvious
(setv ind (dec x)) ; sets index to x-1

;; Bad
(setv ind (dec x));typing words for fun

Coding Style
· As a convention, try not to use def for anything other than global variables; use setv
inside functions, loops, etc.

;; Good (and preferred)
(def *limit* 400000)

(defn fibs [a b]
(while true
(yield a)
(setv (, a b) (, b (+ a b)))))

;; Bad (and not preferred)
(defn fibs [a b]
(while true
(yield a)
(def (, a b) (, b (+ a b)))))

· Do not use s-expression syntax where vector syntax is intended. For instance, the fact
that the former of these two examples works is just because the compiler isn't overly
strict. In reality, the correct syntax in places such as this is the latter.

;; Bad (and evil)
(defn foo (x) (print x))
(foo 1)

;; Good (and preferred)
(defn foo [x] (print x))
(foo 1)

· Use the threading macro or the threading tail macros when encountering deeply nested
s-expressions. However, be judicious when using them. Do use them when clarity and
readability improves; do not construct convoluted, hard to understand expressions.

;; Preferred
(def *names*
(with [f (open "names.txt")]
(-> (.read f) (.strip) (.replace "\"" "") (.split ",") (sorted))))

;; Not so good
(def *names*
(with [f (open "names.txt")]
(sorted (.split "," (.replace "\"" "" (.strip (.read f)))))))

;; Probably not a good idea
(defn square? [x]
(->> 2 (pow (int (sqrt x))) (= x)))

· Clojure-style dot notation is preferred over the direct call of the object's method,
though both will continue to be supported.

;; Good
(with [fd (open "/etc/passwd")]
(print (.readlines fd)))

;; Not so good
(with [fd (open "/etc/passwd")]
(print (fd.readlines)))

“Fashions fade, style is eternal” —Yves Saint Laurent

This guide is just a set of community guidelines, and obviously, community guidelines do
not make sense without an active community. Contributions are welcome. Join us at #hy in
freenode, blog about it, tweet about it, and most importantly, have fun with Hy.

· This guide is heavily inspired from @paultag 's blog post Hy Survival Guide

· The Clojure Style Guide



Command Line Interface
Command Line Options
-c <command>
Execute the Hy code in command.

$ hy -c "(print (+ 2 2))"

-i <command>
Execute the Hy code in command, then stay in REPL.

-m <module>
Execute the Hy code in module, including defmain if defined.

The -m flag terminates the options list so that all arguments after the module name
are passed to the module in sys.argv.

New in version 0.10.2.

--spy Print equivalent Python code before executing. For example:

=> (defn salutationsnm [name] (print (+ "Hy " name "!")))
def salutationsnm(name):
return print(((u'Hy ' + name) + u'!'))
=> (salutationsnm "YourName")
Hy YourName!

New in version 0.9.11.

Print extended tracebacks for Hy exceptions.

New in version 0.9.12.

-v Print the Hy version number and exit.

Command Line Options
file[, fileN]
Compile Hy code to Python bytecode. For example, save the following code as

(defn hy-hy [name]
(print (+ "Hy " name "!")))

(hy-hy "Afroman")

Then run:

$ hyc hyname.hy
$ python hyname.pyc
Hy Afroman!

New in version 0.10.1.

Command Line Options

Show the parsed source structure.


Show the generated AST.


Do not show the Python code generated from the AST.

Hy (the language)
This is incomplete; please consider contributing to the documentation effort.

Theory of Hy
Hy maintains, over everything else, 100% compatibility in both directions with Python
itself. All Hy code follows a few simple rules. Memorize this, as it's going to come in

These rules help ensure that Hy code is idiomatic and interfaceable in both languages.

· Symbols in earmufs will be translated to the upper-cased version of that string. For
example, foo will become FOO.

· UTF-8 entities will be encoded using punycode and prefixed with hy_. For instance,
will become hy_w7h, will become hy_g6h, and i♥u will become hy_iu_t0x.

· Symbols that contain dashes will have them replaced with underscores. For example,
render-template will become render_template. This means that symbols with dashes will
shadow their underscore equivalents, and vice versa.

Hy features a number of special forms that are used to help generate correct Python AST.
The following are "special" forms, which may have behavior that's slightly unexpected in
some situations.

New in version 0.10.0.

. is used to perform attribute access on objects. It uses a small DSL to allow quick
access to attributes and items in a nested data structure.

For instance,

(. foo bar baz [(+ 1 2)] frob)

Compiles down to:

foo.bar.baz[1 + 2].frob

. compiles its first argument (in the example, foo) as the object on which to do the
attribute dereference. It uses bare symbols as attributes to access (in the example, bar,
baz, frob), and compiles the contents of lists (in the example, [(+ 1 2)]) for indexation.
Other arguments throw a compilation error.

Access to unknown attributes throws an AttributeError. Access to unknown keys throws an
IndexError (on lists and tuples) or a KeyError (on dictionaries).

-> (or the threading macro) is used to avoid nesting of expressions. The threading macro
inserts each expression into the next expression's first argument place. The following
code demonstrates this:

=> (defn output [a b] (print a b))
=> (-> (+ 4 6) (output 5))
10 5

->> (or the threading tail macro) is similar to the threading macro, but instead of
inserting each expression into the next expression's first argument, it appends it as the
last argument. The following code demonstrates this:

=> (defn output [a b] (print a b))
=> (->> (+ 4 6) (output 5))
5 10

apply is used to apply an optional list of arguments and an optional dictionary of kwargs
to a function.

Usage: (apply fn-name [args] [kwargs])


(defn thunk []
"hy there")

(apply thunk)
;=> "hy there"

(defn total-purchase [price amount &optional [fees 1.05] [vat 1.1]]
(* price amount fees vat))

(apply total-purchase [10 15])
;=> 173.25

(apply total-purchase [10 15] {"vat" 1.05})
;=> 165.375

(apply total-purchase [] {"price" 10 "amount" 15 "vat" 1.05})
;=> 165.375

and is used in logical expressions. It takes at least two parameters. If all parameters
evaluate to True, the last parameter is returned. In any other case, the first false value
will be returned. Example usage:

=> (and True False)

=> (and True True)

=> (and True 1)

=> (and True [] False True)

and short-circuits and stops evaluating parameters as soon as the first false is

=> (and False (print "hello"))

assert is used to verify conditions while the program is running. If the condition is not
met, an AssertionError is raised. assert may take one or two parameters. The first
parameter is the condition to check, and it should evaluate to either True or False. The
second parameter, optional, is a label for the assert, and is the string that will be
raised with the AssertionError. For example:

(assert (= variable expected-value))

(assert False)
; AssertionError

(assert (= 1 2) "one should equal two")
; AssertionError: one should equal two

assoc is used to associate a key with a value in a dictionary or to set an index of a list
to a value. It takes at least three parameters: the data structure to be modified, a key
or index, and a value. If more than three parameters are used, it will associate in pairs.

Examples of usage:

=>(let [[collection {}]]
... (assoc collection "Dog" "Bark")
... (print collection))
{u'Dog': u'Bark'}

=>(let [[collection {}]]
... (assoc collection "Dog" "Bark" "Cat" "Meow")
... (print collection))
{u'Cat': u'Meow', u'Dog': u'Bark'}

=>(let [[collection [1 2 3 4]]]
... (assoc collection 2 None)
... (print collection))
[1, 2, None, 4]

assoc modifies the datastructure in place and returns None.

break is used to break out from a loop. It terminates the loop immediately. The following
example has an infinite while loop that is terminated as soon as the user enters k.

(while True (if (= "k" (raw-input "? "))
(print "Try again")))

cond can be used to build nested if statements. The following example shows the
relationship between the macro and its expansion:

(cond [condition-1 result-1]
[condition-2 result-2])

(if condition-1 result-1
(if condition-2 result-2))

As shown below, only the first matching result block is executed.

=> (defn check-value [value]
... (cond [(< value 5) (print "value is smaller than 5")]
... [(= value 5) (print "value is equal to 5")]
... [(> value 5) (print "value is greater than 5")]
... [True (print "value is something that it should not be")]))

=> (check-value 6)
value is greater than 5

continue returns execution to the start of a loop. In the following example,
(side-effect1) is called for each iteration. (side-effect2), however, is only called on
every other value in the list.

;; assuming that (side-effect1) and (side-effect2) are functions and
;; collection is a list of numerical values

(for [x collection]
(side-effect1 x)
(if (% x 2)
(side-effect2 x)))

dict-comp is used to create dictionaries. It takes three or four parameters. The first
two parameters are for controlling the return value (key-value pair) while the third is
used to select items from a sequence. The fourth and optional parameter can be used to
filter out some of the items in the sequence based on a conditional expression.

=> (dict-comp x (* x 2) [x (range 10)] (odd? x))
{1: 2, 3: 6, 9: 18, 5: 10, 7: 14}

do / progn
do and progn are used to evaluate each of their arguments and return the last one. Return
values from every other than the last argument are discarded. It can be used in lambda or
list-comp to perform more complex logic as shown in one of the following examples.

Some example usage:

=> (if true
... (do (print "Side effects rock!")
... (print "Yeah, really!")))
Side effects rock!
Yeah, really!

;; assuming that (side-effect) is a function that we want to call for each
;; and every value in the list, but whose return value we do not care about
=> (list-comp (do (side-effect x)
... (if (< x 5) (* 2 x)
... (* 4 x)))
... (x (range 10)))
[0, 2, 4, 6, 8, 20, 24, 28, 32, 36]

do can accept any number of arguments, from 1 to n.

def / setv
def and setv are used to bind a value, object, or function to a symbol. For example:

=> (def names ["Alice" "Bob" "Charlie"])
=> (print names)
[u'Alice', u'Bob', u'Charlie']

=> (setv counter (fn [collection item] (.count collection item)))
=> (counter [1 2 3 4 5 2 3] 2)

New classes are declared with defclass. It can takes two optional parameters: a vector
defining a possible super classes and another vector containing attributes of the new
class as two item vectors.

(defclass class-name [super-class-1 super-class-2]
[[attribute value]])

Both values and functions can be bound on the new class as shown by the example below:

=> (defclass Cat []
... [[age None]
... [colour "white"]
... [speak (fn [self] (print "Meow"))]])

=> (def spot (Cat))
=> (setv spot.colour "Black")
=> (.speak spot)

defn / defun
defn and defun macros are used to define functions. They take three parameters: the name
of the function to define, a vector of parameters, and the body of the function:

(defn name [params] body)

Parameters may have the following keywords in front of them:

Parameter is optional. The parameter can be given as a two item list, where the
first element is parameter name and the second is the default value. The parameter
can be also given as a single item, in which case the default value is None.

=> (defn total-value [value &optional [value-added-tax 10]]
... (+ (/ (* value value-added-tax) 100) value))

=> (total-value 100)

=> (total-value 100 1)


Parameter will contain 0 or more keyword arguments.

The following code examples defines a function that will print all keyword
arguments and their values.

=> (defn print-parameters [&kwargs kwargs]
... (for [(, k v) (.items kwargs)] (print k v)))

=> (apply print-parameters [] {"parameter-1" 1 "parameter-2" 2})
parameter-2 2
parameter-1 1

&rest Parameter will contain 0 or more positional arguments. No other positional
arguments may be specified after this one.

The following code example defines a function that can be given 0 to n numerical
parameters. It then sums every odd number and subtracts every even number.

=> (defn zig-zag-sum [&rest numbers]
(let [[odd-numbers (list-comp x [x numbers] (odd? x))]
[even-numbers (list-comp x [x numbers] (even? x))]]
(- (sum odd-numbers) (sum even-numbers))))

=> (zig-zag-sum)
=> (zig-zag-sum 3 9 4)
=> (zig-zag-sum 1 2 3 4 5 6)

defn-alias / defun-alias
New in version 0.10.0.

The defn-alias and defun-alias macros are much like defn, with the distinction that
instead of defining a function with a single name, these can also define aliases. Other
than taking a list of symbols for function names as the first parameter, defn-alias and
defun-alias are no different from defn and defun.

=> (defn-alias [main-name alias] []
... (print "Hello!"))
=> (main-name)
=> (alias)

New in version 0.10.1.

The defmain macro defines a main function that is immediately called with sys.argv as
arguments if and only if this file is being executed as a script. In other words, this:

(defmain [&rest args]
(do-something-with args))

is the equivalent of:

def main(*args):
return 0

if __name__ == "__main__":
import sys
retval = main(*sys.arg)

if isinstance(retval, int):

Note that as you can see above, if you return an integer from this function, this will be
used as the exit status for your script. (Python defaults to exit status 0 otherwise,
which means everything's okay!)

(Since (sys.exit 0) is not run explicitly in the case of a non-integer return from
defmain, it's a good idea to put (defmain) as the last piece of code in your file.)

defmacro is used to define macros. The general format is (defmacro name [parameters]

The following example defines a macro that can be used to swap order of elements in code,
allowing the user to write code in infix notation, where operator is in between the

=> (defmacro infix [code]
... (quasiquote (
... (unquote (get code 1))
... (unquote (get code 0))
... (unquote (get code 2)))))

=> (infix (1 + 1))

defmacro-alias is used to define macros with multiple names (aliases). The general format
is (defmacro-alias [names] [parameters] expr). It creates multiple macros with the same
parameter list and body, under the specified list of names.

The following example defines two macros, both of which allow the user to write code in
infix notation.

=> (defmacro-alias [infix infi] [code]
... (quasiquote (
... (unquote (get code 1))
... (unquote (get code 0))
... (unquote (get code 2)))))

=> (infix (1 + 1))
=> (infi (1 + 1))

New in version 0.9.12.

defmacro/g! is a special version of defmacro that is used to automatically generate gensym
for any symbol that starts with g!.

For example, g!a would become (gensym "a").

Section using-gensym

New in version 0.9.12.

defreader defines a reader macro, enabling you to restructure or modify syntax.

=> (defreader ^ [expr] (print expr))
=> #^(1 2 3 4)
(1 2 3 4)
=> #^"Hello"

Section Reader Macros

New in version 0.9.12.

del removes an object from the current namespace.

=> (setv foo 42)
=> (del foo)
=> foo
Traceback (most recent call last):
File "<console>", line 1, in <module>
NameError: name 'foo' is not defined

del can also remove objects from mappings, lists, and more.

=> (setv test (list (range 10)))
=> test
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
=> (del (slice test 2 4)) ;; remove items from 2 to 4 excluded
=> test
[0, 1, 4, 5, 6, 7, 8, 9]
=> (setv dic {"foo" "bar"})
=> dic
{"foo": "bar"}
=> (del (get dic "foo"))
=> dic

New in version 0.10.1.

doto is used to simplify a sequence of method calls to an object.

=> (doto [] (.append 1) (.append 2) .reverse)
[2 1]

=> (setv collection [])
=> (.append collection 1)
=> (.append collection 2)
=> (.reverse collection)
=> collection
[2 1]

eval evaluates a quoted expression and returns the value.

=> (eval '(print "Hello World"))
"Hello World"

first / car
first and car are macros for accessing the first element of a collection:

=> (first (range 10))

for is used to call a function for each element in a list or vector. The results of each
call are discarded and the for expression returns None instead. The example code iterates
over collection and for each element in collection calls the side-effect function with
element as its argument:

;; assuming that (side-effect) is a function that takes a single parameter
(for [element collection] (side-effect element))

;; for can have an optional else block
(for [element collection] (side-effect element)
(else (side-effect-2)))

The optional else block is only executed if the for loop terminates normally. If the
execution is halted with break, the else block does not execute.

=> (for [element [1 2 3]] (if (< element 3)
... (print element)
... (break))
... (else (print "loop finished")))

=> (for [element [1 2 3]] (if (< element 4)
... (print element)
... (break))
... (else (print "loop finished")))
loop finished

genexpr is used to create generator expressions. It takes two or three parameters. The
first parameter is the expression controlling the return value, while the second is used
to select items from a list. The third and optional parameter can be used to filter out
some of the items in the list based on a conditional expression. genexpr is similar to
list-comp, except it returns an iterable that evaluates values one by one instead of
evaluating them immediately.

=> (def collection (range 10))
=> (def filtered (genexpr x [x collection] (even? x)))
=> (list filtered)
[0, 2, 4, 6, 8]

New in version 0.9.12.

gensym is used to generate a unique symbol that allows macros to be written without
accidental variable name clashes.

=> (gensym)

=> (gensym "x")

Section using-gensym

get is used to access single elements in lists and dictionaries. get takes two parameters:
the data structure and the index or key of the item. It will then return the corresponding
value from the dictionary or the list. Example usage:

=> (let [[animals {"dog" "bark" "cat" "meow"}]
... [numbers ["zero" "one" "two" "three"]]]
... (print (get animals "dog"))
... (print (get numbers 2)))

get raises a KeyError if a dictionary is queried for a non-existing key.

get raises an IndexError if a list or a tuple is queried for an index that is out of

global can be used to mark a symbol as global. This allows the programmer to assign a
value to a global symbol. Reading a global symbol does not require the global keyword --
only assigning it does.

The following example shows how the global symbol a is assigned a value in a function and
is later on printed in another function. Without the global keyword, the second function
would have thrown a NameError.

(defn set-a [value]
(global a)
(setv a value))

(defn print-a []
(print a))

(set-a 5)

if / if-not
New in version 0.10.0: if-not

if is used to conditionally select code to be executed. It has to contain a condition
block and the block to be executed if the condition block evaluates to True. Optionally,
it may contain a final block that is executed in case the evaluation of the condition is

if-not is similar, but the second block will be executed when the condition fails while
the third and final block is executed when the test succeeds -- the opposite order of if.

Example usage:

(if (money-left? account)
(print "let's go shopping")
(print "let's go and work"))

(if-not (money-left? account)
(print "let's go and work")
(print "let's go shopping"))

Python truthiness is respected. None, False, zero of any numeric type, an empty sequence,
and an empty dictionary are considered False; everything else is considered True.

lisp-if / lif and lisp-if-not / lif-not
New in version 0.10.0.

New in version 0.10.2: lisp-if-not / lif-not

For those that prefer a more Lispy if clause, we have lisp-if, or lif. This only considers
None / nil to be false! All other "false-ish" Python values are considered true.
Conversely, we have lisp-if-not and lif-not in parallel to if and if-not which reverses
the comparison.

=> (lisp-if True "true" "false")
=> (lisp-if False "true" "false")
=> (lisp-if 0 "true" "false")
=> (lisp-if nil "true" "false")
=> (lisp-if None "true" "false")
=> (lisp-if-not nil "true" "false")
=> (lisp-if-not None "true" "false")
=> (lisp-if-not False "true" "false")

; Equivalent but shorter
=> (lif True "true" "false")
=> (lif nil "true" "false")
=> (lif-not None "true" "false")

import is used to import modules, like in Python. There are several ways that import can
be used.

;; Imports each of these modules
;; Python:
;; import sys
;; import os.path
(import sys os.path)

;; Import from a module
;; Python: from os.path import exists, isdir, isfile
(import [os.path [exists isdir isfile]])

;; Import with an alias
;; Python: import sys as systest
(import [sys :as systest])

;; You can list as many imports as you like of different types.
(import [tests.resources [kwtest function-with-a-dash]]
[os.path [exists isdir isfile]]
[sys :as systest])

;; Import all module functions into current namespace
(import [sys [*]])

lambda / fn
lambda and fn can be used to define an anonymous function. The parameters are similar to
defn: the first parameter is vector of parameters and the rest is the body of the
function. lambda returns a new function. In the following example, an anonymous function
is defined and passed to another function for filtering output.

=> (def people [{:name "Alice" :age 20}
... {:name "Bob" :age 25}
... {:name "Charlie" :age 50}
... {:name "Dave" :age 5}])

=> (defn display-people [people filter]
... (for [person people] (if (filter person) (print (:name person)))))

=> (display-people people (fn [person] (< (:age person) 25)))

Just as in normal function definitions, if the first element of the body is a string, it
serves as a docstring. This is useful for giving class methods docstrings.

=> (setv times-three
... (fn [x]
... "Multiplies input by three and returns the result."
... (* x 3)))

This can be confirmed via Python's built-in help function:

=> (help times-three)
Help on function times_three:

Multiplies input by three and returns result

New in version 0.10.2.

last can be used for accessing the last element of a collection:

=> (last [2 4 6])

let is used to create lexically scoped variables. They are created at the beginning of the
let form and cease to exist after the form. The following example showcases this

=> (let [[x 5]] (print x)
... (let [[x 6]] (print x))
... (print x))

The let macro takes two parameters: a vector defining variables and the body which gets
executed. variables is a vector where each element is either a single variable or a vector
defining a variable value pair. In the case of a single variable, it is assigned value
None; otherwise, the supplied value is used.

=> (let [x [y 5]] (print x y))
None 5

list-comp performs list comprehensions. It takes two or three parameters. The first
parameter is the expression controlling the return value, while the second is used to
select items from a list. The third and optional parameter can be used to filter out some
of the items in the list based on a conditional expression. Some examples:

=> (def collection (range 10))
=> (list-comp x [x collection])
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

=> (list-comp (* x 2) [x collection])
[0, 2, 4, 6, 8, 10, 12, 14, 16, 18]

=> (list-comp (* x 2) [x collection] (< x 5))
[0, 2, 4, 6, 8]

not is used in logical expressions. It takes a single parameter and returns a reversed
truth value. If True is given as a parameter, False will be returned, and vice-versa.
Example usage:

=> (not True)

=> (not False)

=> (not None)

or is used in logical expressions. It takes at least two parameters. It will return the
first non-false parameter. If no such value exists, the last parameter will be returned.

=> (or True False)

=> (and False False)

=> (and False 1 True False)

or short-circuits and stops evaluating parameters as soon as the first true value is

=> (or True (print "hello"))

print is used to output on screen. Example usage:

(print "Hello world!")

print always returns None.

quasiquote allows you to quote a form, but also selectively evaluate expressions.
Expressions inside a quasiquote can be selectively evaluated using unquote (~). The
evaluated form can also be spliced using unquote-splice (~@). Quasiquote can be also
written using the backquote (`) symbol.

;; let `qux' be a variable with value (bar baz)
`(foo ~qux)
; equivalent to '(foo (bar baz))
`(foo ~@qux)
; equivalent to '(foo bar baz)

quote returns the form passed to it without evaluating it. quote can alternatively be
written using the apostrophe (') symbol.

=> (setv x '(print "Hello World"))
; variable x is set to expression & not evaluated
=> x
(u'print' u'Hello World')
=> (eval x)
Hello World

require is used to import macros from a given module. It takes at least one parameter
specifying the module which macros should be imported. Multiple modules can be imported
with a single require.

The following example will import macros from module-1 and module-2:

(require module-1 module-2)

rest / cdr
rest and cdr return the collection passed as an argument without the first element:

=> (rest (range 10))
[1, 2, 3, 4, 5, 6, 7, 8, 9]

set-comp is used to create sets. It takes two or three parameters. The first parameter is
for controlling the return value, while the second is used to select items from a
sequence. The third and optional parameter can be used to filter out some of the items in
the sequence based on a conditional expression.

=> (setv data [1 2 3 4 5 2 3 4 5 3 4 5])
=> (set-comp x [x data] (odd? x))
{1, 3, 5}

slice can be used to take a subset of a list and create a new list from it. The form
takes at least one parameter specifying the list to slice. Two optional parameters can be
used to give the start and end position of the subset. If they are not supplied, the
default value of None will be used instead. The third optional parameter is used to
control step between the elements.

slice follows the same rules as its Python counterpart. Negative indices are counted
starting from the end of the list. Some example usage:

=> (def collection (range 10))

=> (slice collection)
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

=> (slice collection 5)
[5, 6, 7, 8, 9]

=> (slice collection 2 8)
[2, 3, 4, 5, 6, 7]

=> (slice collection 2 8 2)
[2, 4, 6]

=> (slice collection -4 -2)
[6, 7]

throw / raise
The throw or raise forms can be used to raise an Exception at runtime. Example usage:

; re-rase the last exception

(throw IOError)
; Throw an IOError

(throw (IOError "foobar"))
; Throw an IOError("foobar")

throw can accept a single argument (an Exception class or instance) or no arguments to
re-raise the last Exception.

The try form is used to start a try / catch block. The form is used as follows:

(catch [e ZeroDivisionError] (print "Division by zero"))
(else (print "no errors"))
(finally (print "all done")))

try must contain at least one catch block, and may optionally include an else or finally
block. If an error is raised with a matching catch block during the execution of
error-prone-function, that catch block will be executed. If no errors are raised, the else
block is executed. The finally block will be executed last regardless of whether or not an
error was raised.

The unless macro is a shorthand for writing an if statement that checks if the given
conditional is False. The following shows the expansion of this macro.

(unless conditional statement)

(if conditional
(do statement))

Within a quasiquoted form, unquote forces evaluation of a symbol. unquote is aliased to
the tilde (~) symbol.

(def name "Cuddles")
(quasiquote (= name (unquote name)))
;=> (u'=' u'name' u'Cuddles')

`(= name ~name)
;=> (u'=' u'name' u'Cuddles')

unquote-splice forces the evaluation of a symbol within a quasiquoted form, much like
unquote. unquote-splice can only be used when the symbol being unquoted contains an
iterable value, as it "splices" that iterable into the quasiquoted form. unquote-splice is
aliased to the ~@ symbol.

(def nums [1 2 3 4])
(quasiquote (+ (unquote-splice nums)))
;=> (u'+' 1L 2L 3L 4L)

`(+ ~@nums)
;=> (u'+' 1L 2L 3L 4L)

when is similar to unless, except it tests when the given conditional is True. It is not
possible to have an else block in a when macro. The following shows the expansion of the

(when conditional statement)

(if conditional (do statement))

while is used to execute one or more blocks as long as a condition is met. The following
example will output "Hello world!" to the screen indefinitely:

(while True (print "Hello world!"))

with is used to wrap the execution of a block within a context manager. The context
manager can then set up the local system and tear it down in a controlled manner. The
archetypical example of using with is when processing files. with can bind context to an
argument or ignore it completely, as shown below:

(with [[arg (expr)]] block)

(with [[(expr)]] block)

(with [[arg (expr)] [(expr)]] block)

The following example will open the NEWS file and print its content to the screen. The
file is automatically closed after it has been processed.

(with [[f (open "NEWS")]] (print (.read f)))

with-decorator is used to wrap a function with another. The function performing the
decoration should accept a single value: the function being decorated, and return a new
function. with-decorator takes a minimum of two parameters: the function performing
decoration and the function being decorated. More than one decorator function can be
applied; they will be applied in order from outermost to innermost, ie. the first
decorator will be the outermost one, and so on. Decorators with arguments are called just
like a function call.

(with-decorator decorator-fun
(defn some-function [] ...)

(with-decorator decorator1 decorator2 ...
(defn some-function [] ...)

(with-decorator (decorator arg) ..
(defn some-function [] ...)

In the following example, inc-decorator is used to decorate the function addition with a
function that takes two parameters and calls the decorated function with values that are
incremented by 1. When the decorated addition is called with values 1 and 1, the end
result will be 4 (1+1 + 1+1).

=> (defn inc-decorator [func]
... (fn [value-1 value-2] (func (+ value-1 1) (+ value-2 1))))
=> (defn inc2-decorator [func]
... (fn [value-1 value-2] (func (+ value-1 2) (+ value-2 2))))

=> (with-decorator inc-decorator (defn addition [a b] (+ a b)))
=> (addition 1 1)
=> (with-decorator inc2-decorator inc-decorator
... (defn addition [a b] (+ a b)))
=> (addition 1 1)

New in version 0.9.12.

with-gensym is used to generate a set of gensym for use in a macro. The following code:

(with-gensyms [a b c]

expands to:

(let [[a (gensym)
[b (gensym)
[c (gensym)]]

Section using-gensym

yield is used to create a generator object that returns one or more values. The generator
is iterable and therefore can be used in loops, list comprehensions and other similar

The function random-numbers shows how generators can be used to generate infinite series
without consuming infinite amount of memory.

=> (defn multiply [bases coefficients]
... (for [[(, base coefficient) (zip bases coefficients)]]
... (yield (* base coefficient))))

=> (multiply (range 5) (range 5))
<generator object multiply at 0x978d8ec>

=> (list-comp value [value (multiply (range 10) (range 10))])
[0, 1, 4, 9, 16, 25, 36, 49, 64, 81]

=> (import random)
=> (defn random-numbers [low high]
... (while True (yield (.randint random low high))))
=> (list-comp x [x (take 15 (random-numbers 1 50))])])
[7, 41, 6, 22, 32, 17, 5, 38, 18, 38, 17, 14, 23, 23, 19]

New in version 0.9.13.


yield-from is used to call a subgenerator. This is useful if you want your coroutine to
be able to delegate its processes to another coroutine, say, if using something fancy like

Hy Core
Core Functions
Usage: (butlast coll)

Returns an iterator of all but the last item in coll.

=> (list (butlast (range 10)))
[0, 1, 2, 3, 4, 5, 6, 7, 8]

=> (list (butlast [1]))

=> (list (butlast []))

=> (import itertools)
=> (list (take 5 (butlast (itertools.count 10))))
[10, 11, 12, 13, 14]

New in version 0.10.0.

Usage: (coll? x)

Returns True if x is iterable and not a string.

=> (coll? [1 2 3 4])

=> (coll? {"a" 1 "b" 2})

=> (coll? "abc")

New in version 0.10.0.

Usage: (cons a b)

Returns a fresh cons cell with car a and cdr b.

=> (setv a (cons 'hd 'tl))

=> (= 'hd (car a))

=> (= 'tl (cdr a))

New in version 0.10.0.

Usage: (cons? foo)

Checks whether foo is a cons cell.

=> (setv a (cons 'hd 'tl))

=> (cons? a)

=> (cons? nil)

=> (cons? [1 2 3])

Usage: (dec x)

Returns one less than x. Equivalent to (- x 1). Raises TypeError if (not (numeric? x)).

=> (dec 3)

=> (dec 0)

=> (dec 12.3)

New in version 0.10.0.

Usage: (disassemble tree &optional [codegen false])

Dump the Python AST for given Hy tree to standard output. If codegen is True, the function
prints Python code instead.

=> (disassemble '(print "Hello World!"))
Expr(value=Call(func=Name(id='print'), args=[Str(s='Hello World!')], keywords=[], starargs=None, kwargs=None))])

=> (disassemble '(print "Hello World!") true)
print('Hello World!')

Usage: (empty? coll)

Returns True if coll is empty. Equivalent to (= 0 (len coll)).

=> (empty? [])

=> (empty? "")

=> (empty? (, 1 2))

New in version 0.10.0.

Usage: (every? pred coll)

Returns True if (pred x) is logical true for every x in coll, otherwise False. Return True
if coll is empty.

=> (every? even? [2 4 6])

=> (every? even? [1 3 5])

=> (every? even? [2 4 5])

=> (every? even? [])

Usage: (float? x)

Returns True if x is a float.

=> (float? 3.2)

=> (float? -2)

Usage: (even? x)

Returns True if x is even. Raises TypeError if (not (numeric? x)).

=> (even? 2)

=> (even? 13)

=> (even? 0)

Usage: (identity x)

Returns the argument supplied to the function.

=> (identity 4)

=> (list (map identity [1 2 3 4]))
[1 2 3 4]

Usage: (inc x)

Returns one more than x. Equivalent to (+ x 1). Raises TypeError if (not (numeric? x)).

=> (inc 3)

=> (inc 0)

=> (inc 12.3)

Usage: (instance? class x)

Returns True if x is an instance of class.

=> (instance? float 1.0)

=> (instance? int 7)

=> (instance? str (str "foo"))

=> (defclass TestClass [object])
=> (setv inst (TestClass))
=> (instance? TestClass inst)

Usage: (integer? x)

Returns True if x is an integer. For Python 2, this is either int or long. For Python 3,
this is int.

=> (integer? 3)

=> (integer? -2.4)

New in version 0.10.1.

Usage: (interleave seq1 seq2 ...)

Returns an iterable of the first item in each of the sequences, then the second, etc.

=> (list (interleave (range 5) (range 100 105)))
[0, 100, 1, 101, 2, 102, 3, 103, 4, 104]

=> (list (interleave (range 1000000) "abc"))
[0, 'a', 1, 'b', 2, 'c']

New in version 0.10.1.

Usage: (interpose item seq)

Returns an iterable of the elements of the sequence separated by the item.

=> (list (interpose "!" "abcd"))
['a', '!', 'b', '!', 'c', '!', 'd']

=> (list (interpose -1 (range 5)))
[0, -1, 1, -1, 2, -1, 3, -1, 4]

Usage: (iterable? x)

Returns True if x is iterable. Iterable objects return a new iterator when (iter x) is
called. Contrast with iterator?.

=> ;; works for strings
=> (iterable? (str "abcde"))

=> ;; works for lists
=> (iterable? [1 2 3 4 5])

=> ;; works for tuples
=> (iterable? (, 1 2 3))

=> ;; works for dicts
=> (iterable? {:a 1 :b 2 :c 3})

=> ;; works for iterators/generators
=> (iterable? (repeat 3))

Usage: (iterator? x)

Returns True if x is an iterator. Iterators are objects that return themselves as an
iterator when (iter x) is called. Contrast with iterable?.

=> ;; doesn't work for a list
=> (iterator? [1 2 3 4 5])

=> ;; but we can get an iter from the list
=> (iterator? (iter [1 2 3 4 5]))

=> ;; doesn't work for dict
=> (iterator? {:a 1 :b 2 :c 3})

=> ;; create an iterator from the dict
=> (iterator? (iter {:a 1 :b 2 :c 3}))

Usage: (list* head &rest tail)

Generates a chain of nested cons cells (a dotted list) containing the arguments. If the
argument list only has one element, return it.

=> (list* 1 2 3 4)
(1 2 3 . 4)

=> (list* 1 2 3 [4])
[1, 2, 3, 4]

=> (list* 1)

=> (cons? (list* 1 2 3 4))

New in version 0.10.0.

Usage: (macroexpand form)

Returns the full macro expansion of form.

=> (macroexpand '(-> (a b) (x y)))
(u'x' (u'a' u'b') u'y')

=> (macroexpand '(-> (a b) (-> (c d) (e f))))
(u'e' (u'c' (u'a' u'b') u'd') u'f')

New in version 0.10.0.

Usage: (macroexpand-1 form)

Returns the single step macro expansion of form.

=> (macroexpand-1 '(-> (a b) (-> (c d) (e f))))
(u'_>' (u'a' u'b') (u'c' u'd') (u'e' u'f'))

New in version 0.10.1.

Usage: (merge-with f &rest maps)

Returns a map that consist of the rest of the maps joined onto first. If a key occurs in
more than one map, the mapping(s) from the latter (left-to-right) will be combined with
the mapping in the result by calling (f val-in-result val-in-latter).

=> (merge-with (fn [x y] (+ x y)) {"a" 10 "b" 20} {"a" 1 "c" 30})
{u'a': 11L, u'c': 30L, u'b': 20L}

Usage: (neg? x)

Returns True if x is less than zero. Raises TypeError if (not (numeric? x)).

=> (neg? -2)

=> (neg? 3)

=> (neg? 0)

Usage: (nil? x)

Returns True if x is nil / None.

=> (nil? nil)

=> (nil? None)

=> (nil? 0)

=> (setf x nil)
=> (nil? x)

=> ;; list.append always returns None
=> (nil? (.append [1 2 3] 4))

Usage: (none? x)

Returns True if x is None.

=> (none? None)

=> (none? 0)

=> (setf x None)
=> (none? x)

=> ;; list.append always returns None
=> (none? (.append [1 2 3] 4))

Usage: (nth coll n &optional [default nil])

Returns the n-th item in a collection, counting from 0. Return the default value, nil, if
out of bounds (unless specified otherwise). Raises ValueError if n is negative.

=> (nth [1 2 4 7] 1)

=> (nth [1 2 4 7] 3)

=> (nil? (nth [1 2 4 7] 5))

=> (nth [1 2 4 7] 5 "default")

=> (nth (take 3 (drop 2 [1 2 3 4 5 6])) 2))

=> (nth [1 2 4 7] -1)
Traceback (most recent call last):
ValueError: Indices for islice() must be None or an integer: 0 <= x <= sys.maxsize.

Usage: (numeric? x)

Returns True if x is a numeric, as defined in Python's numbers.Number class.

=> (numeric? -2)

=> (numeric? 3.2)

=> (numeric? "foo")

Usage: (odd? x)

Returns True if x is odd. Raises TypeError if (not (numeric? x)).

=> (odd? 13)

=> (odd? 2)

=> (odd? 0)

Usage: (pos? x)

Returns True if x is greater than zero. Raises TypeError if (not (numeric? x)).

=> (pos? 3)

=> (pos? -2)

=> (pos? 0)

Usage: (second coll)

Returns the second member of coll. Equivalent to (get coll 1).

=> (second [0 1 2])

New in version 0.10.0.

Usage: (some pred coll)

Returns the first logically-true value of (pred x) for any x in coll, otherwise nil.
Return nil if coll is empty.

=> (some even? [2 4 6])

=> (nil? (some even? [1 3 5]))

=> (nil? (some identity [0 "" []]))

=> (some identity [0 "non-empty-string" []])

=> (nil? (some even? []))

Usage: (string? x)

Returns True if x is a string.

=> (string? "foo")

=> (string? -2)

Usage: (symbol? x)

Returns True if x is a symbol.

=> (symbol? 'foo)

=> (symbol? '[a b c])

Usage: (zero? x)

Returns True if x is zero.

=> (zero? 3)

=> (zero? -2)

=> (zero? 0)

Sequence Functions
Sequence functions can either create or operate on a potentially infinite sequence without
requiring the sequence be fully realized in a list or similar container. They do this by
returning a Python iterator.

We can use the canonical infinite Fibonacci number generator as an example of how to use
some of these functions.

(defn fib []
(setv a 0)
(setv b 1)
(while true
(yield a)
(setv (, a b) (, b (+ a b)))))

Note the (while true ...) loop. If we run this in the REPL,

=> (fib)
<generator object fib at 0x101e642d0>

Calling the function only returns an iterator, but does no work until we consume it.
Trying something like this is not recommend as the infinite loop will run until it
consumes all available RAM, or in this case until I killed it.

=> (list (fib))
[1] 91474 killed hy

To get the first 10 Fibonacci numbers, use take. Note that take also returns a generator,
so I create a list from it.

=> (list (take 10 (fib)))
[0, 1, 1, 2, 3, 5, 8, 13, 21, 34]

To get the Fibonacci number at index 9, (starting from 0):

=> (nth (fib) 9)

Usage: (cycle coll)

Returns an infinite iterator of the members of coll.

=> (list (take 7 (cycle [1 2 3])))
[1, 2, 3, 1, 2, 3, 1]

=> (list (take 2 (cycle [1 2 3])))
[1, 2]

Usage: (distinct coll)

Returns an iterator containing only the unique members in coll.

=> (list (distinct [ 1 2 3 4 3 5 2 ]))
[1, 2, 3, 4, 5]

=> (list (distinct []))

=> (list (distinct (iter [ 1 2 3 4 3 5 2 ])))
[1, 2, 3, 4, 5]

Usage: (drop n coll)

Returns an iterator, skipping the first n members of coll. Raises ValueError if n is

=> (list (drop 2 [1 2 3 4 5]))
[3, 4, 5]

=> (list (drop 4 [1 2 3 4 5]))

=> (list (drop 0 [1 2 3 4 5]))
[1, 2, 3, 4, 5]

=> (list (drop 6 [1 2 3 4 5]))

Usage: (drop-last n coll)

Returns an iterator of all but the last n items in coll. Raises ValueError if n is

=> (list (drop-last 5 (range 10 20)))
[10, 11, 12, 13, 14]

=> (list (drop-last 0 (range 5)))
[0, 1, 2, 3, 4]

=> (list (drop-last 100 (range 100)))

=> (import itertools)
=> (list (take 5 (drop-last 100 (itertools.count 10))))
[10, 11, 12, 13, 14]

Usage: (drop-while pred coll)

Returns an iterator, skipping members of coll until pred is False.

=> (list (drop-while even? [2 4 7 8 9]))
[7, 8, 9]

=> (list (drop-while numeric? [1 2 3 None "a"])))
[None, u'a']

=> (list (drop-while pos? [2 4 7 8 9]))

Usage: (filter pred coll)

Returns an iterator for all items in coll that pass the predicate pred.

See also remove.

=> (list (filter pos? [1 2 3 -4 5 -7]))
[1, 2, 3, 5]

=> (list (filter even? [1 2 3 -4 5 -7]))
[2, -4]

New in version 0.9.12.

Usage: (flatten coll)

Returns a single list of all the items in coll, by flattening all contained lists and/or

=> (flatten [1 2 [3 4] 5])
[1, 2, 3, 4, 5]

=> (flatten ["foo" (, 1 2) [1 [2 3] 4] "bar"])
['foo', 1, 2, 1, 2, 3, 4, 'bar']

Usage: (iterate fn x)

Returns an iterator of x, fn(x), fn(fn(x)), etc.

=> (list (take 5 (iterate inc 5)))
[5, 6, 7, 8, 9]

=> (list (take 5 (iterate (fn [x] (* x x)) 5)))
[5, 25, 625, 390625, 152587890625]

Usage: (read &optional [from-file eof])

Reads the next Hy expression from from-file (defaulting to sys.stdin), and can take a
single byte as EOF (defaults to an empty string). Raises EOFError if from-file ends before
a complete expression can be parsed.

=> (read)
(+ 2 2)
('+' 2 2)
=> (eval (read))
(+ 2 2)

=> (import io)
=> (def buffer (io.StringIO "(+ 2 2)\n(- 2 1)"))
=> (eval (apply read [] {"from_file" buffer}))
=> (eval (apply read [] {"from_file" buffer}))

=> ; assuming "example.hy" contains:
=> ; (print "hello")
=> ; (print "hyfriends!")
=> (with [[f (open "example.hy")]]
... (try
... (while true
... (let [[exp (read f)]]
... (do
... (print "OHY" exp)
... (eval exp))))
... (catch [e EOFError]
... (print "EOF!"))))
OHY ('print' 'hello')
OHY ('print' 'hyfriends!')

Usage: (remove pred coll)

Returns an iterator from coll with elements that pass the predicate, pred, removed.

See also filter.

=> (list (remove odd? [1 2 3 4 5 6 7]))
[2, 4, 6]

=> (list (remove pos? [1 2 3 4 5 6 7]))

=> (list (remove neg? [1 2 3 4 5 6 7]))
[1, 2, 3, 4, 5, 6, 7]

Usage: (repeat x)

Returns an iterator (infinite) of x.

=> (list (take 6 (repeat "s")))
[u's', u's', u's', u's', u's', u's']

Usage: (repeatedly fn)

Returns an iterator by calling fn repeatedly.

=> (import [random [randint]])

=> (list (take 5 (repeatedly (fn [] (randint 0 10)))))
[6, 2, 0, 6, 7]

Usage: (take n coll)

Returns an iterator containing the first n members of coll. Raises ValueError if n is

=> (list (take 3 [1 2 3 4 5]))
[1, 2, 3]

=> (list (take 4 (repeat "s")))
[u's', u's', u's', u's']

=> (list (take 0 (repeat "s")))

Usage: (take-nth n coll)

Returns an iterator containing every n-th member of coll.

=> (list (take-nth 2 [1 2 3 4 5 6 7]))
[1, 3, 5, 7]

=> (list (take-nth 3 [1 2 3 4 5 6 7]))
[1, 4, 7]

=> (list (take-nth 4 [1 2 3 4 5 6 7]))
[1, 5]

=> (list (take-nth 10 [1 2 3 4 5 6 7]))

Usage: (take-while pred coll)

Returns an iterator from coll as long as pred returns True.

=> (list (take-while pos? [ 1 2 3 -4 5]))
[1, 2, 3]

=> (list (take-while neg? [ -4 -3 1 2 5]))
[-4, -3]

=> (list (take-while neg? [ 1 2 3 -4 5]))

New in version 0.9.13.

Usage: (zipwith fn coll ...)

Equivalent to zip, but uses a multi-argument function instead of creating a tuple. If
zipwith is called with N collections, then fn must accept N arguments.

=> (import operator)
=> (list (zipwith operator.add [1 2 3] [4 5 6]))
[5, 7, 9]

Reader Macros
Reader macros gives Lisp the power to modify and alter syntax on the fly. You don't want
Polish notation? A reader macro can easily do just that. Want Clojure's way of having a
regex? Reader macros can also do this easily.

=> (defreader ^ [expr] (print expr))
=> #^(1 2 3 4)
(1 2 3 4)
=> #^"Hello"
=> #^1+2+3+4+3+2

Hy has no literal for tuples. Lets say you dislike (, ...) and want something else. This
is a problem reader macros are able to solve in a neat way.

=> (defreader t [expr] `(, ~@expr))
=> #t(1 2 3)
(1, 2, 3)

You could even do it like Clojure and have a literal for regular expressions!

=> (import re)
=> (defreader r [expr] `(re.compile ~expr))
=> #r".*"
<_sre.SRE_Pattern object at 0xcv7713ph15#>

defreader takes a single character as symbol name for the reader macro; anything longer
will return an error. Implementation-wise, defreader expands into a lambda covered with a
decorator. This decorator saves the lambda in a dictionary with its module name and

=> (defreader ^ [expr] (print expr))
;=> (with_decorator (hy.macros.reader ^) (fn [expr] (print expr)))

# expands into (dispatch_reader_macro ...) where the symbol and expression is passed to
the correct function.

=> #^()
;=> (dispatch_reader_macro ^ ())
=> #^"Hello"

Because of a limitation in Hy's lexer and parser, reader macros can't redefine defined
syntax such as ()[]{}. This will most likely be addressed in the future.

Internal Hy Documentation
These bits are mostly useful for folks who hack on Hy itself, but can also be used for
those delving deeper in macro programming.

Hy Models
Introduction to Hy Models
Hy models are a very thin layer on top of regular Python objects, representing Hy source
code as data. Models only add source position information, and a handful of methods to
support clean manipulation of Hy source code, for instance in macros. To achieve that
goal, Hy models are mixins of a base Python class and HyObject.

hy.models.HyObject is the base class of Hy models. It only implements one method, replace,
which replaces the source position of the current object with the one passed as argument.
This allows us to keep track of the original position of expressions that get modified by
macros, be that in the compiler or in pure hy macros.

HyObject is not intended to be used directly to instantiate Hy models, but only as a mixin
for other classes.

Compound Models
Parenthesized and bracketed lists are parsed as compound models by the Hy parser.

hy.models.list.HyList is the base class of "iterable" Hy models. Its basic use is to
represent bracketed [] lists, which, when used as a top-level expression, translate to
Python list literals in the compilation phase.

Adding a HyList to another iterable object reuses the class of the left-hand-side object,
a useful behavior when you want to concatenate Hy objects in a macro, for instance.

hy.models.expression.HyExpression inherits HyList for parenthesized () expressions. The
compilation result of those expressions depends on the first element of the list: the
compiler dispatches expressions between compiler special-forms, user-defined macros, and
regular Python function calls.

hy.models.dict.HyDict inherits HyList for curly-bracketed {} expressions, which compile
down to a Python dictionary literal.

The decision of using a list instead of a dict as the base class for HyDict allows easier
manipulation of dicts in macros, with the added benefit of allowing compound expressions
as dict keys (as, for instance, the HyExpression Python class isn't hashable).

Atomic Models
In the input stream, double-quoted strings, respecting the Python notation for strings,
are parsed as a single token, which is directly parsed as a HyString.

An uninterrupted string of characters, excluding spaces, brackets, quotes, double-quotes
and comments, is parsed as an identifier.

Identifiers are resolved to atomic models during the parsing phase in the following order:

· HyInteger

· HyFloat

· HyComplex (if the atom isn't a bare j)

· HyKeyword (if the atom starts with :)

· HySymbol

hy.models.string.HyString is the base class of string-equivalent Hy models. It also
represents double-quoted string literals, "", which compile down to unicode string
literals in Python. HyStrings inherit unicode objects in Python 2, and string objects in
Python 3 (and are therefore not encoding-dependent).

HyString based models are immutable.

Hy literal strings can span multiple lines, and are considered by the parser as a single
unit, respecting the Python escapes for unicode strings.

Numeric Models
hy.models.integer.HyInteger represents integer literals (using the long type on Python 2,
and int on Python 3).

hy.models.float.HyFloat represents floating-point literals.

hy.models.complex.HyComplex represents complex literals.

Numeric models are parsed using the corresponding Python routine, and valid numeric python
literals will be turned into their Hy counterpart.

hy.models.symbol.HySymbol is the model used to represent symbols in the Hy language. It
inherits HyString.

HySymbol objects are mangled in the parsing phase, to help Python interoperability:

· Symbols surrounded by asterisks (*) are turned into uppercase;

· Dashes (-) are turned into underscores (_);

· One trailing question mark (?) is turned into a leading is_.

Caveat: as the mangling is done during the parsing phase, it is possible to
programmatically generate HySymbols that can't be generated with Hy source code. Such a
mechanism is used by gensym to generate "uninterned" symbols.

hy.models.keyword.HyKeyword represents keywords in Hy. Keywords are symbols starting with
a :. The class inherits HyString.

To distinguish HyKeywords from HySymbols, without the possibility of (involuntary)
clashes, the private-use unicode character "\uFDD0" is prepended to the keyword literal
before storage.

Cons Cells
hy.models.cons.HyCons is a representation of Python-friendly cons cells. Cons cells are
especially useful to mimic features of "usual" LISP variants such as Scheme or Common

A cons cell is a 2-item object, containing a car (head) and a cdr (tail). In some Lisp
variants, the cons cell is the fundamental building block, and S-expressions are actually
represented as linked lists of cons cells. This is not the case in Hy, as the usual
expressions are made of Python lists wrapped in a HyExpression. However, the HyCons
mimicks the behavior of "usual" Lisp variants thusly:

· (cons something nil) is (HyExpression [something])

· (cons something some-list) is ((type some-list) (+ [something] some-list)) (if
some-list inherits from list).

· (get (cons a b) 0) is a

· (slice (cons a b) 1) is b

Hy supports a dotted-list syntax, where '(a . b) means (cons 'a 'b) and '(a b . c) means
(cons 'a (cons 'b 'c)). If the compiler encounters a cons cell at the top level, it raises
a compilation error.

HyCons wraps the passed arguments (car and cdr) in Hy types, to ease the manipulation of
cons cells in a macro context.

Hy Internal Theory
The Hy internals work by acting as a front-end to Python bytecode, so that Hy itself
compiles down to Python Bytecode, allowing an unmodified Python runtime to run Hy code,
without even noticing it.

The way we do this is by translating Hy into an internal Python AST datastructure, and
building that AST down into Python bytecode using modules from the Python standard
library, so that we don't have to duplicate all the work of the Python internals for every
single Python release.

Hy works in four stages. The following sections will cover each step of Hy from source to

Steps 1 and 2: Tokenizing and Parsing
The first stage of compiling Hy is to lex the source into tokens that we can deal with. We
use a project called rply, which is a really nice (and fast) parser, written in a subset
of Python called rpython.

The lexing code is all defined in hy.lex.lexer. This code is mostly just defining the Hy
grammar, and all the actual hard parts are taken care of by rply -- we just define
"callbacks" for rply in hy.lex.parser, which takes the tokens generated, and returns the
Hy models.

You can think of the Hy models as the "AST" for Hy, it's what Macros operate on
(directly), and it's what the compiler uses when it compiles Hy down.

Section Hy Models for more information on Hy models and what they mean.

Step 3: Hy Compilation to Python AST
This is where most of the magic in Hy happens. This is where we take Hy AST (the models),
and compile them into Python AST. A couple of funky things happen here to work past a few
problems in AST, and working in the compiler is some of the most important work we do

The compiler is a bit complex, so don't feel bad if you don't grok it on the first shot,
it may take a bit of time to get right.

The main entry-point to the Compiler is HyASTCompiler.compile. This method is invoked, and
the only real "public" method on the class (that is to say, we don't really promise the
API beyond that method).

In fact, even internally, we don't recurse directly hardly ever, we almost always force
the Hy tree through compile, and will often do this with sub-elements of an expression
that we have. It's up to the Type-based dispatcher to properly dispatch sub-elements.

All methods that preform a compilation are marked with the @builds() decorator. You can
either pass the class of the Hy model that it compiles, or you can use a string for
expressions. I'll clear this up in a second.

First Stage Type-Dispatch
Let's start in the compile method. The first thing we do is check the Type of the thing
we're building. We look up to see if we have a method that can build the type() that we
have, and dispatch to the method that can handle it. If we don't have any methods that can
build that type, we raise an internal Exception.

For instance, if we have a HyString, we have an almost 1-to-1 mapping of Hy AST to Python
AST. The compile_string method takes the HyString, and returns an ast.Str() that's
populated with the correct line-numbers and content.

If we get a HyExpression, we'll attempt to see if this is a known Macro, and push to have
it expanded by invoking hy.macros.macroexpand, then push the result back into

Second Stage Expression-Dispatch
The only special case is the HyExpression, since we need to create different AST depending
on the special form in question. For instance, when we hit an (if true true false), we
need to generate a ast.If, and properly compile the sub-nodes. This is where the @builds()
with a String as an argument comes in.

For the compile_expression (which is defined with an @builds(HyExpression)) will dispatch
based on the string of the first argument. If, for some reason, the first argument is not
a string, it will properly handle that case as well (most likely by raising an Exception).

If the String isn't known to Hy, it will default to create an ast.Call, which will try to
do a runtime call (in Python, something like foo()).

Issues Hit with Python AST
Python AST is great; it's what's enabled us to write such a powerful project on top of
Python without having to fight Python too hard. Like anything, we've had our fair share of
issues, and here's a short list of the common ones you might run into.

Python differentiates between Statements and Expressions.

This might not sound like a big deal -- in fact, to most Python programmers, this will
shortly become a "Well, yeah" moment.

In Python, doing something like:

print for x in range(10): pass, because print prints expressions, and for isn't an
expression, it's a control flow statement. Things like 1 + 1 are Expressions, as is lambda
x: 1 + x, but other language features, such as if, for, or while are statements.

Since they have no "value" to Python, this makes working in Hy hard, since doing something
like (print (if true true false)) is not just common, it's expected.

As a result, we auto-mangle things using a Result object, where we offer up any ast.stmt
that need to get run, and a single ast.expr that can be used to get the value of whatever
was just run. Hy does this by forcing assignment to things while running.

As example, the Hy:

(print (if true true false))

Will turn into:

if True:
_mangled_name_here = True
_mangled_name_here = False

print _mangled_name_here

OK, that was a bit of a lie, since we actually turn that statement into:

print True if True else False

By forcing things into an ast.expr if we can, but the general idea holds.

Step 4: Python Bytecode Output and Runtime
After we have a Python AST tree that's complete, we can try and compile it to Python
bytecode by pushing it through eval. From here on out, we're no longer in control, and
Python is taking care of everything. This is why things like Python tracebacks, pdb and
django apps work.

Hy Macros
Using gensym for Safer Macros
When writing macros, one must be careful to avoid capturing external variables or using
variable names that might conflict with user code.

We will use an example macro nif (see
http://letoverlambda.com/index.cl/guest/chap3.html#sec_5 for a more complete description.)
nif is an example, something like a numeric if, where based on the expression, one of the
3 forms is called depending on if the expression is positive, zero or negative.

A first pass might be something like:

(defmacro nif [expr pos-form zero-form neg-form]
`(let [[obscure-name ~expr]]
(cond [(pos? obscure-name) ~pos-form]
[(zero? obscure-name) ~zero-form]
[(neg? obscure-name) ~neg-form])))

where obsure-name is an attempt to pick some variable name as not to conflict with other
code. But of course, while well-intentioned, this is no guarantee.

The method gensym is designed to generate a new, unique symbol for just such an occasion.
A much better version of nif would be:

(defmacro nif [expr pos-form zero-form neg-form]
(let [[g (gensym)]]
`(let [[~g ~expr]]
(cond [(pos? ~g) ~pos-form]
[(zero? ~g) ~zero-form]
[(neg? ~g) ~neg-form]))))

This is an easy case, since there is only one symbol. But if there is a need for several
gensym's there is a second macro with-gensyms that basically expands to a series of let

(with-gensyms [a b c]

expands to:

(let [[a (gensym)
[b (gensym)
[c (gensym)]]

so our re-written nif would look like:

(defmacro nif [expr pos-form zero-form neg-form]
(with-gensyms [g]
`(let [[~g ~expr]]
(cond [(pos? ~g) ~pos-form]
[(zero? ~g) ~zero-form]
[(neg? ~g) ~neg-form]))))

Finally, though we can make a new macro that does all this for us. defmacro/g! will take
all symbols that begin with g! and automatically call gensym with the remainder of the
symbol. So g!a would become (gensym "a").

Our final version of nif, built with defmacro/g! becomes:

(defmacro/g! nif [expr pos-form zero-form neg-form]
`(let [[~g!res ~expr]]
(cond [(pos? ~g!res) ~pos-form]
[(zero? ~g!res) ~zero-form]
[(neg? ~g!res) ~neg-form]))))

Checking Macro Arguments and Raising Exceptions
Hy Compiler Built-Ins



Anaphoric Macros
New in version 0.9.12.

The anaphoric macros module makes functional programming in Hy very concise and easy to
An anaphoric macro is a type of programming macro that deliberately captures some form
supplied to the macro which may be referred to by an anaphor (an expression referring
to another). — Wikipedia (http://en.wikipedia.org/wiki/Anaphoric_macro)

Usage: (ap-if (foo) (print it))

Evaluates the first form for truthiness, and bind it to it in both the true and false

Usage: (ap-each [1 2 3 4 5] (print it))

Evaluate the form for each element in the list for side-effects.

Usage: (ap-each-while list pred body)

Evaluate the form for each element where the predicate form returns True.

=> (ap-each-while [1 2 3 4 5 6] (< it 4) (print it))

Usage: (ap-map form list)

The anaphoric form of map works just like regular map except that instead of a function
object it takes a Hy form. The special name it is bound to the current object from the
list in the iteration.

=> (list (ap-map (* it 2) [1 2 3]))
[2, 4, 6]

Usage: (ap-map-when predfn rep list)

Evaluate a mapping over the list using a predicate function to determin when to apply the

=> (list (ap-map-when odd? (* it 2) [1 2 3 4]))
[2, 2, 6, 4]

=> (list (ap-map-when even? (* it 2) [1 2 3 4]))
[1, 4, 3, 8]

Usage: (ap-filter form list)

As with ap-map we take a special form instead of a function to filter the elements of the
list. The special name it is bound to the current element in the iteration.

=> (list (ap-filter (> (* it 2) 6) [1 2 3 4 5]))
[4, 5]

Usage: (ap-reject form list)

This function does the opposite of ap-filter, it rejects the elements passing the
predicate . The special name it is bound to the current element in the iteration.

=> (list (ap-reject (> (* it 2) 6) [1 2 3 4 5]))
[1, 2, 3]

Usage (ap-dotimes n body)

This function evaluates the body n times, with the special variable it bound from 0 to
1-n. It is useful for side-effects.

=> (setv n [])
=> (ap-dotimes 3 (.append n it))
=> n
[0, 1, 2]

Usage (ap-first predfn list)

This function returns the first element that passes the predicate or None, with the
special variable it bound to the current element in iteration.

=>(ap-first (> it 5) (range 10))

Usage (ap-last predfn list)

This function returns the last element that passes the predicate or None, with the special
variable it bound to the current element in iteration.

=>(ap-last (> it 5) (range 10))

Usage (ap-reduce form list &optional initial-value)

This function returns the result of applying form to the first 2 elements in the body and
applying the result and the 3rd element etc. until the list is exhausted. Optionally an
initial value can be supplied so the function will be applied to initial value and the
first element instead. This exposes the element being iterated as it and the current
accumulated value as acc.

=>(ap-reduce (+ it acc) (range 10))

New in version 0.10.0.

The loop / recur macro gives programmers a simple way to use tail-call optimization (TCO)
in their Hy code.
A tail call is a subroutine call that happens inside another procedure as its final
action; it may produce a return value which is then immediately returned by the calling
procedure. If any call that a subroutine performs, such that it might eventually lead
to this same subroutine being called again down the call chain, is in tail position,
such a subroutine is said to be tail-recursive, which is a special case of recursion.
Tail calls are significant because they can be implemented without adding a new stack
frame to the call stack. Most of the frame of the current procedure is not needed any
more, and it can be replaced by the frame of the tail call. The program can then jump
to the called subroutine. Producing such code instead of a standard call sequence is
called tail call elimination, or tail call optimization. Tail call elimination allows
procedure calls in tail position to be implemented as efficiently as goto statements,
thus allowing efficient structured programming. — Wikipedia (‐

loop establishes a recursion point. With loop, recur rebinds the variables set in the
recursion point and sends code execution back to that recursion point. If recur is used in
a non-tail position, an exception is thrown.

Usage: (loop bindings &rest body)


(require hy.contrib.loop)

(defn factorial [n]
(loop [[i n] [acc 1]]
(if (zero? i)
(recur (dec i) (* acc i)))))

(factorial 1000)

New in version 0.10.0.

defmulti lets you arity-overload a function by the given number of args and/or kwargs.
Inspired by Clojure's take on defn.

=> (require hy.contrib.multi)
=> (defmulti fun
... ([a] "a")
... ([a b] "a b")
... ([a b c] "a b c"))
=> (fun 1)
=> (fun 1 2)
"a b"
=> (fun 1 2 3)
"a b c"


Join our Hyve!
Please come hack on Hy!

Please come hang out with us on #hy on irc.freenode.net!

Please talk about it on Twitter with the #hy hashtag!

Please blog about it!

Please don't spraypaint it on your neighbor's fence (without asking nicely)!

Do this:

1. Create a virtual environment:

$ virtualenv venv

and activate it:

$ . venv/bin/activate

or use virtualenvwrapper to create and manage your virtual environment:

$ mkvirtualenv hy
$ workon hy

2. Get the source code:

$ git clone https://github.com/hylang/hy.git

or use your fork:

$ git clone [email protected]:<YOUR_USERNAME>/hy.git

3. Install for hacking:

$ cd hy/
$ pip install -e .

4. Install other develop-y requirements:

$ pip install -r requirements-dev.txt

5. Do awesome things; make someone shriek in delight/disgust at what you have wrought.

Tests are located in tests/. We use nose.

To run the tests:

$ nosetests

Write tests---tests are good!

Also, it is good to run the tests for all the platforms supported and for PEP 8 compliant
code. You can do so by running tox:

$ tox

Documentation is located in docs/. We use Sphinx.

To build the docs in HTML:

$ cd docs
$ make html

Write docs---docs are good! Even this doc!

Contributions are welcome & greatly appreciated, every little bit helps in making Hy more

Pull requests are great! We love them; here is a quick guide:

· Fork the repo and create a topic branch for a feature/fix. Avoid making changes directly
on the master branch.

· All incoming features should be accompanied with tests.

· Before you submit a PR, please run the tests and check your code against the style
guide. You can do both of these things at once:

$ make d

· Make commits into logical units, so that it is easier to track & navigate later. Before
submitting a PR, try squashing the commits into changesets that are easy to come back to
later. Also, make sure you don't leave spurious whitespace in the changesets; this
avoids creation of whitespace fix commits later.

· As far as commit messages go, try to adhere to the following:

· Try sticking to the 50 character limit for the first line of Git commit messages.

· For more detail/explainations, follow this up with a blank line and continue
describing the commit in detail.

· Finally, add yourself to the AUTHORS file (as a separate commit): you deserve it :)

· All incoming changes need to be acked by 2 different members of Hylang's core team.
Additional review is clearly welcome, but we need a minimum of 2 signoffs for any

· If a core member is sending in a PR, please find 2 core members that doesn't include the
PR submitter. The idea here is that one can work with the PR author, and a second acks
the entire change set.

· For documentation & other trivial changes, we're good to merge after one ACK. We've got
low coverage, so it'd be great to keep that barrier low.

Core Team
The core development team of Hy consists of following developers:

· Julien Danjou

· Morten Linderud

· J Kenneth King

· Gergely Nagy

· Tuukka Turto

· Karen Rustad

· Abhishek L

· Christopher Allan Webber

· Konrad Hinsen

· Will Kahn-Greene

· Paul Tagliamonte

· Nicolas Dandrimont

· Bob Tolbert

· Berker Peksag

· Clinton N. Dreisbach

· han semaj

Use hy online using onworks.net services

Free Servers & Workstations

Download Windows & Linux apps

  • 1
    ATTENTION: Cumulative update 2.4.3 has
    been released!! The update works for any
    previous 2.x.x version. If upgrading
    from version v1.x.x, please download and
    Download XISMuS
  • 2
    Modular headtracking program that
    supports multiple face-trackers, filters
    and game-protocols. Among the trackers
    are the SM FaceAPI, AIC Inertial Head
    Tracker ...
    Download facetracknoir
  • 3
    PHP QR Code
    PHP QR Code
    PHP QR Code is open source (LGPL)
    library for generating QR Code,
    2-dimensional barcode. Based on
    libqrencode C library, provides API for
    creating QR Code barc...
    Download PHP QR Code
  • 4
    Freeciv is a free turn-based
    multiplayer strategy game, in which each
    player becomes the leader of a
    civilization, fighting to obtain the
    ultimate goal: to bec...
    Download Freeciv
  • 5
    Cuckoo Sandbox
    Cuckoo Sandbox
    Cuckoo Sandbox uses components to
    monitor the behavior of malware in a
    Sandbox environment; isolated from the
    rest of the system. It offers automated
    analysis o...
    Download Cuckoo Sandbox
  • 6
    Play YouTube video on LMS (porting of
    Triode's to YouTbe API v3) This is
    an application that can also be fetched
    Download LMS-YouTube
  • More »

Linux commands

  • 1
    a2crd - attempts the conversion of
    lyrics file into chordii input ...
    Run a2crd
  • 2
    a2j - Wrapper script to simulate
    a2jmidid's non-DBUS behaviour though
    a2jmidid actually being in DBUS mode ...
    Run a2j
  • 3
    coqdoc - A documentation tool for the
    Coq proof assistant ...
    Run coqdoc
  • 4
    coqide - The Coq Proof Assistant
    graphical interface ...
    Run coqide
  • 5
    g.gisenv - Outputs and modifies the
    user�s current GRASS variable settings.
    Prints all defined GRASS variables if no
    option is given. KEYWORDS: general,
    Run g.gisenvgrass
  • 6
    g.gui - Launches a GRASS graphical user
    interface (GUI) session. Optionally
    updates default user interface settings.
    KEYWORDS: general, GUI, user interface ...
    Run g.guigrass
  • More »