Sometimes I want to pass argument-value pairs to a higher-order function, where the value I should pass is determined by the argument I pass. I want to be able to pass the argument without explicitly specifying the accompanying value. In particular, I'm interested in the case where the argument is itself a function.
Generic Example:
Here's a very generic example, where my-foo and my-bar are functions that I'm passing to higher-foo:
(higher-foo my-foo :option4 args) ;good
(higher-foo my-bar :option13 args) ;good
(higher-foo my-foo :option13 args) ;how stupid are you?! my-foo requires :option4!
Question: Is there a "standard" method for making :option4 or :option13 to be inferable by higher-foo so that I can just write (higher-foo my-foo) and (higher-foo my-bar)?
More Specific Example:
Bear in mind that there are better alternatives to the following code, but I'm just trying to put forward a concrete example of what I'm talking about:
(defn seq-has? [f n someseq]
(every? (partial apply f)
(partition n 1 someseq)))
(defn monotonicity [a b]
(<= a b))
(defn generalized-fib [a b c]
(= c (+ a b)))
(seq-has? monotonicity 2 someseq) should return true if the sequence is monotonic, false otherwise. (seq-has? generalized-fib 3 someseq) should return true if the sequence follows the generalized Fibonacci form, false otherwise.
But the "2" and "3" bother me. I could have an arbitrary number of properties to test for, and I don't want to have to remember the appropriate "magic numbers" for such calls.
Note: I know of two ways to do this, and for my own personal use, I suppose they both work. But I'm interested in what is idiomatic or considered best practice in the community. I'll post my answers, but I'm hoping there are more solutions.
Just make the predicate function itself take variadic arguments, and have it do the partitioning / recurring. Your monotonic? for instance already exists in core, and is called <=
(<= 1 2 4 5)
=> true
(<= 1 2 1 5)
=> false
Here's the source for the 1, 2 and variadic arg versions:
(source <=)
(defn <=
"Returns non-nil if nums are in monotonically non-decreasing order,
otherwise false."
{:inline (fn [x y] `(. clojure.lang.Numbers (lte ~x ~y)))
:inline-arities #{2}
:added "1.0"}
([x] true)
([x y] (. clojure.lang.Numbers (lte x y)))
([x y & more]
(if (<= x y)
(if (next more)
(recur y (first more) (next more))
(<= y (first more)))
false)))
You can make a fib? work the same way, have it take variadic arguments and recur over triples:
(defn fib?
[a b & [c & r]]
(if (= c (+ a b))
(if r
(recur b c r)
true)
false))
(fib? 0 1 1)
=> true
(fib? 2 3 5 8 13)
=> true
Since you are asking for a standard way how a function determines a not passed argument from one argument:
(defn f
([arg0] (case a :foo (f a :bar)
:baz (f a :quux)))
([arg0 arg1] ...))
Depending on your use case a different dispatch construct than case may be a better fit.
For your generic example this implies that higher-foo should determine the correct :option in the desired overload like demonstrated above.
In your specific example, you can't determine the n from the passed function. You need a more specific datastructure:
(defn seq-has? [{:keys [f n]} s]
(every? (partial apply f)
(partition n 1 s)))
(def monotonicity
{:f <=
:n 2})
(def generalized-fib
{:f #(= (+ %1 %2) %3)
:n 3})
(seq-has? monotonicity [1 2 3])
;; => true
This solution seems like a hack to me. Is it considered common/idiomatic? Use meta-data on the functions that define the property you are looking for:
(defn higher-foo [foo & args]
(apply foo (:option (meta foo))
args))
(def my-foo
(with-meta
(fn [a b] (println "I'm doing something cool"))
{:option :option4}))
;using it:
user=> (higher-foo my-foo arg)
Related
I'm trying to estimate the mean distance of all pairs of points in a unit square.
This transducer returns a vector of the distances of x randomly selected pairs of points, but the final step would be to take the mean of all values in that vector. Is there a way to use mean as the final reducing function (or to include it in the composition)?
(defn square [x] (* x x))
(defn mean [x] (/ (reduce + x) (count x)))
(defn xform [iterations]
(comp
(partition-all 4)
(map #(Math/sqrt (+ (square (- (first %) (nth % 1)))
(square (- (nth % 2) (nth % 3))))))
(take iterations)))
(transduce (xform 5) conj (repeatedly #(rand)))
[0.5544757422041136
0.4170515673848907
0.7457675423415904
0.5560901974277822
0.6053573945754688]
(transduce (xform 5) mean (repeatedly #(rand)))
Execution error (ArityException) at test.core/eval19667 (form-init9118116578029918666.clj:562).
Wrong number of args (0) passed to: test.core/mean
If you implement your mean function differently, you won't have to collect all the values before computing the mean. Here is how you can implement it, based on this Java code:
(defn mean
([] [0 1]) ;; <-- Construct an empty accumulator
([[mu n]] mu) ;; <-- Get the mean (final step)
([[mu n] x] ;; <-- Accumulate a value to the mean
[(+ mu (/ (- x mu) n)) (inc n)]))
And you use it like this:
(transduce identity mean [1 2 3 4])
;; => 5/2
or like this:
(transduce (xform 5) mean (repeatedly #(rand)))
;; => 0.582883812837961
From the docs of transduce:
If init is not supplied, (f) will be called to produce it. f should be
a reducing step function that accepts both 1 and 2 arguments, if it
accepts only 2 you can add the arity-1 with 'completing'.
To disect this:
Your function needs 0-arity to produce an initial value -- so conj
is fine (it produces an empty vector).
You need to provide a 2-arity function to do the actual redudcing
-- again conj is fine here
You need to provide a 1-arity function to finalize - here you want
your mean.
So as the docs suggest, you can use completing to just provide that:
(transduce (xform 5) (completing conj mean) (repeatedly #(rand)))
; → 0.4723186070904141
If you look at the source of completing you will see how it produces
all of this:
(defn completing
"Takes a reducing function f of 2 args and returns a fn suitable for
transduce by adding an arity-1 signature that calls cf (default -
identity) on the result argument."
{:added "1.7"}
([f] (completing f identity))
([f cf]
(fn
([] (f))
([x] (cf x))
([x y] (f x y)))))
I have a function that produces lazy-sequences called a-function.
If I run the code:
(map a-function a-sequence-of-values)
it returns a lazy sequence as expected.
But when I run the code:
(mapcat a-function a-sequence-of-values)
it breaks the lazyness of my function. In fact it turns that code into
(apply concat (map a-function a-sequence-of-values))
So it needs to realize all the values from the map before concatenating those values.
What I need is a function that concatenates the result of a map function on demand without realizing all the map beforehand.
I can hack a function for this:
(defn my-mapcat
[f coll]
(lazy-seq
(if (not-empty coll)
(concat
(f (first coll))
(my-mapcat f (rest coll))))))
But I can't believe that clojure doesn't have something already done. Do you know if clojure has such feature? Only a few people and I have the same problem?
I also found a blog that deals with the same issue: http://clojurian.blogspot.com.br/2012/11/beware-of-mapcat.html
Lazy-sequence production and consumption is different than lazy evaluation.
Clojure functions do strict/eager evaluation of their arguments. Evaluation of an argument that is or that yields a lazy sequence does not force realization of the yielded lazy sequence in and of itself. However, any side effects caused by evaluation of the argument will occur.
The ordinary use case for mapcat is to concatenate sequences yielded without side effects. Therefore, it hardly matters that some of the arguments are eagerly evaluated because no side effects are expected.
Your function my-mapcat imposes additional laziness on the evaluation of its arguments by wrapping them in thunks (other lazy-seqs). This can be useful when significant side effects - IO, significant memory consumption, state updates - are expected. However, the warning bells should probably be going off in your head if your function is doing side effects and producing a sequence to be concatenated that your code probably needs refactoring.
Here is similar from algo.monads
(defn- flatten*
"Like #(apply concat %), but fully lazy: it evaluates each sublist
only when it is needed."
[ss]
(lazy-seq
(when-let [s (seq ss)]
(concat (first s) (flatten* (rest s))))))
Another way to write my-mapcat:
(defn my-mapcat [f coll] (for [x coll, fx (f x)] fx))
Applying a function to a lazy sequence will force realization of a portion of that lazy sequence necessary to satisfy the arguments of the function. If that function itself produces lazy sequences as a result, those are not realized as a matter of course.
Consider this function to count the realized portion of a sequence
(defn count-realized [s]
(loop [s s, n 0]
(if (instance? clojure.lang.IPending s)
(if (and (realized? s) (seq s))
(recur (rest s) (inc n))
n)
(if (seq s)
(recur (rest s) (inc n))
n))))
Now let's see what's being realized
(let [seq-of-seqs (map range (list 1 2 3 4 5 6))
concat-seq (apply concat seq-of-seqs)]
(println "seq-of-seqs: " (count-realized seq-of-seqs))
(println "concat-seq: " (count-realized concat-seq))
(println "seqs-in-seq: " (mapv count-realized seq-of-seqs)))
;=> seq-of-seqs: 4
; concat-seq: 0
; seqs-in-seq: [0 0 0 0 0 0]
So, 4 elements of the seq-of-seqs got realized, but none of its component sequences were realized nor was there any realization in the concatenated sequence.
Why 4? Because the applicable arity overloaded version of concat takes 4 arguments [x y & xs] (count the &).
Compare to
(let [seq-of-seqs (map range (list 1 2 3 4 5 6))
foo-seq (apply (fn foo [& more] more) seq-of-seqs)]
(println "seq-of-seqs: " (count-realized seq-of-seqs))
(println "seqs-in-seq: " (mapv count-realized seq-of-seqs)))
;=> seq-of-seqs: 2
; seqs-in-seq: [0 0 0 0 0 0]
(let [seq-of-seqs (map range (list 1 2 3 4 5 6))
foo-seq (apply (fn foo [a b c & more] more) seq-of-seqs)]
(println "seq-of-seqs: " (count-realized seq-of-seqs))
(println "seqs-in-seq: " (mapv count-realized seq-of-seqs)))
;=> seq-of-seqs: 5
; seqs-in-seq: [0 0 0 0 0 0]
Clojure has two solutions to making the evaluation of arguments lazy.
One is macros. Unlike functions, macros do not evaluate their arguments.
Here's a function with a side effect
(defn f [n] (println "foo!") (repeat n n))
Side effects are produced even though the sequence is not realized
user=> (def x (concat (f 1) (f 2)))
foo!
foo!
#'user/x
user=> (count-realized x)
0
Clojure has a lazy-cat macro to prevent this
user=> (def y (lazy-cat (f 1) (f 2)))
#'user/y
user=> (count-realized y)
0
user=> (dorun y)
foo!
foo!
nil
user=> (count-realized y)
3
user=> y
(1 2 2)
Unfortunately, you cannot apply a macro.
The other solution to delay evaluation is wrap in thunks, which is exactly what you've done.
Your premise is wrong. Concat is lazy, apply is lazy if its first argument is, and mapcat is lazy.
user> (class (mapcat (fn [x y] (println x y) (list x y)) (range) (range)))
0 0
1 1
2 2
3 3
clojure.lang.LazySeq
note that some of the initial values are evaluated (more on this below), but clearly the whole thing is still lazy (or the call would never have returned, (range) returns an endless sequence, and will not return when used eagerly).
The blog you link to is about the danger of recursively using mapcat on a lazy tree, because it is eager on the first few elements (which can add up in a recursive application).
I am trying to delay various calculations. I have functions of the following form,
(defn a-fn [a b]
(let [a (if (fn? a)
a
#(identity a))
b (if (fn? b)
b
#(identity b))]
(+ (a) (b))))
this allows me to pass a-fn, a value or a function that returns the value,
(a-fn 1 2)
(defn x [] 1)
(defn y [] 2)
(a-fn x y)
(a-fn x 2)
what I do is build a list of functions (like the one above) to operate on some data, fns may use other fns to retrieve their arguments or in some cases things don't change and they are assigned values as arguments. I was wondering is there a better way to achive this kind of behavior?
You can use delay and force:
user=> (defn a-fn [a b] (+ (force a) (force b)))
#'user/a-fn
user=> (a-fn 1 2)
3
user=> (def x (delay 1))
#'user/x
user=> (def y (delay 2))
#'user/y
user=> (a-fn x y)
3
user=> (a-fn x 2)
3
If you try something like (delay (prn :hello) 1) to test when the computation is done, note that printing the Delay object forces it; so (def x (delay ...)) is safe, but typing a plain (delay ...) in the REPL prompt is not.
There might be a more elegant way to do what you want, but here's at least a more generic version of it:
(defn delayed [reducer & fs]
(apply reducer (for [f fs] (if (fn? f) (f) f))))
(def a-fn (partial delayed +))
So delayed takes an arbitrary function and a list of function/values. If expands all the args and then applies the function to them. Then we use partial to define your a-fn using +:
user=> (a-fn 1 2)
3
user=> (a-fn (constantly 1) 2)
3
user=> (a-fn (constantly 1) 2 4)
7
Alternatively, it might make sense for delayed to return a function rather than using partial. Note sure which is better.
A better name than "delayed" is welcome :)
At the home site of Clojure, there is the following statement:
Strings, numbers, characters, true,
false, nil and keywords evaluate to
themselves.
Is there a single combined predicate that tests for any of these, combining string?, number?, char?, true?, false?, nil?, and keyword?. Should I just use (complement symbol?)?
Maybe I'm missing something, but you could use the following to test for any of those conditions and return true if one is true:
(defn self-eval?
[x]
(or (string? x)
(number? x)
(char? x)
(keyword? x)
(true? x)
(false? x)
(nil? x)))
It's easy enough to write a macro that asks "does the given expression evaluate to itself". In fact this is a good example of tasks that can only be done with a macro because they need to see the argument both evaluated and unevaluated.
(defmacro selfp [a] `(= ~a (quote ~a)))
#'user/selfp
user> (selfp 1)
true
user> (selfp +)
false
user> (selfp [1 2])
true
user> (selfp '(+ 1 2 3))
false
While strings, numbers, characters, keywords, and the booleans are all self-evaluating, other things such as [1 2] are as well,so this may not be a useful test in general.
Another option is to create a function that uses a map:
(defn myclassifier? [x]
(let [types-I-care-about #{java.lang.Sring ...}]
(if (types-I-care-about (type x))
true
false)))
Another option which may have better performance is to use java's dynamism:
(extend-type Object
IMyClassifier
(myclassifier? [x]
(let [c (.getClass x)]
(if (types-I-care-about (type c))
(do
(extend-type (.getClass x)
IMyClassifier
(myclassifier? [x] true))
true)
false))))
where types-I-care-about is a set of types you care about.
I want to apply a series of tests on my list and make sure that all the tests are passed.
Is there a function similar to "andmap" in Clojure?
You could use every?:
user=> (every? string? '("hi" 1))
false
Here's the documentation on every?.
Clojure 1.3 will add every-pred (and the related some-fn for the "or" version).
clojure.core/every-pred
([p] [p1 p2] [p1 p2 p3] [p1 p2 p3 & ps])
Takes a set of predicates and returns a function f that returns true if all of its
composing predicates return a logical true value against all of its arguments, else it returns
false. Note that f is short-circuiting in that it will stop execution on the first
argument that triggers a logical false result against the original predicates.
A naive implementation might be:
(defn every-pred [& preds] (fn [& args] (every? #(every? % args) preds)))
but the actual implementation will have better performance.
I wrote andmap as a macro which takes predicates as its arguments and builds a function that "wraps an and around the predicates", i.e.,
(andmap integer? odd?)
==>
(fn [x] (and (integer? x)
(odd? x)))
(it doesn't expand to exactly this, but it expands to something equivalent to this)
This has the advantage that it shortcuircuts on the predicates so you can write
(every? (andmap integer? odd?) [1 3 "a string"])
without getting a runtime exception as you would get with Arthurs answer.
Here is the definition of andmap:
(defmacro andmap
([] `(fn [& x#] true))
([p & ps] `(fn [& x#] (and (apply ~p x#)
(apply (andmap ~#ps) x#)))))
It is also possible to define andmap as an function which also short-circuits on it's predicates due to lazyness:
(defn andmap [& ps]
(fn [& x]
(every? true? (map (fn [p] (apply p x)) ps))))
The predicates to andmap can take an arbitrary number of arguments, so it is possible to write
(map (andmap #(and (integer? %1)
(integer? %2))
#(and (odd? %1)
(even? %2))
<)
[1 3 9]
[2 6 "string"])
which evaluates to (true true false).
every? will ask "Does this one function return true for each member of the seq", which is close to what I think you are asking for. An improvement on every? would take a list of functions and ask "Are all these predicates true for every member of this seq".
Here is a first attempt:
(defn andmap? [data tests]
(every? true? (for [d data, f tests]
(f d))))
user> (andmap? '(2 4 8) [even? pos?])
true
user> (andmap? '(2 4 8) [even? odd?])
false