I have to write a recursive macro for list addition in Common Lisp (homework). What I have so far is :
(defmacro matrix-add-row (r1 r2 sum_row)
(if (not (and r1 r2)) `sum_row
(progn
`(matrix-add-row (cdr r1) (cdr r2) (cons sum_row (+ (car r1) (car r2))))
(reverse sum_row)
)
)
)
I call this function with
(matrix-add-row `(1 2) `(3 4) ())
and as an output I get unvaluated code instead of numbers (which leads going to infinite loop).
How to put , ` properly (or call the macro properly)?
Firstly, to me this seems a rather bizarre thing to do with a macro. I assume the point is that you use the macro to transform (matrix-add-row '(1 2) '(3 4)) to an explicit list of sums like (list (+ 1 3) (+ 2 4)).
Also, what you have written has several problems which look like you don't quite understand how the backtick works. So I think the easiest way to help is to solve an example for you.
Since this is homework, I'm going to solve a different (but similar) question. You should be able to take the answer and use it for your example. Suppose I want to solve the following:
Write a macro, diffs, which computes all differences of pairs of successive elements in a list. For example,
(diffs '(1 2 3)) should expand to (list (- 2 1) (- 3 2)), which will then evaluate to (1 1).
Note that my macro won't do the actual subtraction, so I can use it even if I don't know some of the numbers until runtime. (The reason I think this sort of question is a bit weird is that it does need to know the length of the list at compile time).
My solution is going to be used as a macro with one argument but if I want to use recursion I'll need to pass in an accumulator too, which I can start with nil. So I write something like this:
(defmacro diffs (lst &optional accumulator)
...)
Now what do I do with lst? If lst is nil, I want to bottom out and just return the accumulator, with a call to list at the front, which will be code to make my list. Something like this:
(defmacro diffs (lst &optional accumulator)
(cond
((null lst)
;; You could write `(list ,#accumulator) instead, but that seems
;; unnecessarily obfuscated.
(cons 'list accumulator))
(t
(error "Aargh. Unhandled"))))
Let's try it!
CL-USER> (diffs nil)
NIL
Not hugely exciting, but it looks plausible. Now use macroexpand, which just does the expansion without the evaluation:
CL-USER> (macroexpand '(diffs nil))
(LIST)
T
And what if we'd already got some stuff from a recursion?
CL-USER> (macroexpand '(diffs nil ((- a b) (- b c))))
(LIST (- A B) (- B C))
T
Looks good! Now we need to deal with the case when there's an actual list there. The test you want is consp and (for my example) it only makes sense when there's at least two elements.
(defmacro diffs (lst &optional accumulator)
(cond
;; A list of at least two elements
((and (consp lst) (consp (cdr lst)))
(list 'diffs (cdr lst)
(cons (list '- (cadr lst) (car lst)) accumulator)))
;; A list with at most one element
((listp lst)
(cons 'list accumulator))
(t
(error "Aargh. Unhandled"))))
This seems almost to work:
CL-USER> (macroexpand '(diffs (3 4 5)))
(LIST (- 5 4) (- 4 3))
T
but for two problems:
The list comes out backwards
The code is a bit horrible when we actually construct the recursive expansion
Let's fix the second part first by using the backtick operator:
(defmacro diffs (lst &optional accumulator)
(cond
;; A list of at least two elements
((and (consp lst) (consp (cdr lst)))
`(diffs ,(cdr lst)
,(cons `(- ,(cadr lst) ,(car lst)) accumulator)))
;; A list with at most one element
((listp lst)
(cons 'list accumulator))
(t
(error "Aargh. Unhandled"))))
Hmm, it's not actually much shorter, but I think it's clearer.
For the second part, we could proceed by adding each item to the end of the accumulator rather than the front, but that's not particularly quick in Lisp because lists are singly linked. Better is to construct the accumulator backwards and then reverse it at the end:
(defmacro diffs (lst &optional accumulator)
(cond
;; A list of at least two elements
((and (consp lst) (consp (cdr lst)))
`(diffs ,(cdr lst)
,(cons `(- ,(cadr lst) ,(car lst)) accumulator)))
;; A list with at most one element
((listp lst)
(cons 'list (reverse accumulator)))
(t
(error "Aargh. Unhandled"))))
Now we get:
CL-USER> (macroexpand '(diffs (3 4 5)))
(LIST (- 4 3) (- 5 4))
T
Much better!
Two last things. Firstly, I still have an error clause in my macro. Can you see how to trigger it? Can you think of a better behaviour than just outputting an error? (Your macro is going to have to deal with the same problem)
Secondly, for debugging recursive macros like this, I recommend using macroexpand-1 which just unfolds one level at once. For example:
CL-USER> (macroexpand-1 '(diffs (3 4 5)))
(DIFFS (4 5) ((- 4 3)))
T
CL-USER> (macroexpand-1 *)
(DIFFS (5) ((- 5 4) (- 4 3)))
T
CL-USER> (macroexpand-1 *)
(LIST (- 4 3) (- 5 4))
T
There are two problems with your logic. First you are calling reverse on each iteration instead of at the end of the iteration. Then you are accumulating the new values, through cons, in the cdr of the cons cell as opposed to the car.
Also I don't see why this have to be a macro so using a function.
(defun matrix-add-row (r1 r2 sum-row)
(if (or (endp r1) (endp r2))
(reverse sum-row)
(matrix-add-row (cdr r1)
(cdr r2)
(cons (+ (car r1) (car r2))
sum-row))))
(matrix-add-row '(1 2) '(3 4) ())
;; => (4 6)
Related
I'm working through a textbook on programming languages, and one of the exercises was to make a function in Scheme that flips tuples in a list. Here's my code:
; invert : Listof(List(Int,Int)) -> Listof(List(Int,int))
; usage: (invert '((a 1) (a 2) (1 b) (2 b))) -> ((1 a) (2 a) (b 1) (b 2))
(define invert
(lambda (lst)
(if (null? lst)
'()
(cons
(flip (car lst))
(invert (cdr lst))))))
; flip : List(Int,Int) -> List(Int,int)
; usage: (flip '(a 1)) -> (1 a)
(define flip
(lambda (tuple)
(if (not (eqv? (length (tuple)) 2))
(eopl:error 'flip
"Tuple is not length 2~%")
(cons (cdr tuple) (car tuple)))))
I tried testing my program in chez-scheme. When I use the test case in the usage comment, I get this error: Exception: attempt to apply non-procedure (a 1). I've never worked with Scheme before, so I'd greatly appreciate any help and advice. Thanks!
You have a coupe of errors in flip, this should fix them:
(define flip
(lambda (tuple)
(if (not (= (length tuple) 2))
(eopl:error 'flip "Tuple is not length 2~%")
(list (cadr tuple) (car tuple)))))
In particular:
The specific error reported was because of this expression: (tuple). We must not surround variables with (), unless they're procedures that we intend to call.
We should use = for comparing numbers, not eqv?.
In this expression: (cons (cdr tuple) (car tuple)) there are two issues, for building a list of two elements we use list, not cons. And for accessing the second element we use cadr, not cdr - you should read a bit about how cons, car and cdr are used for building lists.
Notice that there's a simpler way to solve this problem if we use map; I'll skip error checking for simplicity:
(define (invert lst)
(map (lambda (tuple) (list (cadr tuple) (car tuple)))
lst))
I'm trying to write a simple procedure for finding the n:th prime but I don't think I understand how to reference variables correctly in racket.
What I want is for the inner procedure sieve-iter to add primes to the list primlst which is in the namespace of sieve but I get an infinite loop. My guess is that primlst within sieve-iter is causing issues.
(define (sieve n) ;; returns the n:th prime (n>0)
(let [(primlst '(2))
(cand 3)]
(define (sieve-iter i lst)
(cond ((null? lst) (and (cons i primlst) (sieve-iter (+ i 2) primlst))) ;;prime
((= (length primlst) n) (car primlst)) ;;end
((= (modulo i (car lst)) 0) (sieve-iter (+ i 2) primlst)) ;;non-prime
(#t (sieve-iter n (cdr lst))))) ;;unclear if prime
(sieve-iter cand primlst)))
Any help is appreciated!
First of all, you shouldn't refer to primlist at all within the sieve-iter function. Instead, you should refer to lst.
Second of all, you appear to be mistaken on the effect of this expression:
(and (cons i primlst) (sieve-iter (+ i 2) primlst))
You seem to be interpreting that as meaning "Add i to the primlist and then start the next iteration."
(cons i primlist) changes nothing. Instead, it creates a new list consisting of primlist with i in front of it and then evaluates to that value. The original primlist (which should have been lst anyway) is left untouched.
Also, and is for Boolean logic, not for stringing commands together. It evaluates each of its subexpressions separately until it finds one that evaluates to #f and then it stops.
You should replace that whole expression with this:
(sieve-iter (+ i 2) (cons i lst))
...which passes the new list created by cons to the next run of sieve-iter.
Your trying to do too much in one function, let prime-iter just worry about the iteration to build up the list of primes. Make another internal function to recurse down the existing primes to test the new candidate.
(define (sieve n) ;; returns the n:th prime (n>0)
(define (sieve-iter i lst remaining)
(cond ;((null? lst) (and (cons i primlst) (sieve-iter (+ i 2) primlst))) ;;should never be null, we are building the list up
((<= remaining 0) (car lst)) ;;checking length every time gets expensive added a variable to the function
((sieve-prime? i lst) ;;if prime add to lst and recurse on next i
(sieve-iter (+ i 2) (cons i lst) (- remaining 1)))
(else
(sieve-iter (+ i 2) lst remaining)))) ; else try next
(define (sieve-prime? i lst)
(cond ((null? lst) #t)
((= 0 (modulo i (car lst))) #f)
(else (sieve-prime? i (cdr lst)))))
(let ((primlst '(2)) ;;you generally don't modify these,
(cand 3)) ;mostly they just bind values to name for convenience or keep from having to re-calculate the same thing more than once
(sieve-iter cand primlst (- n 1))))
You could have used set! to modify primlist where it was before, but the procedure is no longer obviously a pure function.
There is another low-handing optimization possible here, when calling sieve-prime? filter the lst argument to remove values larger than the square root of i.
I am totally new to Lisp.
How to find the difference between elements in an arithmetic progression series?
e.g.
(counted-by-N '(20 10 0))
Return -10
(counted-by-N '(20 10 5))
(counted-by-N '(2))
(counted-by-N '())
Returns Nil
In Python/C and other languages, it is very straightforward... Kinda stuck here in Lisp.
My pseudo algorithm would be something like this:
function counted-by-N(L):
if len(L) <= 1:
return Nil
else:
diff = L[second] - L[first]
for (i = second; i < len(L) - 1; i++):
if L[i+1] - L[i] != diff
return Nil
return diff
Current work:
(defun count-by-N (L)
(if (<= (length L) 1) Nil
(
(defvar diff (- (second L) (first L)))
; How to do the loop part?
))
)
(flet ((by-n (list &aux
(e1 (first list))
(e2 (second list))
(difference (and e1 e2 (- e2 e1))))
(and difference
(loop for (one two) on list
while (and one two)
when (/= (- two one) difference)
do (return-from by-n nil)))
difference))
(by-n '(20 10 0)))
or
(flet ((by-n (list &aux
(e1 (first list))
(e2 (second list))
(difference (and e1 e2 (- e2 e1))))
(when difference
(loop for (one two) on list
while (and one two)
when (/= (- two one) difference)
do (return-from by-n nil))
difference)))
(by-n '(20 10 0)))
As far as you said on the second answer the best choice you have to do this example is implement it recursively.
Example Using List Processing (good manners)
That way, you have some ways to do this example on the recursively and simple way:
(defun count-by-N-1 (lst)
(if (equal NIL lst)
NIL
(- (car (cdr lst)) (car lst))
)
(count-by-N-1 (cdr lst))
)
On this first approach of the function count-by-N-1 I am using the simple car and cdr instructions to simplify the basics of Common Lisp List transformations.
Example Using List Processing Shortcuts (best implementation)
However you can resume by using some shortcuts of the car and cdr instructions like when you want to do a a car of a cdr, like I did on this example:
(defun count-by-N-2 (lst)
(if (equal NIL lst)
NIL
(- (cadr lst) (car lst))
)
(count-by-N-2 (cdr lst))
)
If you have some problems to understand this kind of questions using basic instructions of Common Lisp List transformation as well as car and cdr, you still can choose the first, second and rest approach. However I recommend you to see some of this basic instructions first:
http://www.gigamonkeys.com/book/they-called-it-lisp-for-a-reason-list-processing.html
Example Using Accessors (best for understand)
(defun count-by-N-3 (lst)
(if (equal NIL lst)
NIL
(- (first (rest lst)) (first lst))
)
(count-by-N-3 (rest lst))
)
This last one, the one that I will explain more clearly since it is the most understandable, you will do a recursion list manipulation (as in the others examples), and like the others, until the list is not NIL it will get the first element of the rest of the list and subtract the first element of the same list. The program will do this for every element till the list is "clean". And at last returns the list with the subtracted values.
That way if you read and study the similarities between using first, second and rest approach against using car and cdr, you easily will understand the both two first examples that I did put here.
Here is my final answer of this question which uses recursion:
(defun diff (N)
(- (second N) (first N))
)
(defun count-by-N (L)
(cond
((null L) nil)
((= (length L) 1) nil)
((= (length L) 2) (diff L))
((= (diff L) (diff (rest L))) (count-by-N (rest L)))
(T nil)
)
)
I am watching SICP video lectures and i came to a section where tutors are showing procedures to work with lists, so, here is one of them:
(define (map p l)
(if (null? l)
(list)
(cons (p (car l))
(map p (cdr l)))))
What i want to ask is: is there a way to define map in iterative way, or that cons requires lazy evaluation to be executed right?
You original code is almost tail recursive.. the only thing that makes it not is the cons part. If Scheme had equal requirement for having TRMC optimization as it has TCO requirement you could leave your code as is and the implementation would have made it tail recursive for you.
Since it isn't a requirement we need to do our own TRMC optimization. Usually when iterating a list in a loop and having it tail recursive by using an accumulator you get the result in the opposite order, thus you can do linear update reverse:
(define (map proc lst)
(let loop ((lst lst) (acc '()))
(cond ((null? lst) (reverse! acc) acc)
(else (loop (cdr lst)
(cons (proc (car lst)) acc))))))
Or you can do it all in one pass:
(define (map proc lst)
(define head (list 1))
(let loop ((tail head) (lst lst))
(cond ((null? lst) (cdr head))
(else (set-cdr! tail (list (proc (car lst))))
(loop (cdr tail) (cdr lst))))))
Now in both cases you mutate only the structure the procedure has itself created, thus for the user it might as well be implemented in the same manner as your example.
When you use higher order procedures like map from your implementation it could happen it has been implemented like this. It's easy to find out by comparing performance on the supplied map with the different implementations with a very long list. The difference between the executions would tell you if it's TRMCO or how the supplied map probably has been implemented.
You need to embrace recursion in order to appreciate SICP and Scheme in general, so try to get used to it, you will appreciate it later, promised.
But yes, you can:
(define (iterative-map f lst)
(define res null)
(do ((i (- (length lst) 1) (- i 1))) ((= i -1))
(set! res (cons (f (list-ref lst i)) res)))
res)
(iterative-map (lambda (x) (+ x 1)) '(1 3 5))
=> '(2 4 6)
but using set! is considered bad style if avoidable.
In Racket you have a different set of loops that are more elegant:
(define (for-map f lst)
(for/list ((i lst))
(f i)))
(for-map add1 '(1 3 5))
=> '(2 4 6)
I'm new to Scheme (via Racket) and (to a lesser extent) functional programming, and could use some advise on the pros and cons of accumulation via variables vs recursion. For the purposes of this example, I'm trying to calculate a moving average. So, for a list '(1 2 3 4 5), the 3 period moving average would be '(1 2 2 3 4). The idea is that any numbers before the period are not yet part of the calculation, and once we reach the period length in the set, we start averaging the subset of the list according the chosen period.
So, my first attempt looked something like this:
(define (avg lst)
(cond
[(null? lst) '()]
[(/ (apply + lst) (length lst))]))
(define (make-averager period)
(let ([prev '()])
(lambda (i)
(set! prev (cons i prev))
(cond
[(< (length prev) period) i]
[else (avg (take prev period))]))))
(map (make-averager 3) '(1 2 3 4 5))
> '(1 2 2 3 4)
This works. And I like the use of map. It seems composible and open to refactoring. I could see in the future having cousins like:
(map (make-bollinger 5) '(1 2 3 4 5))
(map (make-std-deviation 2) '(1 2 3 4 5))
etc.
But, it's not in the spirit of Scheme (right?) because I'm accumulating with side effects. So I rewrote it to look like this:
(define (moving-average l period)
(let loop ([l l] [acc '()])
(if (null? l)
l
(let* ([acc (cons (car l) acc)]
[next
(cond
[(< (length acc) period) (car acc)]
[else (avg (take acc period))])])
(cons next (loop (cdr l) acc))))))
(moving-average '(1 2 3 4 5) 3)
> '(1 2 2 3 4)
Now, this version is more difficult to grok at first glance. So I have a couple questions:
Is there a more elegant way to express the recursive version using some of the built in iteration constructs of racket (like for/fold)? Is it even tail recursive as written?
Is there any way to write the first version without the use of an accumulator variable?
Is this type of problem part of a larger pattern for which there are accepted best practices, especially in Scheme?
It's a little strange to me that you're starting before the first of the list but stopping sharply at the end of it. That is, you're taking the first element by itself and the first two elements by themselves, but you don't do the same for the last element or the last two elements.
That's somewhat orthogonal to the solution for the problem. I don't think the accumulator is making your life any easier here, and I would write the solution without it:
#lang racket
(require rackunit)
;; given a list of numbers and a period,
;; return a list of the averages of all
;; consecutive sequences of 'period'
;; numbers taken from the list.
(define ((moving-average period) l)
(cond [(< (length l) period) empty]
[else (cons (mean (take l period))
((moving-average period) (rest l)))]))
;; compute the mean of a list of numbers
(define (mean l)
(/ (apply + l) (length l)))
(check-equal? (mean '(4 4 1)) 3)
(check-equal? ((moving-average 3) '(1 3 2 7 6)) '(2 4 5))
Well, as a general rule, you want to separate the manner in which you recurse and/or iterate from the content of the iteration steps. You mention fold in your question, and this points in the right step: you want some form of higher-order function that will handle the list traversal mechanics, and call a function you supply with the values in the window.
I cooked this up in three minutes; it's probably wrong in many ways, but it should give you an idea:
;;;
;;; Traverse a list from left to right and call fn with the "windows"
;;; of the list. fn will be called like this:
;;;
;;; (fn prev cur next accum)
;;;
;;; where cur is the "current" element, prev and next are the
;;; predecessor and successor of cur, and accum either init or the
;;; accumulated result from the preceeding call to fn (like
;;; fold-left).
;;;
;;; The left-edge and right-edge arguments specify the values to use
;;; as the predecessor of the first element of the list and the
;;; successor of the last.
;;;
;;; If the list is empty, returns init.
;;;
(define (windowed-traversal fn left-end right-end init list)
(if (null? list)
init
(windowed-traversal fn
(car list)
right-end
(fn left-end
(car list)
(if (null? (cdr list))
right-end
(second list))
init)
(cdr list))))
(define (moving-average list)
(reverse!
(windowed-traversal (lambda (prev cur next list-accum)
(cons (avg (filter true? (list prev cur next)))
list-accum))
#f
#f
'()
list)))
Alternately, you could define a function that converts a list into n-element windows and then map average over the windows.
(define (partition lst default size)
(define (iter lst len result)
(if (< len 3)
(reverse result)
(iter (rest lst)
(- len 1)
(cons (take lst 3) result))))
(iter (cons default (cons default lst))
(+ (length lst) 2)
empty))
(define (avg lst)
(cond
[(null? lst) 0]
[(/ (apply + lst) (length lst))]))
(map avg (partition (list 1 2 3 4 5) 0 3))
Also notice that the partition function is tail-recursive, so it doesn't eat up stack space -- this is the point of result and the reverse call. I explicitly keep track of the length of the list to avoid either repeatedly calling length (which would lead to O(N^2) runtime) or hacking together a at-least-size-3 function. If you don't care about tail recursion, the following variant of partition should work:
(define (partition lst default size)
(define (iter lst len)
(if (< len 3)
empty
(cons (take lst 3)
(iter (rest lst)
(- len 1)))))
(iter (cons default (cons default lst))
(+ (length lst) 2)))
Final comment - using '() as the default value for an empty list could be dangerous if you don't explicitly check for it. If your numbers are greater than 0, 0 (or -1) would probably work better as a default value - they won't kill whatever code is using the value, but are easy to check for and can't appear as a legitimate average