define: expected a variable, but found a part - racket

I am getting define error from the function neighbours-in-grid below, do I need to use "let" to declare from inside the define???
;; Purpose: Produce a maze, being a list of whole numbers (cells)
;; between 0 and the square of grid-size (grid-size^2)
;; The list prepresenting the maze will start with 0 and be expanded
;; by randomly picking a cell already in the maze
;; then randomly selecting a neighbouring cell (horz and vert axis
;; only) following tests to ensure that the new neighbouring cell is:
; a) within the grid/maze area
; b) not already in the maze
; c) not adjacent to a cell already in the maze (otherwide could block the maze path)
; d) will likely add tests here to make sure maze does not consume itself
; Grid will lay out as follows (assuming grid-size is 3;
; 0 1 2
; 3 4 5
; 6 7 8
(define grid-size 15)
(define a-maze (list 0))
;Is returned value part of the existing a-maze
(check-expect (member? (random-cell a-maze) a-maze)
#true)
;Is returned vale a number?
(check-expect (number? (random-cell a-maze))
#true)
;Is returned value within the limits of the grid
(check-expect (<= 0 (random-cell a-maze) (sqr grid-size))
#true)
;random-cell: non-empty list -> random number from list
(define (random-cell a-list)
(list-ref a-list (random (length a-list))))
(check-expect (member? 15 (neighbours (random-cell a-maze))) #true)
(check-expect (member? 1 (neighbours (random-cell a-maze))) #true)
(check-expect (member? -1 (neighbours (random-cell a-maze))) #true)
(check-expect (member? -15 (neighbours (random-cell a-maze))) #true)
(check-expect (= (length (neighbours-in-grid (random-cell a-maze))) 2) #true)
;neighbours: non-empty whole number representing a cell -> list of
;neighbouring cells - use horz and vert axis only (no diagonals)
(define (neighbours member-cell)
(list (+ member-cell grid-size) ;cell below
(+ member-cell 1) ;cell immediate right
(- member-cell 1) ;cell immediate left
(- member-cell grid-size) ;cell above
)
)
;neighbours-in-grid: non-empty list of potential neighbours -> narrowed list of validated neighbours that are in the grid
(define (neighbours-in-grid (neighbours (random-cell a-maze)))
(cond [(< 0 (first neighbours) > grid-size) (remove (first neighbours))]
[(< 0 (second neighbours) > grid-size) (remove (second neighbours))]
[(< 0 (third neighbours) > grid-size) (remove (third neighbours))]
[(< 0 (fourth neighbours) > grid-size) (remove (fourth neighbours))]
))

The problem here has to do with the first line of your neighbours-in-grid function.
Specifically, a function is defined using this shape:
(define (<name-of-function> <name-of-argument> ...)
<body-of-function>)
Your function's first line looks like this:
(define (neighbours-in-grid (neighbours (random-cell a-maze)))
(cond ...))
This doesn't fit the pattern. What are the names of the arguments? Based on the code that follows this line, it looks to me like you want a single argument, named neighbours. If so, this code should instead read:
(define (neighbours-in-grid neighbours)
(cond ...))
I think part of the confusion here may stem from the fact that separately, you have a function named neighbours.

Related

Simplifying a Racket function

I have the following function "change" which takes a certain amount of money to be paid, the size of the bill/coin used to pay, and returns a list with the number of "coins" ($50, $20 $10 $5 $2 and $1) one would receive after completing the transaction:
(define (change total payment)
(let [(x (- payment total))]
(change-aux x '(50 20 10 5 2 1))))
(define (change-aux change coins)
(cond
[(empty? coins) empty]
[else (let [(num-coins (floor (/ change (car coins))))]
(append (list num-coins)
(change-aux (- change (* (car coins) num-coins)) (cdr coins))))]))
So, if I input these parameters:
> (change 44 200)
It returns the output:
'(3 0 0 1 0 1)
That's 200-44 = 156, which corresponds to 3 coins worth $50, 1 worth $5 and 1 worth $1.
My question would be if there's a more elegant, simplified way to write a similar procedure without relying on auxiliary functions, and rather use lambda, filter, map, foldr, foldl etc?
Thanks in advance.
Here is a solution in a different Lisp dialect which shows how to do it with a left fold (reduce) without any mutation of an accumulator variable, as a kind of functional counterpoint to the existing solution.
(defun change (amount coins)
(reduce-left (tb ((counts . rem) next-coin)
(let* ((coin-count (floor rem next-coin))
(coin-value (* coin-count next-coin)))
(cons (cons coin-count counts)
(- rem coin-value))))
coins
(cons '() amount)))
3> (change 156 '(50 20 10 5 2 1))
((1 0 1 0 0 3) . 0)
4> (change 88 '(50 20 10 5 2 1))
((1 1 1 1 1 1) . 0)
Note that the values end up reported in reverse order and wrapped in an extra cons cell; a "porcelain" function could be used around this "plumbing" to report the result in the expected form.
The idea is that we have an accumulator which looks like this: (counts . remainder). The counts part of the accumulator stored in the car is the list of coins accumulated so far. When the reduce is done, this holds the final list. The cdr field holds the remaining amount to be processed; since the last coin is 1, this will always emerge as zero.
Using this accumulator structure, we process the list of coins.
On each call to our reduce kernel function, the left argument is the accumulator, and the right argument, next-coin, is the next coin denomination value.
I used a macro called tb ("tree bind") macro, which is a kind of lambda that provides built-in destructuring, to make it look like we have three parameters.
The initial value for the reduce job is the starting accumulator, which has an empty list of coins, and the full original amount: (cons nil amount) (which I rewrote to (cons '() amount) for better Scheme compatibility).
The reduce function is very simple: greedily calculate how many of the next coin value are needed to represent the remainder, and then calculate the new remainder, packaging these up into a new accumulator cons cell that is returned, and is passed to the next invocation of the function, or returned when the list of coin values has been processed.
Hopefully this points the way to "a more elegant, simplified way to write a similar procedure without relying on auxiliary functions, and rather use lambda, filter, map, foldr, foldl etc" that you can work out in Racket. Good luck!
Sure, you can.
Final solution
(define (change total payment (coins '(50 20 10 5 2 1)))
(let ((rest-diff (- payment total)))
(map (lambda (coin)
(let ((num-coins (floor (/ rest-diff coin))))
(set! rest-diff (- rest-diff (* num-coins coin)))
num-coins))
coins)))
Step by step
First of all, using inner define, you can get rid of the auxiliary function from the global namespace.
(define (change total payment)
(define (change-aux change coins)
(cond
[(empty? coins) empty]
[else (let [(num-coins (floor (/ change (car coins))))]
(append (list num-coins)
(change-aux (- change (* (car coins) num-coins)) (cdr coins))))]))
(let [(x (- payment total))]
(change-aux x '(50 20 10 5 2 1))))
Then, you can pull some variables of the helper function's to the global function's lambda list.
(define (change total payment (coins '(50 20 10 5 2 1)))
(define (change-aux change) ;; eliminate coins in the inner lambda list
(cond
[(empty? coins) empty] ;; coins in function body looked up from outer arguments
[else (let [(num-coins (floor (/ change (car coins))))]
(append (list num-coins)
(change-aux (- change (* (car coins) num-coins)) (cdr coins))))]))
(let [(x (- payment total))]
(change-aux x))) ;; eliminate coins in the call
Then, looking at the code of change-aux, one understands this is actually
a looping through and trying to fit maximal multiples of current value
into the rest of the difference remaining - and collecting those reults. One could loop using map and use set! to mutate the rest.
(define (change total payment (coins '(50 20 10 5 2 1)))
(let ((rest-diff (- payment total)))
(map (lambda (coin)
(let ((num-coins (floor (/ rest-diff coin))))
(set! rest-diff (- rest-diff (* num-coins coin)))
num-coins))
coins)))
Then, you call like above:
(change 44 200)
;; '(3 0 0 1 0 1)

Object appears at least twice in list?

Let's say we have a: (list 1 2 3 5 2 6 7 4)
I want to know if 2 appears at least twice.
member? checks if it appears at all. I suppose I could check member? then run remove then check member? again. Is there a more efficient method?
You should use count, which receives a lambda for checking any condition you want and a list to be checked. It returns the number of elements in the list that meet the condition:
(count (lambda (n) (= n 2)) (list 1 2 3 5 2 6 7 4))
=> 2
Then it's a simple matter to verify if the count fulfills the requirement that the number appears at least twice.
One can use member function with recursion to check if first element is the desired number and is also part of rest of the list- then loop again with rest of the list:
(define (is_duplicated n L)
(let loop ((L L))
(cond
[(empty? L) #f]
[(and (= n (first L))
(member n (rest L))) #t]
[else (loop (rest L))]
)))
(is_duplicated 2 (list 1 2 3 5 2 6 7 4))
Note that it is "member" and not "member?" which is part of Racket base functions (https://docs.racket-lang.org/search/index.html?q=member%3F).

Racket function that returns all numbers whose sum is <= given number?

I'm taking an intro to computer science course and one question needs me to write a function that takes a list of numbers and a number and returns the numbers in the list whose sum is less than the given number. I've written the function signature, definition, and check-expects, but I'm stuck. The function needs to assume intermediate student with lambda. I don't want any direct answers here; just help so that I can reach the answer myself.
I know it needs to use recursion. Perhaps a helper function would be needed.
;; sum-up-to: lon, number -> lon
;; consumes a list of numbers and a number and
;; returns the numbers in the list whose sum is
;; less than or equal to the given number
(define the-numbers (list 1 2 3 4 5 6 7 8 9))
(check-expect (sum-up-to the-numbers 7) (list 1 2 3))
(check-expect (sum-up-to the-numbers 18) (list 1 2 3 4 5))
(check-expect (sum-up-to the-numbers 45) the-numbers)
This problem can be simplified if we sort the list first and if we define a helper function that keeps track of the accumulated sum. Here's a skeleton, fill-in the blanks with the missing expressions and you'll have the solution:
(define (sum-up-to lst n)
(helper <???> n 0)) ; sort the input list, pass it to the helper
(define (helper lst n sum)
(cond (<???> '()) ; if the list is empty, end the recursion
((> <???> n) '()) ; also end recursion if sum + current element > n
(else
(cons <???> ; otherwise cons current element
(helper <???> ; advance recursion over list
n
(+ <???> <???>)))))) ; update sum
Following recursive method keeps adding numbers from the list sequentially to an initially empty outlist, till the sum is reached:
(define the-numbers (list 1 2 3 4 5 6 7 8 9))
(define (f lst sum)
(let loop ((lst lst)
(ol '()))
(if (or (..ENTER CONDITION FOR EMPTY LIST..)
(..ENTER CONDITION WHEN SUM IS REACHED..)
(..ENTER HOW TO PUT THE NEW LIST OUT..)
(loop (..ENTER ARGUMENTS TO BE SENT TO NEXT LOOP..)
))))
(f the-numbers 7)
(f the-numbers 18)
(f the-numbers 45)
Output:
'(1 2 3)
'(1 2 3 4 5)
'(1 2 3 4 5 6 7 8 9)

Test if all elements of a list are different from each other

I have a list of lists and want to test if all elements are different from each other, i.e. equal should return nil for all combinations of list elements.
E.g.
(defparameter feld '((1 0 0 5 5 0)
(0 0 0 0 0 0)
(1 1 5 5 0 0)
(0 1 0 1 5 5)
(5 5 1 0 1 0)
(1 0 1 0 5 5)))
I thought of using reduce but as far as I understand it only tests the equality of neighbors, as would do a loop construct like:
(loop for i below (length feld)
for j from 1
if (equal (nth i feld) (nth j feld)) return t)
Is there a simple way using a standard construct which I do not see at the moment or do I have to create a recursive function?
The whole data structure represents a "board game" where every list is a line on the board and each element in the inside-lists is a value of this very field. The three numerical values (0, 1 and 5) are something like empty, Symbol A and Symbol B. A valid board cannot have two identical lines. This is why I want to identify those.
Basically, it is like remove-duplicates without removing. In the meantime I was thinking about something like this:
(defun duplicates-p (lst)
(cond ((null lst) '())
((member (car lst) (cdr lst)) t)
(t (duplicates-p (rest lst)))))
Something like this:
(defun unique (lsts &aux (h (make-hash-table :test 'equal)))
(loop :for lst :in lsts
:never (gethash lst h)
:do (setf (gethash lst h) t)))

Lisp, sub total of a numbers in a nested list

i have a problem that i just cant work out,
the user enters a list ie
(total-cost
'((anItem 2 0.01)
(item 3 0.10)
(anotherItem 4 4.10)
(item 5 2.51)))
i need to add the number on the end together and then return the result
my current code returns the code after each addition. and also throws a error about unexpected type
(defun total-cost (list)
(loop with sum = 0
for x in list
collect (setf sum (+ sum (last x)))
)
)
Error: (0.01)' is not of the expected typeNUMBER'
Any help is appreciated
Thanks Dale
Using LOOP:
CL-USER 19 > (loop for (nil nil number) in '((anItem 2 0.01)
(item 3 0.10)
(anotherItem 4 4.10)
(item 5 2.51))
sum number)
6.72
REDUCE is another option:
CL-USER 20 > (reduce '+
'((anItem 2 0.01)
(item 3 0.10)
(anotherItem 4 4.10)
(item 5 2.51))
:key 'third)
6.72
Loop has a keyword sum for summing so you don't have to have an explicit variable nor use setf:
(defun total-cost (list)
(loop for x in list sum (third x)))
As Chris said, use (car (last x)) if the number you're looking for is always the last one. Or you can use (third x) as in my example if it's always the third one.
Also, note that the use of collectis wrong if your aim is to return the sum only; your example (corrected) returns
(0.01 0.11 4.21 6.7200003)
whereas mine returns
6.7200003
Note that if you want so escape the rounding errors as much as possible you need to use an exponent marker to make them double-floats for example:
(total-cost '((anItem 2 0.01D0)
(item 3 0.10D0)
(anotherItem 4 4.10D0)
(item 5 2.51D0)))
=> 6.72D0
last returns the last cons cell in the list, not its value. You need to use (car (last x)) instead.
Just in case you want the code to give you a precise result rather then being short:
(defun kahan-sum (floats)
(loop
:with sum := 0.0 :and error := 0.0
:for float :in floats
:for epsilon := (- float error)
:for corrected-sum := (+ sum epsilon) :do
(setf error (- corrected-sum sum epsilon) sum corrected-sum)
:finally (return sum)))
(defun naive-sum (floats) (loop :for float :in floats :sum float))
(let ((floats (loop :repeat 1000 :collect (- (random 1000000.0) 1000000.0))))
(format t "~&naive sum: ~f, kahan sum: ~f" (naive-sum floats) (kahan-sum floats)))
;; naive sum: -498127420.0, kahan sum: -498127600.0
Read more about why it works like this here: http://en.wikipedia.org/wiki/Kahan_summation_algorithm
Coming late to the party... How about a little lisping instead of looping? ;-)
(defun sum-3rd (xs)
(let ((sum 0))
(dolist (x xs sum)
(incf sum (nth 2 x)))))