Predicates

lseq? x -> boolean

Returns #t iff x is an lseq, otherwise #f. This procedure may return #t if x is an improper list whose last car is a procedure that requires arguments, since there is no portable way to examine a procedure to determine how many arguments it requires.

lseq= elt= lseq₁ ... -> boolean

Determines lseq equality, given an element-equality procedure. The lseq A equals the lseq B if they are of the same length, and their corresponding elements are equal, as determined by elt=. If the element-comparison procedure's first argument is from lseq_i, then its second argument is from lseq_i+1, i.e. it is always called as (elt= a b) for a an element of lseq A, and b an element of lseq B.

In the n-ary case, every lseq_i is compared to lseq_i+1 (as opposed, for example, to comparing lseq₁ to lseq_i, for i>1). If there are no lseq arguments at all, lseq= simply returns true.

The dynamic order in which the elt= procedure is applied to pairs of elements is not specified. For example, if lseq= is applied to three lseqs, A, B, and C, it may first completely compare A to B, then compare B to C, or it may compare the first elements of A and B, then the first elements of B and C, then the second elements of A and B, and so forth.

The equality procedure must be consistent with eq?. That is, it must be the case that

(eq? x y) => (elt= x y).

Note that this implies that two lseqs which are eq? are always lseq=, as well; implementations may exploit this fact to "short-cut" the element-by-element comparisons.

(lseq= eq?) => #t       ; Trivial cases
(lseq= eq? '(a)) => #t

Selectors

lazy-car   lseq -> object

lazy-rest  lseq -> object

These procedures are synonymous. They return the first element of lseq except for the first element. They are included for completeness, as they are the same as car. It is an error to apply them to an empty lseq.

lazy-cdr   lseq -> object

lazy-rest  lseq -> object

These procedures are synonymous. They return an lseq with the contents of lseq except for the first element. The exact behavior is as follows:

• If lseq is a pair whose cdr is a procedure, then the procedure is invoked with no arguments to produce an object obj.
• If obj is an end-of-file object, then the cdr of lseq is set to the empty list, which is returned.
• If obj is any other object, then a new pair is allocated whose car is obj and whose cdr is the cdr of lseq (i.e. the procedure). The cdr of lseq is set to the newly allocated pair, which is returned.
• If lseq is a pair whose cdr is not a procedure, then the cdr is returned.
• If lseq is not a pair, it is an error.

Implementations that inline cdr are advised to inline lazy-cdr if possible.

lazy-second  lseq -> object
lazy-third   lseq -> object
lazy-fourth  lseq -> object
lazy-fifth   lseq -> object
lazy-sixth   lseq -> object
lazy-seventh lseq -> object
lazy-eighth  lseq -> object
lazy-ninth   lseq -> object
lazy-tenth   lseq -> object

Returns the second, third, ..., or tenth element of lseq.

(lazy-third '(a b c d e)) => c

lazy-ref lseq i -> value

Returns the ith element of lseq. (This is the same as (lazy-first (lazy-drop lseq i)).) It is an error if i >= n, where n is the length of lseq.

    
(lazy-ref '(a b c d) 2) => c

lazy-take lseq i -> list
lazy-drop lseq i -> lseq

lazy-take returns the first i elements of lseq as a proper list.
lazy-drop returns all but the first i elements of lseq.


(lazy-take '(a b c d e)  2) => (a b)
(lazy-drop '(a b c d e)  2) => (c d e)

lazy-drop is exactly equivalent to performing i lazy-rest operations on lseq.

lazy-split-at  lseq i -> [lseq object]

Splits the lseq lseq at index i, returning two values, a list of the first i elements, and an lseq of the remaining elements. It is equivalent to

(values (lazy-take lseq i) (lazy-drop lseq i))

The whole lazy sequence

lazy-realize lseq

Repeatedly applies lazy-cdr to lseq until its generator (if there is one) has been exhausted, and returns lseq, which is now a proper list. This procedure can be called on an arbitrary lseq before passing it to a procedure which only accepts lists. However, if the generator never returns an end-of-file object, lazy-realize will never return.

(lazy-zip '(one two three) 
     '(1 2 3)
     '(odd even odd even odd even odd even))
    => ((one 1 odd) (two 2 even) (three 3 odd))

(lazy-zip '(1 2 3)) => ((1) (2) (3))

(lazy-unzip2 (list '(1 one) '(2 two) '(3 three))) =>
    (1 2 3) 
    (one two three)

Folding and mapping

It is an error if all the lseq arguments to these procedures can be indefinitely expanded: for example, if they are built from circular generators or procedures that never return an end-of-file object.

lazy-fold kons knil lseq₁ lseq₂ ... -> value

The fundamental lseq iterator.

First, consider the single lazy-parameter case. If lseq₁ = (e₁ e₂ ... e_n), then this procedure returns

(kons en ... (kons e2 (kons e1 knil)) ... )

That is, it obeys the (tail) recursion

(lazy-fold kons knil lis) = (lazy-fold kons (kons (lazy-first lis) knil) (lazy-last lis))
(lazy-fold kons knil '()) = knil

Examples:

(lazy-fold + 0 lis)			; Add up the elements of LIS.

(lazy-fold lseq* (lseq) lseq)		; Reverse lseq.

(lazy-fold lseq* tail rev-head)	; See append-reverse.

;; How many symbols in LIS?
(lazy-fold (lambda (x count) (if (symbol? x) (+ count 1) count))
      0
      lseq)

;; Length of the longest string in lseq:
(lazy-fold (lambda (s max-len) (max max-len (string-length s)))
      0
      lseq)

If n lseq arguments are provided, then the kons function must take n+1 parameters: one element from each lseq, and the "seed" or fold state, which is initially knil. The fold operation terminates when the shortest lseq runs out of values:

(lazy-fold lseq* '() '(a b c) '(1 2 3 4 5)) => (c 3 b 2 a 1)

lazy-reduce f ridentity lseq -> value

lazy-reduce is a variant of lazy-fold.

ridentity should be a "right identity" of the procedure f — that is, for any value x acceptable to f,

(f x ridentity) = x

lazy-reduce has the following definition:

If lseq = (), return ridentity;
Otherwise, return (lazy-fold f (lazy-first lseq) (lazy-last lseq)).

...in other words, we compute (lazy-fold f ridentity lseq).

Note that ridentity is used only in an empty-lseq case. You typically use lazy-reduce when applying f is expensive and you'd like to avoid the extra application incurred when lazy-fold applies f to the head of lseq and the identity value, redundantly producing the same value passed in to f. For example, if f involves searching a file directory or performing a database query, this can be significant. In general, however, lazy-fold is useful in many contexts where lazy-reduce is not (consider the examples given in the lazy-fold definition — only one of the folds uses a function with a right identity. The other four may not be performed with lazy-reduce).

;; take the max of an lseq of non-negative integers.
(lazy-reduce max 0 nums)

lazy-map proc lseq₁ lseq₂ ... -> lseq

proc is a procedure taking as many arguments as there are lseq arguments and returning a single value. lazy-map applies proc element-wise to the elements of the lseqs and returns an lseq of the results, in order. The dynamic order in which proc is applied to the elements of the lseqs is unspecified.

(lazy-map icadr '((a b) (d e) (g h))) =>  (b e h)

(lazy-map (lambda (n) (expt n n))
     '(1 2 3 4 5))
    =>  (1 4 27 256 3125)

(lazy-map + '(1 2 3) '(4 5 6)) =>  (5 7 9)

(let ((count 0))
  (lazy-map (lambda (lazy-gnored)
         (set! count (+ count 1))
         count)
       '(a b))) =>  (1 2) or (2 1)

lazy-for-each proc lseq₁ lseq₂ ... -> unspecified

lazy-pair-for-each proc lseq₁ lseq₂ ... -> unspecified

The arguments to lazy-for-each are like the arguments to lazy-map, but lazy-for-each calls proc for its side effects rather than for its values. Unlike lazy-map, lazy-for-each is guaranteed to call proc on the elements of the lseqs in order from the first element(s) to the last, and the value returned by lazy-for-each is unspecified.

The procedure lazy-pair-for-each is the same as lazy-for-each, except that it calls proc on each pair rather than each element.

(let ((v (make-vector 5)))
  (lazy-for-each (lambda (lazy-)
              (vector-set! v i (* i i)))
            '(0 1 2 3 4))
  v)  =>  #(0 1 4 9 16)

Searching

The following procedures all search lseqs for the leftmost element satisfying some criteria.

(lazy-find even? '(3 1 4 1 5 9)) => 4

(cond ((lazy-find-tail pred lseq) => (lambda (lseq) ...))
      (else ...)) ; Search failed.

(lazy-find-tail even? '(3 1 37 -8 -5 0 0)) => (-8 -5 0 0)
(lazy-find-tail even? '(3 1 37 -5)) => #f

;; imember x lseq:
(lazy-find-tail (lambda (elt) (equal? x elt)) lseq)

(lazy-take-while even? '(2 18 3 10 22 9)) => (2 18)

(lazy-drop-while even? '(2 18 3 10 22 9)) => (3 10 22 9)

(values (lazy-take-while pred lseq) 
        (lazy-drop-while pred lseq))

(lazy-span even? '(2 18 3 10 22 9)) =>
  (2 18)
  (3 10 22 9)

(lazy-break even? '(3 1 4 1 5 9)) =>
  (3 1)
  (4 1 5 9)

(lazy-any integer? '(a 3 b 2.7))   => #t
(lazy-any integer? '(a 3.1 b 2.7)) => #f
(lazy-any < '(3 1 4 1 5)
       '(2 7 1 8 2)) => #t

(lazy-index even? '(3 1 4 1 5 9)) => 2
(lazy-index < '(3 1 4 1 5 9 2 5 6) '(2 7 1 8 2)) => 1
(lazy-index = '(3 1 4 1 5 9 2 5 6) '(2 7 1 8 2)) => #f

    (lazy-memq 'a '(a b c))           =>  (a b c)
    (lazy-memq 'b '(a b c))           =>  (b c)
    (lazy-memq 'a '(b c d))           =>  #f
    (lazy-memq (lseq 'a) '(b (a) c)) =>  #f
    (lazy-member (lseq 'a)
            '(b (a) c))           =>  ((a) c)
    (lazy-memq 101 '(100 101 102))    =>  *unspecified*
    (lazy-memv 101 '(100 101 102))    =>  (101 102)

(lazy-find-tail even? lseq) ; Find the first elt with an even key.

Lazy association lists

An "lazy association list" (or "lazy alist") is an lseq of pairs. The car of each pair contains a key value, and the cdr contains the associated data value. They can be used to construct simple look-up tables in Scheme. Note that lazy alists are probably inappropriate for performance-critical use on large data; in these cases, immutable maps or some other alternative should be employed.

lazy-assoc key lazy-alist [=] -> lseq or #f

lazy-assq key lazy-alist -> lseq or #f

lazy-assv key lazy-alist -> lseq or #f

These procedures find the first pair in lazy-alist whose car field is key, and returns that pair. If no pair in lazy-alist has key as its car, then #f is returned. lazy-assq uses eq? to compare key with the car fields of the ipairs in lazy-alist, while lazy-assv uses eqv? and lazy-assoc uses =, which defaults to equal?.

(define e '((a 1) (b 2) (c 3)))
(lazy-assq 'a e)                              =>  (a 1)
(lazy-assq 'b e)                              =>  (b 2)
(lazy-assq 'd e)                              =>  #f
(lazy-assq (lseq 'a) '(((a)) ((b)) ((c))))  =>  #f
(lazy-assoc (lseq 'a) '(((a)) ((b)) ((c)))) =>  ((a))
(lazy-assq 5 '((2 3) (5 7) (11 13)))	      =>  *unspecified*
(lazy-assv 5 '((2 3) (5 7) (11 13)))	      =>  (5 7)

The comparison procedure is used to compare the elements e_i of lseq to the key parameter in this way:

(= key (lazy-first ei)) ; lseq is (E1 ... En)

That is, the first argument is always key, and the second argument is one of the lseq elements. Thus one can reliably find the first entry of lazy-alist whose key is greater than five with (lazy-assoc 5 lazy-alist <)

Note that fully general lazy alist searching may be performed with the lazy-find-tail and lazy-find procedures, e.g.

;; Look up the first association in lazy alist with an even key:
(lazy-find (lambda (a) (even? (lazy-first a))) lazy alist)

Comparators

lazy-pair-comparator

The lazy-comparator object is a SRFI-114 comparator suitable for comparing lseqs. Note that it is not a procedure. It compares lseqs using default-comparator on their first elements. If they are not equal, that value is returned. If they are equal, lazy-comparator is used on the lazy-rest of the lseqs and that value is returned.

make-lazy-comparator comparator -> comparator

The make-lazy-comparator procedure returns a comparator suitable for comparing lseqs using element-comparator to compare the elements.

Sample Implementation

The files in the implementation are as follows:

FIXME

Acknowledgements

Without the work of Olin Shivers on SRFI 1, this SRFI would not exist. Everyone acknowledged there is transitively acknowledged here. This is not to imply that either Olin or anyone else necessarily endorses the final results, of course.