"""Abstract supertype of [[categories|Category]] whose
elements may be iterated. Iterable categories are often
called _streams_. A stream need not be finite, but its
elements must be countable. That is, for any given
element of the stream, every iterator of the stream must
eventually return the element, even if the iterator
itself is not exhaustible.
A stream may have null elements. That is, an iterator for
the stream may produce the value [[null]] one or more
times. For every non-null `element` of a given stream
`it`, the expression `element in it` must evaluate to
`true`. Thus, a stream is a `Category` of its non-null
elements.
A given stream might not have a well-defined order, and
so the order in which elements are produced by the
stream's iterator may not be _stable_. That is, the order
may be different for two different iterators of the
stream. However, a stream has a well-defined set of
elements, and any two iterators for an immutable finite
stream should eventually return the same elements.
Furthermore, any two iterators for an immutable finite
stream should eventually return exactly the same number
of elements, which must be the [[size]] of the stream.
A given stream may not be _finite_, in which case an
iterator for the stream is never exhaustible, and certain
operations of this interface either never terminate or
result in an [[AssertionError]]. It may not, in general,
be possible to even determine if an `Iterable` is finite.
The type `Iterable<Element,Null>`, usually abbreviated
`{Element*}`, represents a possibly-empty iterable
container. The type `Iterable<Element,Nothing>`, usually
abbreviated `{Element+}`, represents a nonempty iterable
container.
A value list in braces produces a new instance of
`Iterable`:
{String+} words = { "hello", "world" };
An instance of `Iterable` may be iterated using a `for`
loop:
for (c in "hello world") { ... }
Comprehensions provide a convenient syntax for
transforming streams:
{Integer+} lengths = { for (w in words) w.size };
The `*.` operator may be used to evaluate an attribute
or invoke a method of the elements of the stream,
producing a new stream:
{Integer+} lengths = words*.size;
`Iterable` and its subtypes define various operations
that return other iterable objects. Such operations come
in two flavors:
- _Lazy_ operations return a *view* of the receiving
iterable object. If the underlying iterable object is
mutable, then changes to the underlying object will be
reflected in the resulting view. Lazy operations are
usually efficient, avoiding memory allocation or
iteration of the receiving iterable object.
- _Eager_ operations return an immutable object. If the
receiving iterable object is mutable, changes to this
object will not be reflected in the resulting immutable
object. Eager operations are often expensive, involving
memory allocation and iteration of the receiving
iterable object.
Lazy operations are generally preferred, because they can
be efficiently chained. For example:
string.filter((c) => c.letter||c.digit)
.map(Character.uppercased)
is much less expensive than:
string.select((c) => c.letter||c.digit)
.collect(Character.uppercased)
Furthermore, it is always easy to produce a new
immutable iterable object given the view produced by a
lazy operation. For example:
[ *string.filter((c) => c.letter||c.digit)
.map(Character.uppercased) ]
However, there are certain scenarios where an eager
operation is more useful, more convenient, or no more
expensive than a lazy operation, including:
- sorting operations, for example [[sort]], which are
eager by nature,
- operations which result in a subset or subrange of the
receiving stream, where structural sharing would or
could result in unnecessary memory retention.
Certain operations come in both lazy and eager flavors,
for example:
- [[map]] vs [[collect]],
- [[filter]] vs [[select]],
- [[List.sublist]] vs [[List.measure]].
Lazy operations normally return an instance of `Iterable`,
or even a [[List]], [[Map]], or [[Set]]. Eager operations
usually return a [[sequence|Sequential]]. The method
[[sequence]] materializes the current elements of a
stream into a sequence.
There is no meaningful generic definition of equality for
streams. For some streams—for example,
`List`s—order is significant; for others—for
example, `Set`s—order is not significant. Therefore,
unlike [[Collection]], `Iterable` does not define or
require any form of [[value equality|Object.equals]], and
some streams do not support value equality. It follows
that the `==` operator should not be used to compare
generic streams, unless the streams are known to share
some additional structure.
To compare two streams, taking order into account, use
the function [[corresponding]].
{Float*} xs = ... ;
{Float*} ys = ... ;
Boolean same = corresponding(xs, ys);"""
see (`interface Collection`, `function corresponding`)
by ("Gavin")
tagged("Streams")
shared interface Iterable<out Element=Anything,
out Absent=Null>
satisfies Category<>
given Absent satisfies Null {
"An iterator for the elements belonging to this stream."
shared formal Iterator<Element> iterator();
"Returns `true` if the iterator for this stream produces
the given element, or `false` otherwise. In the case of
an infinite stream, this operation might never terminate;
furthermore, this default implementation iterates all
the elements until found (or not), which might be very
expensive."
shared actual default Boolean contains(Object element)
=> any((e) => if (exists e) then e==element else false);
"Determines if the stream is empty, that is to say, if
the iterator returns no elements."
shared default Boolean empty
=> iterator().next() is Finished;
"The number of elements returned by the [[iterator]] of
this stream, if the iterator terminates. In the case of
an infinite stream, this operation never terminates."
shared default Integer size => count((e) => true);
"Determines if this stream has more elements than the
given [[length]]. This is an efficient operation for
streams with many elements."
see (`value size`)
shared default Boolean longerThan(Integer length) {
if (length<0) {
return true;
}
variable value count=0;
for (element in this) {
if (count++==length) {
return true;
}
}
return false;
}
"Determines if this stream has fewer elements than the
given [[length]]. This is an efficient operation for
streams with many elements."
see (`value size`)
shared default Boolean shorterThan(Integer length) {
if (length<=0) {
return false;
}
variable value count=0;
for (element in this) {
if (++count==length) {
return false;
}
}
return true;
}
"The first element returned by the iterator, if any, or
`null` if this stream is empty. For a stream with an
unstable iteration order, a different value might be
produced each time `first` is evaluated."
shared default Absent|Element first {
if (!is Finished first = iterator().next()) {
return first;
}
else {
"iterator for nonempty iterable must produce at
least one element"
assert (is Absent null);
return null;
}
}
"The last element returned by the iterator, if any, or
`null` if this stream is empty. In the case of an
infinite stream, this operation never terminates;
furthermore, this default implementation iterates all
elements, which might be very expensive."
shared default Absent|Element last {
variable Absent|Element e = first;
for (x in this) {
e = x;
}
return e;
}
"The [[index]]th element returned by an iterator of this
stream, or `null` if there are fewer than `index+1`
elements in the stream. For a stream with an unstable
iteration order, a different value might be produced
each time `getFromFirst(index)` is called for a given
integer `index`."
shared default Element? getFromFirst(Integer index) {
variable value current = 0;
for (element in this) {
if (current++==index) {
return element;
}
}
else {
return null;
}
}
"A [[sequence|Sequential]] containing all the elements
of this stream, in the same order they occur in this
stream. This operation eagerly evaluates and collects
every element of the stream."
shared default Element[] sequence()
=> let (array = Array(this))
if (array.empty)
then []
else ArraySequence(array);
"A [[Range]] containing all indexes of this stream, or
`[]` if this list is empty. The resulting range is
equal to `0:size`."
shared default Range<Integer>|[] indexes() => 0:size;
"A stream containing all but the first element of this
stream. For a stream with an unstable iteration order,
a different stream might be produced each time `rest`
is evaluated.
Therefore, if the stream `i` has an unstable iteration
order, the stream `{ i.first, *i.rest }` might not have
the same elements as `i`."
see (`value first`)
shared default {Element*} rest => skip(1);
"A stream containing all but the last element of this
stream. For a stream with an unstable iteration order,
a different stream might be produced each time
`exceptLast` is evaluated."
shared default {Element*} exceptLast
=> object satisfies {Element*} {
iterator()
=> let (iter = outer.iterator())
object satisfies Iterator<Element> {
variable value current = iter.next();
shared actual Element|Finished next() {
if (!is Finished next = iter.next()) {
value result = current;
current = next;
return result;
}
else {
return finished;
}
}
};
};
"Call the given [[function|step]] for each element of
this stream, passing the elements in the order they
occur in this stream.
For example:
words.each((word) {
print(word.lowercased);
print(word.uppercased);
});
Has the same effect as the following `for` loop:
for (word in words) {
print(word.lowercased);
print(word.uppercased);
}
_For certain streams this method is highly efficient,
surpassing the performance of `for` loops on the JVM.
Thus, `each()` is sometimes preferred in highly
performance-critical low-level code._"
shared default void each(
"The function to be called for each element in the
stream."
void step(Element element)) {
for (element in this) {
step(element);
}
}
"Produces a stream containing the results of applying
the given [[mapping|collecting]] to the elements of
this stream.
For any empty stream, `map()` returns an empty stream:
{}.map(f) == {}
For any nonempty stream `it`, and mapping function `f`,
the result of `map()` may be obtained according to this
recursive definition:
it.map(f).first == f(it.first)
it.map(f).rest == it.rest.map(f)
Alternatively, and in practice, `map()` may be defined
by this comprehension:
it.map(f) == { for (e in it) f(e) }
For example, the expression
(0..4).map(10.power)
results in the stream `{ 1, 10, 100, 1000, 10000 }`."
see (`function collect`)
shared default
Iterable<Result,Absent> map<Result>(
"The mapping to apply to the elements."
Result collecting(Element element))
=> { for (elem in this) collecting(elem) };
"Given a [[mapping function|collecting]] that accepts an
[[Element]] and returns a stream of [[Result]]s,
produces a new stream containing all elements of every
`Result` stream that results from applying the function
to the elements of this stream.
For example, the expression
{ \"Hello\", \"World\" }.flatMap(String.lowercased)
results in this stream:
{ 'h', 'e', 'l', 'l', 'o', 'w', 'o', 'r,' 'l', 'd' }
The expression
{ \"hello\"->\"hola\", \"world\"->\"mundo\" }
.flatMap(Entry<String,String>.pair)
produces this stream:
{ \"hello\", \"hola\", \"world\", \"mundo\" }"
see (`function expand`)
shared default
Iterable<Result,Absent|OtherAbsent>
flatMap<Result,OtherAbsent>(
"The mapping function to apply to the elements of
this stream, that produces a new stream of
[[Result]]s."
Iterable<Result,OtherAbsent> collecting(Element element))
given OtherAbsent satisfies Null
=> expand(map(collecting));
"Produces a stream containing the elements of this
stream that satisfy the given [[predicate
function|selecting]].
For any empty stream, `filter()` returns an empty
stream:
{}.filter(p) == {}
For any nonempty stream `it`, and predicate `p`, the
result of `filter()` may be obtained according to this
recursive definition:
it.filter(p) == { if (p(it.first)) it.first }.chain(it.rest.filter(f))
Alternatively, and in practice, `filter()` may be
defined by this comprehension:
it.filter(p) == { for (e in it) if (p(e)) e };
For example, the expression
(1..100).filter(13.divides)
results in the stream `{ 13, 26, 39, 52, 65, 78, 91 }`."
see (`function select`)
shared default
{Element*} filter(
"The predicate the elements must satisfy. The
elements which satisfy the predicate are included
in the resulting stream."
Boolean selecting(Element element))
=> { for (elem in this) if (selecting(elem)) elem };
"Produces a stream containing the elements of this
stream that are instances of the given [[type|Type]].
For example, the expression
{ 1, 2, null, 3 }.narrow<Object>()
results in the stream `{ 1, 2, 3 }` of type `{Integer*}`.
If the type argument `Type` is not explicitly specified,
[[Nothing]] is inferred, and the resulting stream is
empty."
shared default
{Element&Type*} narrow<Type>()
=> { for (elem in this) if (is Type elem) elem };
"Beginning with a given [[initial value|initial]], apply
the given [[combining function|accumulating]] to each
element of this stream in turn, progressively
accumulating a single result.
For an empty stream, `fold()` returns the given initial
value `z`:
{}.fold(z)(f) == z
For a given nonempty stream `it`, initial value `z`,
and combining function `f`, the result of `fold()` is
obtained according to the following recursive
definition:
it.fold(z)(f) == f(it.exceptLast.fold(z)(f), it.last)
For example, the expression
(1..100).fold(0)(plus)
results in the integer `5050`."
see (`function reduce`,
`function scan`)
shared default
Result fold<Result>(Result initial)(
"The accumulating function that accepts an
[[intermediate result|partial]], and the
[[next element|element]]."
Result accumulating(Result partial,
Element element)) {
variable value partial = initial;
for (elem in this) {
partial = accumulating(partial, elem);
}
return partial;
}
"Beginning with the [[first]] element of this stream,
apply the given [[combining function|accumulating]] to
each element of this stream in turn, progressively
accumulating a single result.
For an empty stream, `reduce()` always returns `null`.
For a stream with one element, `reduce()` returns that
element:
{ first }.reduce(f) == first
For a given stream `it` with more than one element, and
combining function `f`, the result of `reduce()` is
obtained according to the following recursive
definition:
it.reduce(f) == f(it.exceptLast.reduce(f), it.last)
For example, the expression
(1..100).reduce(plus)
results in the integer `5050`."
see (`function fold`)
shared default
Result|Element|Absent reduce<Result>(
"The accumulating function that accepts an
[[intermediate result|partial]], and the
[[next element|element]]."
Result accumulating(Result|Element partial,
Element element)) {
value it = iterator();
if (!is Finished initial = it.next()) {
variable Result|Element partial = initial;
while (!is Finished next = it.next()) {
partial = accumulating(partial, next);
}
return partial;
}
else {
"iterable must be empty"
assert (is Absent null);
return null;
}
}
"The stream of intermediate results obtained by
beginning with a given [[initial value|initial]] and
iteratively applying the given
[[combining function|accumulating]] to each element of
this stream in turn.
For an empty stream, `scan()` returns a stream
containing just the given initial value `z`:
{}.scan(z)(f) == { z }
For a given nonempty stream `it`, initial value `z`,
and combining function `f`, the result of `scan()` is
obtained according to the following recursive
definition:
it.scan(z)(f).last == f(it.exceptLast.scan(z)(f).last, it.last)
it.scan(z)(f).exceptLast == it.exceptLast.scan(z)(f)
The following identities explain the relationship
between `scan` and [[fold]]:
it.scan(z)(f).getFromFirst(n) == it.take(n).fold(z)(f)
it.scan(z)(f).last == it.fold(z)(f)
it.scan(z)(f).first == {}.fold(z)(f) == z
For example, the expression
(1..4).scan(0)(plus)
results in the stream `{ 0, 1, 3, 6, 10 }`."
see (`function fold`)
shared default
{Result+} scan<Result>(Result initial)(
"The accumulating function that accepts the
[[running total|partial]] and the
[[next element|element]]."
Result accumulating(Result partial,
Element element))
=> object satisfies {Result+} {
empty => false;
first => initial;
size => 1 + outer.size;
iterator()
=> let (iter = outer.iterator())
object satisfies Iterator<Result> {
variable value returnInitial = true;
variable value partial = initial;
shared actual Result|Finished next() {
if (returnInitial) {
returnInitial = false;
return initial;
}
else if (!is Finished element = iter.next()) {
partial = accumulating(partial, element);
return partial;
}
else {
return finished;
}
}
string => outer.string + ".iterator()";
};
};
"The first element of this stream which satisfies the
[[given predicate function|selecting]], if any, or
`null` if there is no such element. For an infinite
stream, this method might not terminate.
For example, the expression
(-10..10).find(Integer.positive)
evaluates to `1`."
see (`function findLast`, `function locate`)
shared default
Element? find(
"The predicate the element must satisfy."
Boolean selecting(Element&Object element)) {
for (elem in this) {
if (exists elem, selecting(elem)) {
return elem;
}
}
return null;
}
"The last element of this stream which satisfies the
[[given predicate function|selecting]], if any, or
`null` if there is no such element. For an infinite
stream, this method will not terminate.
For example, the expression
(-10..10).findLast(3.divides)
evaluates to `9`."
see (`function find`, `function locateLast`)
shared default
Element? findLast(
"The predicate the element must satisfy."
Boolean selecting(Element&Object element)) {
variable Element? last = null;
for (elem in this) {
if (exists elem, selecting(elem)) {
last = elem;
}
}
return last;
}
"The first element of this stream which satisfies the
[[given predicate function|selecting]], if any,
together with its position in the stream, or `null` if
there is no such element. For an infinite stream, this
method might not terminate.
For example, the expression
(-10..10).locate(Integer.positive)
evaluates to `11->1`."
see (`function locateLast`, `function locations`,
`function find`,
`function List.firstIndexWhere`)
shared default
<Integer->Element&Object>? locate(
"The predicate the element must satisfy."
Boolean selecting(Element&Object element)) {
variable value index = 0;
for (elem in this) {
if (exists elem, selecting(elem)) {
return index->elem;
}
index++;
}
return null;
}
"The last element of this stream which satisfies the
[[given predicate function|selecting]], if any,
together with its position in the stream, or `null` if
there is no such element. For an infinite stream, this
method might not terminate.
For example, the expression
(-10..10).locateLast(3.divides)
evaluates to `19->9`."
see (`function locate`, `function locations`,
`function findLast`,
`function List.lastIndexWhere`)
shared default
<Integer->Element&Object>? locateLast(
"The predicate the element must satisfy."
Boolean selecting(Element&Object element)) {
variable <Integer->Element&Object>? last = null;
variable value index = 0;
for (elem in this) {
if (exists elem, selecting(elem)) {
last = index->elem;
}
index++;
}
return last;
}
"A stream producing all elements of this stream which
satisfy the [[given predicate function|selecting]],
together with their positions in the stream.
For example, the expression
(-5..5).locations(3.divides)
evaluates to the stream `{ 2->-3, 5->0, 8->3 }`.
Note that this method is more efficient than the
alternative of applying [[filter]] to an [[indexed]]
stream."
see (`function locate`, `function locateLast`,
`function List.indexesWhere`)
shared default
{<Integer->Element&Object>*} locations(
"The predicate the element must satisfy."
Boolean selecting(Element&Object element))
=> object satisfies {<Integer->Element&Object>*} {
iterator()
=> let (iter = outer.iterator())
object satisfies Iterator<Integer->Element&Object> {
variable value i=0;
shared actual <Integer->Element&Object>|Finished next() {
while (!is Finished next = iter.next()) {
if (exists next, selecting(next)) {
return i++->next;
}
else {
i++;
}
}
return finished;
}
string => outer.string + ".iterator()";
};
};
"Return the largest value in the stream, as measured by
the given [[comparator function|comparing]] imposing a
partial order upon the elements of the stream, or `null`
if this stream is empty.
For example, the expression
{-10.0, -1.0, 5.0}.max(byIncreasing(Float.magnitude))
evaluates to `-10`.
For any nonempty stream `it`, and comparator function
`c`, `it.max(c)` evaluates to the first element of `it`
such that for every element `e` of `it`,
`c(e, it.max(c)) != larger`.
Note that the toplevel functions [[ceylon.language::max]]
and [[ceylon.language::min]] may be used to find the
largest and smallest values in a stream of [[Comparable]]
values, according to the natural order of its elements."
see (`function package.max`, `function package.min`)
shared default
Element|Absent max(
"The function comparing pairs of elements."
Comparison comparing(Element x, Element y)) {
value it = iterator();
if (!is Finished first = it.next()) {
variable value max = first;
while (!is Finished val = it.next()) {
if (comparing(val,max)==larger) {
max = val;
}
}
return max;
}
else {
"iterable must be empty"
assert (is Absent null);
return null;
}
}
"Given a [[method]] of the element type [[Element]],
return a function that, when supplied with a list of
method arguments, produces a new iterable object that
applies the `method` to each element of this iterable
object in turn.
{Boolean+}(Object) fun = (-1..1).spread(Object.equals);
print(fun(0)); //prints { false, true, false }"
shared default
Iterable<Result,Absent>(*Args)
spread<Result,Args>(Result(*Args) method(Element element))
given Args satisfies Anything[]
//=> flatten((Args args) => map(shuffle(method)(*args)));
=> flatten((Args args)
=> { for (elem in this) method(elem)(*args) });
/*"Produce a new [[sequence|Sequential]] containing the
elements of this stream, in the reverse order to the
order in which they occur in this stream.
That is, if a stream `i` is stable, and if `x` and `y`
are elements `i`, and `x` is produced before `y` by the
[[iterator]] for `i`, then `y` occurs before `x` in the
sequence `i.reverse()`.
This operation is an eager counterpart to
[[List.reversed]]."
see (`value List.reversed`)
shared default List<Element> reverse() {
value array = Array(this);
if (array.empty) {
return [];
}
else {
array.reverseInPlace();
return ArraySequence(array);
}
}*/
"Produce a new [[sequence|Sequential]] containing the
elements of this stream, sorted according to the given
[[comparator function|comparing]] imposing a partial
order upon the elements of the stream.
For convenience, the functions [[byIncreasing]] and
[[byDecreasing]] produce suitable comparator functions.
For example, this expression
\"Hello World!\".sort(byIncreasing(Character.lowercased))
evaluates to the sequence
`[ , !, d, e, H, l, l, l, o, o, r, W].`
This operation is eager by nature.
Note that the toplevel function [[ceylon.language::sort]]
may be used to sort a stream of [[Comparable]] values
according to the natural order of its elements."
see (`function increasing`, `function decreasing`,
`function byIncreasing`, `function byDecreasing`)
shared default
Element[] sort(
"The function comparing pairs of elements."
Comparison comparing(Element x, Element y)) {
value array = Array(this);
if (array.empty) {
return [];
}
else {
array.sortInPlace(comparing);
return ArraySequence(array);
}
}
"Produce a new [[sequence|Sequential]] containing the
results of applying the given [[mapping|collecting]] to
the elements of this stream.
This operation is an eager counterpart to [[map]]. For
any stream `it`, and mapping `f`:
it.collect(f) == [*it.map(f)]"
see (`function map`)
shared default
Result[] collect<Result>(
"The transformation applied to the elements."
Result collecting(Element element))
=> map(collecting).sequence();
"Produce a new [[sequence|Sequential]] containing all
elements of this stream that satisfy the given
[[predicate function|selecting]], in the order in which
they occur in this stream.
This operation is an eager counterpart to [[filter]].
For any stream `it`, and predicate `p`:
it.select(p) == [*it.filter(p)]"
see (`function filter`)
shared default
Element[] select(
"The predicate the elements must satisfy."
Boolean selecting(Element element))
=> filter(selecting).sequence();
"Produces the number of elements in this stream that
satisfy the [[given predicate function|selecting]].
For an infinite stream, this method never terminates."
shared default
Integer count(
"The predicate satisfied by the elements to be
counted."
Boolean selecting(Element element)) {
variable value count=0;
for (elem in this) {
if (selecting(elem)) {
count++;
}
}
return count;
}
"Determines if there is at least one element of this
stream that satisfies the given [[predicate
function|selecting]]. If the stream is empty, returns
`false`. For an infinite stream, this operation might
not terminate."
see (`function every`)
shared default
Boolean any(
"The predicate that at least one element must
satisfy."
Boolean selecting(Element element)) {
for (e in this) {
if (selecting(e)) {
return true;
}
}
return false;
}
"Determines if all elements of this stream satisfy the
given [[predicate function|selecting]]. If the stream
is empty, return `true`. For an infinite stream, this
operation might not terminate."
see (`function any`)
shared default
Boolean every(
"The predicate that all elements must satisfy."
Boolean selecting(Element element)) {
for (e in this) {
if (!selecting(e)) {
return false;
}
}
return true;
}
"Produces a stream containing the elements of this
stream, after skipping the first [[skipping]] elements
produced by its iterator.
If this stream does not contain more elements than the
specified number of elements to skip, the resulting
stream has no elements. If the specified number of
elements to skip is zero or fewer, the resulting stream
contains the same elements as this stream."
see (`function List.sublistFrom`,
`function skipWhile`,
`function take`)
shared default
{Element*} skip(Integer skipping) {
if (skipping <= 0) {
return this;
}
else {
return object satisfies {Element*} {
shared actual Iterator<Element> iterator() {
value iter = outer.iterator();
variable value i=0;
while (i++<skipping &&
!iter.next() is Finished) {}
return iter;
}
string => outer.string + ".iterator()";
};
}
}
"Produces a stream containing the first [[taking]]
elements of this stream.
If the specified number of elements to take is larger
than the number of elements of this stream, the
resulting stream contains the same elements as this
stream. If the specified number of elements to take is
fewer than one, the resulting stream has no elements."
see (`function List.sublistTo`,
`function List.initial`,
`function takeWhile`,
`function skip`)
shared default
{Element*} take(Integer taking) {
if (taking <= 0) {
return {};
}
else {
return object satisfies {Element*} {
shared actual Iterator<Element> iterator() {
value iter = outer.iterator();
return object satisfies Iterator<Element> {
variable value i=0;
next() => ++i>taking
then finished
else iter.next();
string => outer.string + ".iterator()";
};
}
first => outer.first;
};
}
}
"Produces a stream containing the elements of this
stream, after skipping the leading elements until the
given [[predicate function|skipping]] returns `false`."
see (`function skip`,
`function takeWhile`)
shared default
{Element*} skipWhile(
"The function that returns `false` when the
resulting stream should stop skipping elements from
the stream."
Boolean skipping(Element element))
=> object satisfies {Element*} {
shared actual Iterator<Element> iterator() {
value iter = outer.iterator();
while (!is Finished elem = iter.next()) {
if (!skipping(elem)) {
return object satisfies Iterator<Element> {
variable Boolean first=true;
actual shared Element|Finished next() {
if (first) {
first = false;
return elem;
}
else {
return iter.next();
}
}
string => outer.string + ".iterator()";
};
}
}
return emptyIterator;
}
};
"Produces a stream containing the leading elements of
this stream until the given [[predicate function|taking]]
returns `false`."
see (`function take`,
`function skipWhile`)
shared default
{Element*} takeWhile(
"The function that returns `false` when the
resulting stream should stop taking elements from
this stream."
Boolean taking(Element element))
=> object satisfies {Element*} {
iterator()
=> let (iter = outer.iterator())
object satisfies Iterator<Element> {
variable Boolean alive = true;
actual shared Element|Finished next() {
if (alive,
!is Finished next = iter.next()) {
if (taking(next)) {
return next;
}
else {
alive = false;
}
}
return finished;
}
string => outer.string + ".iterator()";
};
};
"Produces a stream formed by repeating the elements of
this stream the given [[number of times|times]], or an
empty stream if `times<=0`.
For example, the expression
{ 1, 2 }.repeat(3)
evaluates to the stream `{ 1, 2, 1, 2, 1, 2 }`."
see (`value cycled`)
shared default
{Element*} repeat(Integer times)
=> object satisfies {Element*} {
size => times * outer.size;
string => "(``outer.string``).repeat(``times``)";
iterator() => CycledIterator(outer,times);
};
"Produces a stream containing every [[step]]th element
of this stream. If the step size is `1`, the resulting
stream contains the same elements as this stream.
For example, the expression
(0..10).by(3)
results in the stream `{ 0, 3, 6, 9 }`.
The step size must be greater than zero."
throws (`class AssertionError`,
"if the given step size is nonpositive,
i.e. `step<1`")
shared default
Iterable<Element,Absent> by(Integer step) {
"step size must be greater than zero"
assert (step > 0);
if (step == 1) {
return this;
}
else {
return object satisfies Iterable<Element,Absent> {
string => "(``outer.string``).by(``step``)";
iterator()
=> let (iter = outer.iterator())
object satisfies Iterator<Element> {
shared actual Element|Finished next() {
value next = iter.next();
variable value i=0;
while (++i<step &&
!iter.next() is Finished) {}
return next;
}
string => outer.string + ".iterator()";
};
};
}
}
"Produces a stream containing the elements of this
stream, in the order in which they occur in this stream,
after replacing every `null` element in the stream with
the [[given default value|defaultValue]]. The value
`null` does not ocur in the resulting stream.
For example, the expression
{ \"123\", \"abc\", \"456\" }.map(parseInteger).defaultNullElements(0)
results in the stream `{ 123, 0, 456 }`."
see (`value coalesced`)
shared default
Iterable<Element&Object|Default,Absent>
defaultNullElements<Default>(
"A default value that replaces `null` elements."
Default defaultValue)
given Default satisfies Object
=> map((elem) => elem else defaultValue);
"The non-null elements of this stream, in the order in
which they occur in this stream. For null elements of
the original stream, there is no entry in the resulting
stream.
For example, the expression
{ \"123\", \"abc\", \"456\"}.map(parseInteger).coalesced
results in the stream `{ 123, 456 }`."
see (`function defaultNullElements`)
shared default
{Element&Object*} coalesced
=> { for (e in this) if (exists e) e };
"A stream containing all [[entries|Entry]] of form
`index->element` where `element` is an element of this
stream, and `index` is the position at which `element`
occurs in this stream, ordered by increasing `index`.
For example, the expression
{ \"hello\", null, \"world\" }.indexed
results in the stream `{ 0->\"hello\", 1->null, 2->\"world\" }`."
see (`function locations`)
shared default
Iterable<<Integer->Element>,Absent> indexed
=> object
satisfies Iterable<<Integer->Element>,Absent> {
iterator()
=> let (iter = outer.iterator())
object satisfies Iterator<Integer->Element> {
variable value i=0;
next() => switch (next = iter.next())
case (finished) finished
else i++ -> next;
string => outer.string + ".iterator()";
};
};
"A stream containing whose elements are pairs (2-tuples)
comprising an element of this stream paired with the
next element in the stream. The resulting stream has
one fewer elements than this stream.
For example, the expression
(1..5).paired
results in the stream
`{ [1, 2], [2, 3], [3, 4], [4, 5] }`.
This expression determines if a stream is monotonically
increasing:
every { for ([x, y] in nums.paired) x < y }
For any stable `stream`, this operation is equivalent
to `zipPairs(stream,stream.rest)`."
shared default
{[Element,Element]*} paired
=> object satisfies {[Element,Element]*} {
size => let (size = outer.size-1)
if (size<0) then 0 else size;
empty => outer.size<2;
iterator()
=> let (iter = outer.iterator())
object satisfies Iterator<[Element,Element]> {
variable value previous = iter.next();
shared actual [Element,Element]|Finished next() {
if (!is Finished head = previous,
!is Finished tip = iter.next()) {
previous = tip;
return [head, tip];
}
else {
return finished;
}
}
};
};
"Produces a stream of sequences of the given [[length]],
containing elements of this stream. Each sequence in
the stream contains the next [[length]] elements of
this sequence that have not yet been assigned to a
previous sequence, in the same order that they occur in
this stream. The very last sequence in the stream may
be shorter than the given `length`.
For example, the expression
\"hello\".partition(2)
results in the stream `{ ['h','e'], ['l','l'], ['o'] }.`
For any `stream` and for any strictly positive integer
[[length]]:
expand { stream.partition(length) } == stream"
throws (`class AssertionError`,
"if `length<=0`")
shared default
Iterable<[Element+],Absent> partition(
"The length of the sequences in the resulting stream,
which must be strictly positive."
Integer length) {
"length must be strictly positive"
assert (length>0);
return object satisfies Iterable<[Element+],Absent> {
size => let (outerSize = outer.size,
quotient = outerSize/length)
if (length.divides(outerSize))
then quotient
else quotient+1;
empty => outer.empty;
iterator()
=> let (iter = outer.iterator())
object satisfies Iterator<[Element+]> {
shared actual [Element+]|Finished next() {
if (!is Finished next = iter.next()) {
value array =
Array.ofSize(length, next);
variable value index = 0;
while (++index<length) {
if (!is Finished current
= iter.next()) {
array.set(index, current);
}
else {
return ArraySequence(
array[...index-1]);
}
}
return ArraySequence(array);
}
else {
return finished;
}
}
};
};
}
"Produces a stream with a given [[initial element|head]],
followed by the elements of this stream, in the order
in which they occur in this stream.
For example, the expression
(1..3).follow(0)
evaluates to the stream `{ 0, 1, 2, 3 }`.
Note that the expression `stream.follow(head)` eagerly
evaluates `head`, and therefore is not precisely the
same as this enumeration expression, where `head` is
evaluated lazily:
{ head, *stream }"
see (`function chain`)
shared default
{Element|Other+} follow<Other>(Other head)
=> { head, *this };
"The elements of this stream, in the order in which they
occur in this stream, followed by the elements of the
[[given stream|other]] in the order in which they occur
in the given stream.
For example, the expression
(1..3).chain(\"abc\")
evaluates to the stream `{ 1, 2, 3, 'a', 'b', 'c' }`."
see (`function expand`)
shared default
Iterable<Element|Other,Absent&OtherAbsent>
chain<Other,OtherAbsent>
(Iterable<Other,OtherAbsent> other)
given OtherAbsent satisfies Null
=> object
satisfies Iterable<Element|Other,
Absent&OtherAbsent> {
iterator() => ChainedIterator(outer, other);
};
"A stream of pairs of elements of this stream and the
the given stream, where for each element `x` of this
stream, and element `y` of the given stream, the pair
`[x,y]` belongs to the resulting stream. The pairs are
sorted first by the position of `x` in this stream, and
then by the position of `y` in the given stream.
For example, this expression
(1..3).product(\"ab\")
evaluates to the stream
`{ [1,'a'], [1,'b'], [2,'a'], [2,'b'], [3,'a'], [3,'b'] }`."
shared default
Iterable<[Element,Other],Absent|OtherAbsent>
product<Other,OtherAbsent>
(Iterable<Other,OtherAbsent> other)
given OtherAbsent satisfies Null
=> { for (x in this) for (y in other) [x,y] };
"An infinite stream that produces the elements of this
stream, repeatedly.
For example, the expression
{6, 9}.cycled.take(5)
evaluates to the stream `{ 6, 9, 6, 9, 6 }`.
If this stream is empty, the resulting stream also
empty."
see (`function repeat`)
shared default
Iterable<Element,Absent> cycled
=> object satisfies Iterable<Element,Absent> {
value orig => outer;
string => "(``outer.string``).cycled";
shared actual Integer size {
"stream is infinite"
assert (false);
}
iterator()
=> object satisfies Iterator<Element> {
variable Iterator<Element> iter
= emptyIterator;
shared actual Element|Finished next() {
if (!is Finished next = iter.next()) {
return next;
}
else {
iter = orig.iterator();
return iter.next();
}
}
string => outer.string + ".iterator()";
};
};
"A stream that contains the given [[element]] interposed
between blocks of [[step]] elements of this stream. The
resulting stream starts with the [[first]] element of
this stream and ends with the [[last]] element of this
stream. Elements of this stream occur in the resulting
stream in the same order they occur in this stream.
For example, the expression
String(\"hello\".interpose(' '))
evaluates to the string `\"h e l l o\"`."
throws (`class AssertionError`, "if `step<1`")
see (`function interleave`)
shared default
Iterable<Element|Other,Absent> interpose<Other>(
"The value to interpose between blocks of
elements of this stream."
Other element,
"The step size that determines how often the
given [[element]] occurs in the resulting
stream. The `element` occurs after each block
of size `step` of elements of this stream. If
`step==1`, the `element` occurs at every second
position. The step size must be strictly positive."
Integer step=1) {
"step must be strictly positive"
assert (step>=1);
return object
satisfies Iterable<Element|Other,Absent> {
shared actual Integer size {
value outerSize = outer.size;
return if (outerSize>0)
then outerSize + (outerSize-1)/step
else 0;
}
empty => outer.empty;
first => outer.first;
last => outer.last;
iterator()
=> let (iter = outer.iterator())
object satisfies Iterator<Element|Other> {
variable value current = iter.next();
variable value count = 0;
shared actual Element|Other|Finished next() {
if (!is Finished curr = current) {
if ((step+1).divides(++count)) {
return element;
}
else {
current = iter.next();
return curr;
}
}
else {
return finished;
}
}
string => outer.string + ".iterator()";
};
};
}
"A stream that produces every element produced by this
stream exactly once. Duplicate elements of this stream
are eliminated. Two elements are considered distinct
unless they are both [[null|Null]], or unless they are
both non-null and [[equal|Object.equals]].
For example:
String(\"hello world\".distinct)
is the string `\"helo wrd\"`.
This is a lazy operation and the resulting stream
reflects changes to this stream."
see(`function set`)
shared default Iterable<Element,Absent> distinct
=> object satisfies Iterable<Element,Absent> {
iterator()
=> let (elements=outer)
object satisfies Iterator<Element> {
value it = elements.iterator();
variable value count = 0;
variable value store
= Array.ofSize {
size = 16;
element = null of ElementEntry<Element>?;
};
function hash(Element element, Integer size)
=> if (exists element)
then element.hash.magnitude % size
else 0;
function rebuild(Array<ElementEntry<Element>?> store) {
value newStore
= Array.ofSize {
size = store.size*2;
element = null of ElementEntry<Element>?;
};
for (entries in store) {
variable value entry = entries;
while (exists e = entry) {
value index =
hash(e.element,
newStore.size);
newStore.set(index,
ElementEntry {
next = newStore[index];
element = e.element;
});
entry = e.next;
}
}
return newStore;
}
shared actual Element|Finished next() {
while (true) {
switch (element = it.next())
case (is Finished) {
return element;
}
else {
value index = hash(element, store.size);
value entry = store[index];
if (exists entry, entry.has(element)) {
//keep iterating
}
else {
store.set(index,
ElementEntry {
next = entry;
element = element;
});
count++;
if (count>store.size*2) {
store = rebuild(store);
}
return element;
}
}
}
}
};
};
/*"Create a new immutable [[Set]] containing every element
produced by this stream that is not null.
For example:
{0, 1, 1, 2, 3, null, 3, 3}.elements()
is the set `{0, 1, 2, 3}`.
This is an eager operation and the resulting set does
not reflect changes to this stream."
aliased("set")
see(`value distinct`)
shared default Set<Element&Object> elements()
=> object extends Object()
satisfies Set<Element&Object> {
value elements =
outer.coalesced
.summarize(identity,
(Boolean? _, e) => true);
clone() => this;
iterator() => elements.keys.iterator();
contains(Object element) => elements.defines(element);
shared actual Set<Element&Object>
complement<Other>(Set<Other> set)
given Other satisfies Object
=> filter((e) => !e in set)
.elements();
shared actual Set<Element&Object|Other>
exclusiveUnion<Other>(Set<Other> set)
given Other satisfies Object
=> filter((e) => !e in set)
.chain(set.filter((e) => !e in this))
.elements();
shared actual Set<Element&Other&Object>
intersection<Other>(Set<Other> set)
given Other satisfies Object
=> filter((e) => e in set)
.narrow<Other>()
.elements();
shared actual Set<Element&Object|Other>
union<Other>(Set<Other> set)
given Other satisfies Object
=> chain(set)
.elements();
};*/
"Produce a [[Map]] mapping elements to frequencies where
each [[entry|Entry]] maps a distinct non-null element
of this stream to the number of times the element was
produced by the stream. Elements are considered
distinct if they are not [[equal|Object.equals]]. Null
elements of this stream are simply discarded.
This is an eager operation, and the resulting map does
not reflect changes to this stream."
shared Map<Element&Object,Integer> frequencies()
=> coalesced.summarize(identity,
(Integer? count, _)
=> if (exists count)
then count+1 else 1);
"Produces a [[Map]] mapping elements to items where each
[[entry|Entry]] maps a distinct non-null element of
this stream to the item produced by the given
[[function|collecting]]. Elements are considered
distinct if they are not [[equal|Object.equals]]. Null
elements of this stream are simply discarded.
This is an eager operation, and the resulting map does
not reflect changes to the given stream."
shared Map<Element&Object,Result> tabulate<Result>(
"A function that produces an item for the given
[[key]], an element of this stream."
Result collecting(Element key))
=> coalesced.summarize(identity,
(Result? item, key)
=> if (exists item)
then item else collecting(key));
"Classifies the elements of this stream into a new
immutable [[Map]] where each key is a value produced by
the given [[grouping function|grouping]] and each
corresponding item is [[sequence|Sequence]] of all
elements that produced the key when passed as arguments
to the grouping function.
Within each group, the sequence elements occur in the
same order they occurred in this stream.
For example:
(0..10).group((i) => i.even then \"even\" else \"odd\")
produces the map
`{ even->[0, 2, 4, 6, 8, 10], odd->[1, 3, 5, 7, 9] }`.
This is an eager operation, and the resulting map does
not reflect changes to this stream."
see(`function summarize`)
shared Map<Group,[Element+]> group<Group>(
"The grouping function that assigns a key to the
given [[element]]. Multiple elements may be
assigned to the same key, indicating that they
belong to the same [[Group]] in the resulting map."
Group grouping(Element element))
given Group satisfies Object
=> summarize<Group,ElementEntry<Element>>
(grouping, ElementEntry)
.mapItems((_, item) => item.reversed);
"Efficiently [[group]] and [[fold]] the elements of this
stream in a single step.
For example, the expression:
(1..10)
.summarize((i) => i%3,
(Integer[2]? pair, i)
=> if (exists [sum, product] = pair)
then [sum+i, product*i] else [i,i])
produces the map
`{ 0->[18, 162], 1->[22, 280], 2->[15, 80] }`, being
equivalent to, but much more efficient than, the
following expression written using `group()`,
`mapItems()` and `fold()`:
(1..10)
.group((i) => i%3)
.mapItems((_, item)
=> item.fold([0,1])
((pair, i)
=> let ([sum, product] = pair)
[sum+i, product*i]))
This is an eager operation, and the resulting map does
not reflect changes to this stream."
see(`function group`, `function fold`)
shared Map<Group,Result> summarize<Group,Result>(
"The grouping function that assigns a key to the
given [[element]]. Multiple elements may be
assigned to the same key, indicating that they
should be aggregated by calling [[accumulating]]."
Group grouping(Element element),
"The accumulating function that accepts an
[[intermediate result|partial]] for a key, and the
[[next element]] with that key."
Result accumulating(Result? partial, Element element))
given Group satisfies Object
=> object extends Object()
satisfies Map<Group,Result> {
variable value store
= Array.ofSize {
size = 16;
element = null of GroupEntry<Group,Result>?;
};
function hash(Object group, Integer size)
=> group.hash.magnitude % size;
function rebuild
(Array<GroupEntry<Group,Result>?> store) {
value newStore
= Array.ofSize {
size = store.size*2;
element = null of GroupEntry<Group,Result>?;
};
for (groups in store) {
variable value group = groups;
while (exists g = group) {
value index =
hash(g.group, newStore.size);
newStore.set(index,
GroupEntry {
next = newStore[index];
group = g.group;
elements = g.elements;
});
group = g.next;
}
}
return newStore;
}
variable value count = 0;
for (element in outer) {
value group = grouping(element);
value index = hash(group, store.size);
value entries = store[index];
if (exists entries,
exists entry = entries.get(group)) {
entry.elements
= accumulating(entry.elements,
element);
//keep iterating
}
else {
store.set(index,
GroupEntry {
next = entries;
group = group;
elements = accumulating(null, element);
});
count++;
if (count>store.size*2) {
store = rebuild(store);
}
}
}
size => count;
iterator()
=> object satisfies Iterator<Group->Result> {
variable value index = 0;
variable GroupEntry<Group,Result>? entry = null;
shared actual <Group->Result>|Finished next() {
GroupEntry<Group,Result> result;
if (exists e = entry) {
entry = e.next;
result = e;
}
else {
while (true) {
if (index>=store.size) {
return finished;
}
else {
entry = store[index++];
if (exists e = entry) {
entry = e.next;
result = e;
break;
}
}
}
}
return result.group -> result.elements;
}
};
clone() => this;
function group(Object key)
=> store[hash(key, store.size)]?.get(key);
defines(Object key) => group(key) exists;
get(Object key) => group(key)?.elements;
shared actual Result|Default getOrDefault<Default>
(Object key, Default default)
=> if (exists group = group(key))
then group.elements else default;
};
"A string of form `\"{ x, y, z }\"` where `x`, `y`, and
`z` are the `string` representations of the elements of
this collection, as produced by the iterator of the
stream, or the string `\"{}\"` if this stream is empty.
If the stream is very long, the list of elements might
be truncated, as indicated by an ellipse."
shared actual default String string
=> let (elements = take(31).sequence())
if (elements.empty) then
"{}"
else if (elements.size==31) then
"{ ``commaList(elements.take(30))``, ... }"
else
"{ ``commaList(elements)`` }";
}
String commaList({Anything*} elements)
=> ", ".join { for (e in elements) stringify(e) };
class ElementEntry<Element>(next, element)
extends Object()
satisfies [Element+] {
shared Element element;
shared ElementEntry<Element>? next;
first => element;
shared Boolean has(Anything element) {
variable ElementEntry<Element>? entry = this;
while (exists e = entry) {
if (exists element) {
if (exists ee = e.element,
element == ee) {
return true;
}
}
else {
if (!e.element exists) {
return true;
}
}
entry = e.next;
}
return false;
}
shared actual Element? getFromFirst(Integer index) {
if (index<0) {
return null;
}
else {
variable ElementEntry<Element> entry = this;
for (i in 0:index) {
if (exists next = entry.next) {
entry = next;
}
else {
return null;
}
}
return entry.element;
}
}
rest => next else [];
shared actual Element last {
variable ElementEntry<Element> entry = this;
while (exists next = entry.next) {
entry = next;
}
return entry.element;
}
shared actual Integer size {
variable value count = 1;
variable ElementEntry<Element> entry = this;
while (exists next = entry.next) {
entry = next;
count++;
}
return count;
}
iterator()
=> object satisfies Iterator<Element> {
variable ElementEntry<Element>? entry = outer;
shared actual Element|Finished next() {
if (exists e = entry) {
entry = e.next;
return e.element;
}
else {
return finished;
}
}
};
}
class GroupEntry<Group,Result>(next, group, elements)
given Group satisfies Object {
shared Group group;
shared variable Result elements;
shared GroupEntry<Group,Result>? next;
shared GroupEntry<Group,Result>? get(Object group) {
variable GroupEntry<Group,Result>? entry = this;
while (exists e = entry) {
if (group == e.group) {
return e;
}
entry = e.next;
}
return null;
}
}