mirror of
https://github.com/dlang/phobos.git
synced 2025-04-29 14:40:30 +03:00
Add splitter functionality that allows to keep the sentinels
Signed-off-by: Robert Aron <aronrobert293@gmail.com> Update the Return: section of DDoc Fix whitespace
This commit is contained in:
Robert Aron
The Dlang Bot
parent
fb6267b6b3
commit
d8f3d3a815
1 changed files with 282 additions and 43 deletions
|
|
@ -75,7 +75,7 @@ module std.algorithm.iteration;
|
|||
import std.functional : unaryFun, binaryFun;
|
||||
import std.range.primitives;
|
||||
import std.traits;
|
||||
import std.typecons : Flag;
|
||||
import std.typecons : Flag, Yes, No;
|
||||
|
||||
private template aggregate(fun...)
|
||||
if (fun.length >= 1)
|
||||
|
|
@ -4824,6 +4824,7 @@ Params:
|
|||
s = The element (or range) to be treated as the separator
|
||||
between range segments to be split.
|
||||
isTerminator = The predicate for deciding where to split the range when no separator is passed
|
||||
keepSeparators = The flag for deciding if the separators are kept
|
||||
|
||||
Constraints:
|
||||
The predicate `pred` needs to accept an element of `r` and the
|
||||
|
|
@ -4836,6 +4837,9 @@ Returns:
|
|||
the returned range will be likewise.
|
||||
When a range is used a separator, bidirectionality isn't possible.
|
||||
|
||||
If keepSeparators is equal to Yes.keepSeparators the output will also contain the
|
||||
separators.
|
||||
|
||||
If an empty range is given, the result is an empty range. If a range with
|
||||
one separator is given, the result is a range with two empty elements.
|
||||
|
||||
|
|
@ -4844,7 +4848,10 @@ See_Also:
|
|||
$(REF _split, std,array) for a version that splits eagerly and
|
||||
$(LREF splitWhen), which compares adjacent elements instead of element against separator.
|
||||
*/
|
||||
auto splitter(alias pred = "a == b", Range, Separator)(Range r, Separator s)
|
||||
auto splitter(alias pred = "a == b",
|
||||
Flag!"keepSeparators" keepSeparators = No.keepSeparators,
|
||||
Range,
|
||||
Separator)(Range r, Separator s)
|
||||
if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
||||
&& ((hasSlicing!Range && hasLength!Range) || isNarrowString!Range))
|
||||
{
|
||||
|
|
@ -4870,6 +4877,11 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
enum _separatorLength = 1;
|
||||
}
|
||||
|
||||
static if (keepSeparators)
|
||||
{
|
||||
bool _wasSeparator = true;
|
||||
}
|
||||
|
||||
static if (isBidirectionalRange!Range)
|
||||
{
|
||||
size_t lastIndexOf(Range haystack, Separator needle)
|
||||
|
|
@ -4911,10 +4923,27 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
@property Range front()
|
||||
{
|
||||
assert(!empty, "Attempting to fetch the front of an empty splitter.");
|
||||
if (_frontLength == _unComputed)
|
||||
static if (keepSeparators)
|
||||
{
|
||||
auto r = _input.find!pred(_separator);
|
||||
_frontLength = _input.length - r.length;
|
||||
if (!_wasSeparator)
|
||||
{
|
||||
_frontLength = _separatorLength;
|
||||
_wasSeparator = true;
|
||||
}
|
||||
else if (_frontLength == _unComputed)
|
||||
{
|
||||
auto r = _input.find!pred(_separator);
|
||||
_frontLength = _input.length - r.length;
|
||||
_wasSeparator = false;
|
||||
}
|
||||
}
|
||||
else
|
||||
{
|
||||
if (_frontLength == _unComputed)
|
||||
{
|
||||
auto r = _input.find!pred(_separator);
|
||||
_frontLength = _input.length - r.length;
|
||||
}
|
||||
}
|
||||
return _input[0 .. _frontLength];
|
||||
}
|
||||
|
|
@ -4928,18 +4957,35 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
}
|
||||
assert(_frontLength <= _input.length, "The front position must"
|
||||
~ " not exceed the input.length");
|
||||
if (_frontLength == _input.length)
|
||||
static if (keepSeparators)
|
||||
{
|
||||
// no more input and need to fetch => done
|
||||
_frontLength = _atEnd;
|
||||
if (_frontLength == _input.length && !_wasSeparator)
|
||||
{
|
||||
_frontLength = _atEnd;
|
||||
|
||||
// Probably don't need this, but just for consistency:
|
||||
_backLength = _atEnd;
|
||||
_backLength = _atEnd;
|
||||
}
|
||||
else
|
||||
{
|
||||
_input = _input[_frontLength .. _input.length];
|
||||
_frontLength = _unComputed;
|
||||
}
|
||||
}
|
||||
else
|
||||
{
|
||||
_input = _input[_frontLength + _separatorLength .. _input.length];
|
||||
_frontLength = _unComputed;
|
||||
if (_frontLength == _input.length)
|
||||
{
|
||||
// no more input and need to fetch => done
|
||||
_frontLength = _atEnd;
|
||||
|
||||
// Probably don't need this, but just for consistency:
|
||||
_backLength = _atEnd;
|
||||
}
|
||||
else
|
||||
{
|
||||
_input = _input[_frontLength + _separatorLength .. _input.length];
|
||||
_frontLength = _unComputed;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -4958,16 +5004,40 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
@property Range back()
|
||||
{
|
||||
assert(!empty, "Attempting to fetch the back of an empty splitter.");
|
||||
if (_backLength == _unComputed)
|
||||
static if (keepSeparators)
|
||||
{
|
||||
immutable lastIndex = lastIndexOf(_input, _separator);
|
||||
if (lastIndex == -1)
|
||||
if (!_wasSeparator)
|
||||
{
|
||||
_backLength = _input.length;
|
||||
_backLength = _separatorLength;
|
||||
_wasSeparator = true;
|
||||
}
|
||||
else
|
||||
else if (_backLength == _unComputed)
|
||||
{
|
||||
_backLength = _input.length - lastIndex - 1;
|
||||
immutable lastIndex = lastIndexOf(_input, _separator);
|
||||
if (lastIndex == -1)
|
||||
{
|
||||
_backLength = _input.length;
|
||||
}
|
||||
else
|
||||
{
|
||||
_backLength = _input.length - lastIndex - 1;
|
||||
}
|
||||
_wasSeparator = false;
|
||||
}
|
||||
}
|
||||
else
|
||||
{
|
||||
if (_backLength == _unComputed)
|
||||
{
|
||||
immutable lastIndex = lastIndexOf(_input, _separator);
|
||||
if (lastIndex == -1)
|
||||
{
|
||||
_backLength = _input.length;
|
||||
}
|
||||
else
|
||||
{
|
||||
_backLength = _input.length - lastIndex - 1;
|
||||
}
|
||||
}
|
||||
}
|
||||
return _input[_input.length - _backLength .. _input.length];
|
||||
|
|
@ -4983,16 +5053,32 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
}
|
||||
assert(_backLength <= _input.length, "The end index must not"
|
||||
~ " exceed the length of the input");
|
||||
if (_backLength == _input.length)
|
||||
static if (keepSeparators)
|
||||
{
|
||||
// no more input and need to fetch => done
|
||||
_frontLength = _atEnd;
|
||||
_backLength = _atEnd;
|
||||
if (_backLength == _input.length && !_wasSeparator)
|
||||
{
|
||||
_frontLength = _atEnd;
|
||||
_backLength = _atEnd;
|
||||
}
|
||||
else
|
||||
{
|
||||
_input = _input[0 .. _input.length - _backLength];
|
||||
_backLength = _unComputed;
|
||||
}
|
||||
}
|
||||
else
|
||||
{
|
||||
_input = _input[0 .. _input.length - _backLength - _separatorLength];
|
||||
_backLength = _unComputed;
|
||||
if (_backLength == _input.length)
|
||||
{
|
||||
// no more input and need to fetch => done
|
||||
_frontLength = _atEnd;
|
||||
_backLength = _atEnd;
|
||||
}
|
||||
else
|
||||
{
|
||||
_input = _input[0 .. _input.length - _backLength - _separatorLength];
|
||||
_backLength = _unComputed;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
|
@ -5013,6 +5099,20 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
assert(a.splitter(0).equal(w));
|
||||
}
|
||||
|
||||
/// Basic splitting with characters and numbers and keeping sentinels.
|
||||
@safe unittest
|
||||
{
|
||||
import std.algorithm.comparison : equal;
|
||||
import std.typecons : Yes;
|
||||
|
||||
assert("a|bc|def".splitter!("a == b", Yes.keepSeparators)('|')
|
||||
.equal([ "a", "|", "bc", "|", "def" ]));
|
||||
|
||||
int[] a = [1, 0, 2, 3, 0, 4, 5, 6];
|
||||
int[][] w = [ [1], [0], [2, 3], [0], [4, 5, 6] ];
|
||||
assert(a.splitter!("a == b", Yes.keepSeparators)(0).equal(w));
|
||||
}
|
||||
|
||||
/// Adjacent separators.
|
||||
@safe unittest
|
||||
{
|
||||
|
|
@ -5029,6 +5129,27 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
assert(a.splitter(0).equal(w));
|
||||
}
|
||||
|
||||
/// Adjacent separators and keeping sentinels.
|
||||
@safe unittest
|
||||
{
|
||||
import std.algorithm.comparison : equal;
|
||||
import std.typecons : Yes;
|
||||
|
||||
assert("|ab|".splitter!("a == b", Yes.keepSeparators)('|')
|
||||
.equal([ "", "|", "ab", "|", "" ]));
|
||||
assert("ab".splitter!("a == b", Yes.keepSeparators)('|')
|
||||
.equal([ "ab" ]));
|
||||
|
||||
assert("a|b||c".splitter!("a == b", Yes.keepSeparators)('|')
|
||||
.equal([ "a", "|", "b", "|", "", "|", "c" ]));
|
||||
assert("hello world".splitter!("a == b", Yes.keepSeparators)(' ')
|
||||
.equal([ "hello", " ", "", " ", "world" ]));
|
||||
|
||||
auto a = [ 1, 2, 0, 0, 3, 0, 4, 5, 0 ];
|
||||
auto w = [ [1, 2], [0], [], [0], [3], [0], [4, 5], [0], [] ];
|
||||
assert(a.splitter!("a == b", Yes.keepSeparators)(0).equal(w));
|
||||
}
|
||||
|
||||
/// Empty and separator-only ranges.
|
||||
@safe unittest
|
||||
{
|
||||
|
|
@ -5040,6 +5161,20 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
assert("||".splitter('|').equal([ "", "", "" ]));
|
||||
}
|
||||
|
||||
/// Empty and separator-only ranges and keeping sentinels.
|
||||
@safe unittest
|
||||
{
|
||||
import std.algorithm.comparison : equal;
|
||||
import std.typecons : Yes;
|
||||
import std.range : empty;
|
||||
|
||||
assert("".splitter!("a == b", Yes.keepSeparators)('|').empty);
|
||||
assert("|".splitter!("a == b", Yes.keepSeparators)('|')
|
||||
.equal([ "", "|", "" ]));
|
||||
assert("||".splitter!("a == b", Yes.keepSeparators)('|')
|
||||
.equal([ "", "|", "", "|", "" ]));
|
||||
}
|
||||
|
||||
/// Use a range for splitting
|
||||
@safe unittest
|
||||
{
|
||||
|
|
@ -5060,6 +5195,32 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
assert(a.splitter([0, 0]).equal([ [], [1] ]));
|
||||
}
|
||||
|
||||
/// Use a range for splitting
|
||||
@safe unittest
|
||||
{
|
||||
import std.algorithm.comparison : equal;
|
||||
import std.typecons : Yes;
|
||||
|
||||
assert("a=>bc=>def".splitter!("a == b", Yes.keepSeparators)("=>")
|
||||
.equal([ "a", "=>", "bc", "=>", "def" ]));
|
||||
assert("a|b||c".splitter!("a == b", Yes.keepSeparators)("||")
|
||||
.equal([ "a|b", "||", "c" ]));
|
||||
assert("hello world".splitter!("a == b", Yes.keepSeparators)(" ")
|
||||
.equal([ "hello", " ", "world" ]));
|
||||
|
||||
int[] a = [ 1, 2, 0, 0, 3, 0, 4, 5, 0 ];
|
||||
int[][] w = [ [1, 2], [0, 0], [3, 0, 4, 5, 0] ];
|
||||
assert(a.splitter!("a == b", Yes.keepSeparators)([0, 0]).equal(w));
|
||||
|
||||
a = [ 0, 0 ];
|
||||
assert(a.splitter!("a == b", Yes.keepSeparators)([0, 0])
|
||||
.equal([ (int[]).init, [0, 0], (int[]).init ]));
|
||||
|
||||
a = [ 0, 0, 1 ];
|
||||
assert(a.splitter!("a == b", Yes.keepSeparators)([0, 0])
|
||||
.equal([ [], [0, 0], [1] ]));
|
||||
}
|
||||
|
||||
/// Custom predicate functions.
|
||||
@safe unittest
|
||||
{
|
||||
|
|
@ -5073,6 +5234,21 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
assert(w.splitter!"a.front == b"(1).equal([ [[0]], [[2]] ]));
|
||||
}
|
||||
|
||||
/// Custom predicate functions.
|
||||
@safe unittest
|
||||
{
|
||||
import std.algorithm.comparison : equal;
|
||||
import std.typecons : Yes;
|
||||
import std.ascii : toLower;
|
||||
|
||||
assert("abXcdxef".splitter!("a.toLower == b", Yes.keepSeparators)('x')
|
||||
.equal([ "ab", "X", "cd", "x", "ef" ]));
|
||||
|
||||
auto w = [ [0], [1], [2] ];
|
||||
assert(w.splitter!("a.front == b", Yes.keepSeparators)(1)
|
||||
.equal([ [[0]], [[1]], [[2]] ]));
|
||||
}
|
||||
|
||||
/// Use splitter without a separator
|
||||
@safe unittest
|
||||
{
|
||||
|
|
@ -5105,6 +5281,18 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
assert("ab".splitter('|').equal([ "ab" ]));
|
||||
}
|
||||
|
||||
/// Leading separators, trailing separators, or no separators.
|
||||
@safe unittest
|
||||
{
|
||||
import std.algorithm.comparison : equal;
|
||||
import std.typecons : Yes;
|
||||
|
||||
assert("|ab|".splitter!("a == b", Yes.keepSeparators)('|')
|
||||
.equal([ "", "|", "ab", "|", "" ]));
|
||||
assert("ab".splitter!("a == b", Yes.keepSeparators)('|')
|
||||
.equal([ "ab" ]));
|
||||
}
|
||||
|
||||
/// Splitter returns bidirectional ranges if the delimiter is a single element
|
||||
@safe unittest
|
||||
{
|
||||
|
|
@ -5113,6 +5301,16 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
assert("a|bc|def".splitter('|').retro.equal([ "def", "bc", "a" ]));
|
||||
}
|
||||
|
||||
/// Splitter returns bidirectional ranges if the delimiter is a single element
|
||||
@safe unittest
|
||||
{
|
||||
import std.algorithm.comparison : equal;
|
||||
import std.typecons : Yes;
|
||||
import std.range : retro;
|
||||
assert("a|bc|def".splitter!("a == b", Yes.keepSeparators)('|')
|
||||
.retro.equal([ "def", "|", "bc", "|", "a" ]));
|
||||
}
|
||||
|
||||
/// Splitting by word lazily
|
||||
@safe unittest
|
||||
{
|
||||
|
|
@ -5221,7 +5419,10 @@ if (is(typeof(binaryFun!pred(r.front, s)) : bool)
|
|||
}
|
||||
|
||||
/// ditto
|
||||
auto splitter(alias pred = "a == b", Range, Separator)(Range r, Separator s)
|
||||
auto splitter(alias pred = "a == b",
|
||||
Flag!"keepSeparators" keepSeparators = No.keepSeparators,
|
||||
Range,
|
||||
Separator)(Range r, Separator s)
|
||||
if (is(typeof(binaryFun!pred(r.front, s.front)) : bool)
|
||||
&& (hasSlicing!Range || isNarrowString!Range)
|
||||
&& isForwardRange!Separator
|
||||
|
|
@ -5238,15 +5439,38 @@ if (is(typeof(binaryFun!pred(r.front, s.front)) : bool)
|
|||
// _frontLength == size_t.max means empty
|
||||
size_t _frontLength = size_t.max;
|
||||
|
||||
static if (keepSeparators)
|
||||
{
|
||||
bool _wasSeparator = true;
|
||||
}
|
||||
|
||||
@property auto separatorLength() { return _separator.length; }
|
||||
|
||||
void ensureFrontLength()
|
||||
{
|
||||
if (_frontLength != _frontLength.max) return;
|
||||
assert(!_input.empty, "The input must not be empty");
|
||||
// compute front length
|
||||
_frontLength = (_separator.empty) ? 1 :
|
||||
static if (keepSeparators)
|
||||
{
|
||||
assert(!_input.empty || _wasSeparator, "The input must not be empty");
|
||||
if (_wasSeparator)
|
||||
{
|
||||
_frontLength = _input.length -
|
||||
find!pred(_input, _separator).length;
|
||||
_wasSeparator = false;
|
||||
}
|
||||
else
|
||||
{
|
||||
_frontLength = separatorLength();
|
||||
_wasSeparator = true;
|
||||
}
|
||||
}
|
||||
else
|
||||
{
|
||||
assert(!_input.empty, "The input must not be empty");
|
||||
// compute front length
|
||||
_frontLength = (_separator.empty) ? 1 :
|
||||
_input.length - find!pred(_input, _separator).length;
|
||||
}
|
||||
}
|
||||
|
||||
public:
|
||||
|
|
@ -5271,7 +5495,14 @@ if (is(typeof(binaryFun!pred(r.front, s.front)) : bool)
|
|||
{
|
||||
@property bool empty()
|
||||
{
|
||||
return _frontLength == size_t.max && _input.empty;
|
||||
static if (keepSeparators)
|
||||
{
|
||||
return _frontLength == size_t.max && _input.empty && !_wasSeparator;
|
||||
}
|
||||
else
|
||||
{
|
||||
return _frontLength == size_t.max && _input.empty;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -5279,24 +5510,32 @@ if (is(typeof(binaryFun!pred(r.front, s.front)) : bool)
|
|||
{
|
||||
assert(!empty, "Attempting to popFront an empty splitter.");
|
||||
ensureFrontLength();
|
||||
if (_frontLength == _input.length)
|
||||
|
||||
static if (keepSeparators)
|
||||
{
|
||||
// done, there's no separator in sight
|
||||
_input = _input[_frontLength .. _frontLength];
|
||||
_frontLength = _frontLength.max;
|
||||
return;
|
||||
_input = _input[_frontLength .. _input.length];
|
||||
}
|
||||
if (_frontLength + separatorLength == _input.length)
|
||||
else
|
||||
{
|
||||
// Special case: popping the first-to-last item; there is
|
||||
// an empty item right after this.
|
||||
_input = _input[_input.length .. _input.length];
|
||||
_frontLength = 0;
|
||||
return;
|
||||
if (_frontLength == _input.length)
|
||||
{
|
||||
// done, there's no separator in sight
|
||||
_input = _input[_frontLength .. _frontLength];
|
||||
_frontLength = _frontLength.max;
|
||||
return;
|
||||
}
|
||||
if (_frontLength + separatorLength == _input.length)
|
||||
{
|
||||
// Special case: popping the first-to-last item; there is
|
||||
// an empty item right after this.
|
||||
_input = _input[_input.length .. _input.length];
|
||||
_frontLength = 0;
|
||||
return;
|
||||
}
|
||||
// Normal case, pop one item and the separator, get ready for
|
||||
// reading the next item
|
||||
_input = _input[_frontLength + separatorLength .. _input.length];
|
||||
}
|
||||
// Normal case, pop one item and the separator, get ready for
|
||||
// reading the next item
|
||||
_input = _input[_frontLength + separatorLength .. _input.length];
|
||||
// mark _frontLength as uninitialized
|
||||
_frontLength = _frontLength.max;
|
||||
}
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue