package norm

import "golang.org/x/text/unicode/norm"

Package norm contains types and functions for normalizing Unicode strings.

Index ¶

• Constants
• type Form
• func (f Form) Append(out []byte, src ...byte) []byte
• func (f Form) AppendString(out []byte, src string) []byte
• func (f Form) Bytes(b []byte) []byte
• func (f Form) FirstBoundary(b []byte) int
• func (f Form) FirstBoundaryInString(s string) int
• func (f Form) IsNormal(b []byte) bool
• func (f Form) IsNormalString(s string) bool
• func (f Form) LastBoundary(b []byte) int
• func (f Form) NextBoundary(b []byte, atEOF bool) int
• func (f Form) NextBoundaryInString(s string, atEOF bool) int
• func (f Form) Properties(s []byte) Properties
• func (f Form) PropertiesString(s string) Properties
• func (f Form) QuickSpan(b []byte) int
• func (f Form) QuickSpanString(s string) int
• func (f Form) Reader(r io.Reader) io.Reader
• func (Form) Reset()
• func (f Form) Span(b []byte, atEOF bool) (n int, err error)
• func (f Form) SpanString(s string, atEOF bool) (n int, err error)
• func (f Form) String(s string) string
• func (f Form) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error)
• func (f Form) Writer(w io.Writer) io.WriteCloser
• type Iter
• func (i *Iter) Done() bool
• func (i *Iter) Init(f Form, src []byte)
• func (i *Iter) InitString(f Form, src string)
• func (i *Iter) Next() []byte
• func (i *Iter) Pos() int
• func (i *Iter) Seek(offset int64, whence int) (int64, error)
• type Properties
• func (p Properties) BoundaryAfter() bool
• func (p Properties) BoundaryBefore() bool
• func (p Properties) CCC() uint8
• func (p Properties) Decomposition() []byte
• func (p Properties) LeadCCC() uint8
• func (p Properties) Size() int
• func (p Properties) TrailCCC() uint8

Examples ¶

• Form.NextBoundary
• Iter

Constants ¶

const (
	// Version is the Unicode edition from which the tables are derived.
	Version = "15.0.0"

	// MaxTransformChunkSize indicates the maximum number of bytes that Transform
	// may need to write atomically for any Form. Making a destination buffer at
	// least this size ensures that Transform can always make progress and that
	// the user does not need to grow the buffer on an ErrShortDst.
	MaxTransformChunkSize = 35 + maxNonStarters*4
)

const GraphemeJoiner = "\u034F"

GraphemeJoiner is inserted after maxNonStarters non-starter runes.

const MaxSegmentSize = maxByteBufferSize

MaxSegmentSize is the maximum size of a byte buffer needed to consider any sequence of starter and non-starter runes for the purpose of normalization.

Types ¶

type Form ¶

type Form int

A Form denotes a canonical representation of Unicode code points. The Unicode-defined normalization and equivalence forms are:

NFC   Unicode Normalization Form C
NFD   Unicode Normalization Form D
NFKC  Unicode Normalization Form KC
NFKD  Unicode Normalization Form KD

For a Form f, this documentation uses the notation f(x) to mean the bytes or string x converted to the given form. A position n in x is called a boundary if conversion to the form can proceed independently on both sides:

f(x) == append(f(x[0:n]), f(x[n:])...)

References: https://unicode.org/reports/tr15/ and https://unicode.org/notes/tn5/.

const (
	NFC Form = iota
	NFD
	NFKC
	NFKD
)

func (Form) Append ¶

func (f Form) Append(out []byte, src ...byte) []byte

Append returns f(append(out, b...)). The buffer out must be nil, empty, or equal to f(out).

func (Form) AppendString ¶

func (f Form) AppendString(out []byte, src string) []byte

AppendString returns f(append(out, []byte(s))). The buffer out must be nil, empty, or equal to f(out).

func (Form) Bytes ¶

func (f Form) Bytes(b []byte) []byte

Bytes returns f(b). May return b if f(b) = b.

func (Form) FirstBoundary ¶

func (f Form) FirstBoundary(b []byte) int

FirstBoundary returns the position i of the first boundary in b or -1 if b contains no boundary.

func (Form) FirstBoundaryInString ¶

func (f Form) FirstBoundaryInString(s string) int

FirstBoundaryInString returns the position i of the first boundary in s or -1 if s contains no boundary.

func (Form) IsNormal ¶

func (f Form) IsNormal(b []byte) bool

IsNormal returns true if b == f(b).

func (Form) IsNormalString ¶

func (f Form) IsNormalString(s string) bool

IsNormalString returns true if s == f(s).

func (Form) LastBoundary ¶

func (f Form) LastBoundary(b []byte) int

LastBoundary returns the position i of the last boundary in b or -1 if b contains no boundary.

func (Form) NextBoundary ¶

func (f Form) NextBoundary(b []byte, atEOF bool) int

NextBoundary reports the index of the boundary between the first and next segment in b or -1 if atEOF is false and there are not enough bytes to determine this boundary.

package main

import (
	"fmt"

	"golang.org/x/text/unicode/norm"
)

func main() {
	s := norm.NFD.String("Mêlée")

	for i := 0; i < len(s); {
		d := norm.NFC.NextBoundaryInString(s[i:], true)
		fmt.Printf("%[1]s: %+[1]q\n", s[i:i+d])
		i += d
	}
}

M: "M"
ê: "e\u0302"
l: "l"
é: "e\u0301"
e: "e"

func (Form) NextBoundaryInString ¶

func (f Form) NextBoundaryInString(s string, atEOF bool) int

NextBoundaryInString reports the index of the boundary between the first and next segment in b or -1 if atEOF is false and there are not enough bytes to determine this boundary.

func (Form) Properties ¶

func (f Form) Properties(s []byte) Properties

Properties returns properties for the first rune in s.

func (Form) PropertiesString ¶

func (f Form) PropertiesString(s string) Properties

PropertiesString returns properties for the first rune in s.

func (Form) QuickSpan ¶

func (f Form) QuickSpan(b []byte) int

QuickSpan returns a boundary n such that b[0:n] == f(b[0:n]). It is not guaranteed to return the largest such n.

func (Form) QuickSpanString ¶

func (f Form) QuickSpanString(s string) int

QuickSpanString returns a boundary n such that s[0:n] == f(s[0:n]). It is not guaranteed to return the largest such n.

func (Form) Reader ¶

func (f Form) Reader(r io.Reader) io.Reader

Reader returns a new reader that implements Read by reading data from r and returning f(data).

func (Form) Reset ¶

func (Form) Reset()

Reset implements the Reset method of the transform.Transformer interface.

func (Form) Span ¶

func (f Form) Span(b []byte, atEOF bool) (n int, err error)

Span implements transform.SpanningTransformer. It returns a boundary n such that b[0:n] == f(b[0:n]). It is not guaranteed to return the largest such n.

func (Form) SpanString ¶

func (f Form) SpanString(s string, atEOF bool) (n int, err error)

SpanString returns a boundary n such that s[0:n] == f(s[0:n]). It is not guaranteed to return the largest such n.

func (Form) String ¶

func (f Form) String(s string) string

String returns f(s).

func (Form) Transform ¶

func (f Form) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error)

Transform implements the Transform method of the transform.Transformer interface. It may need to write segments of up to MaxSegmentSize at once. Users should either catch ErrShortDst and allow dst to grow or have dst be at least of size MaxTransformChunkSize to be guaranteed of progress.

func (Form) Writer ¶

func (f Form) Writer(w io.Writer) io.WriteCloser

Writer returns a new writer that implements Write(b) by writing f(b) to w. The returned writer may use an internal buffer to maintain state across Write calls. Calling its Close method writes any buffered data to w.

type Iter ¶

type Iter struct {
	// contains filtered or unexported fields
}

An Iter iterates over a string or byte slice, while normalizing it to a given Form.

package main

import (
	"bytes"
	"fmt"
	"io"
	"unicode/utf8"

	"golang.org/x/text/unicode/norm"
)

// EqualSimple uses a norm.Iter to compare two non-normalized
// strings for equivalence.
func EqualSimple(a, b string) bool {
	var ia, ib norm.Iter
	ia.InitString(norm.NFKD, a)
	ib.InitString(norm.NFKD, b)
	for !ia.Done() && !ib.Done() {
		if !bytes.Equal(ia.Next(), ib.Next()) {
			return false
		}
	}
	return ia.Done() && ib.Done()
}

// FindPrefix finds the longest common prefix of ASCII characters
// of a and b.
func FindPrefix(a, b string) int {
	i := 0
	for ; i < len(a) && i < len(b) && a[i] < utf8.RuneSelf && a[i] == b[i]; i++ {
	}
	return i
}

// EqualOpt is like EqualSimple, but optimizes the special
// case for ASCII characters.
func EqualOpt(a, b string) bool {
	n := FindPrefix(a, b)
	a, b = a[n:], b[n:]
	var ia, ib norm.Iter
	ia.InitString(norm.NFKD, a)
	ib.InitString(norm.NFKD, b)
	for !ia.Done() && !ib.Done() {
		if !bytes.Equal(ia.Next(), ib.Next()) {
			return false
		}
		if n := int64(FindPrefix(a[ia.Pos():], b[ib.Pos():])); n != 0 {
			ia.Seek(n, io.SeekCurrent)
			ib.Seek(n, io.SeekCurrent)
		}
	}
	return ia.Done() && ib.Done()
}

var compareTests = []struct{ a, b string }{
	{"aaa", "aaa"},
	{"aaa", "aab"},
	{"a\u0300a", "\u00E0a"},
	{"a\u0300\u0320b", "a\u0320\u0300b"},
	{"\u1E0A\u0323", "\x44\u0323\u0307"},
	// A character that decomposes into multiple segments
	// spans several iterations.
	{"\u3304", "\u30A4\u30CB\u30F3\u30AF\u3099"},
}

func main() {
	for i, t := range compareTests {
		r0 := EqualSimple(t.a, t.b)
		r1 := EqualOpt(t.a, t.b)
		fmt.Printf("%d: %v %v\n", i, r0, r1)
	}
}

0: true true
1: false false
2: true true
3: true true
4: true true
5: true true

func (*Iter) Done ¶

func (i *Iter) Done() bool

Done returns true if there is no more input to process.

func (*Iter) Init ¶

func (i *Iter) Init(f Form, src []byte)

Init initializes i to iterate over src after normalizing it to Form f.

func (*Iter) InitString ¶

func (i *Iter) InitString(f Form, src string)

InitString initializes i to iterate over src after normalizing it to Form f.

func (*Iter) Next ¶

func (i *Iter) Next() []byte

Next returns f(i.input[i.Pos():n]), where n is a boundary of i.input. For any input a and b for which f(a) == f(b), subsequent calls to Next will return the same segments. Modifying runes are grouped together with the preceding starter, if such a starter exists. Although not guaranteed, n will typically be the smallest possible n.

func (*Iter) Pos ¶

func (i *Iter) Pos() int

Pos returns the byte position at which the next call to Next will commence processing.

func (*Iter) Seek ¶

func (i *Iter) Seek(offset int64, whence int) (int64, error)

Seek sets the segment to be returned by the next call to Next to start at position p. It is the responsibility of the caller to set p to the start of a segment.

type Properties ¶

type Properties struct {
	// contains filtered or unexported fields
}

Properties provides access to normalization properties of a rune.

func (Properties) BoundaryAfter ¶

func (p Properties) BoundaryAfter() bool

BoundaryAfter returns true if runes cannot combine with or otherwise interact with this or previous runes.

func (Properties) BoundaryBefore ¶

func (p Properties) BoundaryBefore() bool

BoundaryBefore returns true if this rune starts a new segment and cannot combine with any rune on the left.

func (Properties) CCC ¶

func (p Properties) CCC() uint8

CCC returns the canonical combining class of the underlying rune.

func (Properties) Decomposition ¶

func (p Properties) Decomposition() []byte

Decomposition returns the decomposition for the underlying rune or nil if there is none.

func (Properties) LeadCCC ¶

func (p Properties) LeadCCC() uint8

LeadCCC returns the CCC of the first rune in the decomposition. If there is no decomposition, LeadCCC equals CCC.

func (Properties) Size ¶

func (p Properties) Size() int

Size returns the length of UTF-8 encoding of the rune.

func (Properties) TrailCCC ¶

func (p Properties) TrailCCC() uint8

TrailCCC returns the CCC of the last rune in the decomposition. If there is no decomposition, TrailCCC equals CCC.

Source Files ¶

composition.go forminfo.go input.go iter.go normalize.go readwriter.go tables15.0.0.go transform.go trie.go