Source file src/pkg/strings/strings.go

     1	// Copyright 2009 The Go Authors. All rights reserved.
     2	// Use of this source code is governed by a BSD-style
     3	// license that can be found in the LICENSE file.
     4	
     5	// Package strings implements simple functions to manipulate strings.
     6	package strings
     7	
     8	import (
     9		"unicode"
    10		"unicode/utf8"
    11	)
    12	
    13	// explode splits s into an array of UTF-8 sequences, one per Unicode character (still strings) up to a maximum of n (n < 0 means no limit).
    14	// Invalid UTF-8 sequences become correct encodings of U+FFF8.
    15	func explode(s string, n int) []string {
    16		if n == 0 {
    17			return nil
    18		}
    19		l := utf8.RuneCountInString(s)
    20		if n <= 0 || n > l {
    21			n = l
    22		}
    23		a := make([]string, n)
    24		var size int
    25		var ch rune
    26		i, cur := 0, 0
    27		for ; i+1 < n; i++ {
    28			ch, size = utf8.DecodeRuneInString(s[cur:])
    29			if ch == utf8.RuneError {
    30				a[i] = string(utf8.RuneError)
    31			} else {
    32				a[i] = s[cur : cur+size]
    33			}
    34			cur += size
    35		}
    36		// add the rest, if there is any
    37		if cur < len(s) {
    38			a[i] = s[cur:]
    39		}
    40		return a
    41	}
    42	
    43	// primeRK is the prime base used in Rabin-Karp algorithm.
    44	const primeRK = 16777619
    45	
    46	// hashstr returns the hash and the appropriate multiplicative
    47	// factor for use in Rabin-Karp algorithm.
    48	func hashstr(sep string) (uint32, uint32) {
    49		hash := uint32(0)
    50		for i := 0; i < len(sep); i++ {
    51			hash = hash*primeRK + uint32(sep[i])
    52	
    53		}
    54		var pow, sq uint32 = 1, primeRK
    55		for i := len(sep); i > 0; i >>= 1 {
    56			if i&1 != 0 {
    57				pow *= sq
    58			}
    59			sq *= sq
    60		}
    61		return hash, pow
    62	}
    63	
    64	// Count counts the number of non-overlapping instances of sep in s.
    65	func Count(s, sep string) int {
    66		n := 0
    67		// special cases
    68		switch {
    69		case len(sep) == 0:
    70			return utf8.RuneCountInString(s) + 1
    71		case len(sep) == 1:
    72			// special case worth making fast
    73			c := sep[0]
    74			for i := 0; i < len(s); i++ {
    75				if s[i] == c {
    76					n++
    77				}
    78			}
    79			return n
    80		case len(sep) > len(s):
    81			return 0
    82		case len(sep) == len(s):
    83			if sep == s {
    84				return 1
    85			}
    86			return 0
    87		}
    88		hashsep, pow := hashstr(sep)
    89		h := uint32(0)
    90		for i := 0; i < len(sep); i++ {
    91			h = h*primeRK + uint32(s[i])
    92		}
    93		lastmatch := 0
    94		if h == hashsep && s[:len(sep)] == sep {
    95			n++
    96			lastmatch = len(sep)
    97		}
    98		for i := len(sep); i < len(s); {
    99			h *= primeRK
   100			h += uint32(s[i])
   101			h -= pow * uint32(s[i-len(sep)])
   102			i++
   103			if h == hashsep && lastmatch <= i-len(sep) && s[i-len(sep):i] == sep {
   104				n++
   105				lastmatch = i
   106			}
   107		}
   108		return n
   109	}
   110	
   111	// Contains returns true if substr is within s.
   112	func Contains(s, substr string) bool {
   113		return Index(s, substr) >= 0
   114	}
   115	
   116	// ContainsAny returns true if any Unicode code points in chars are within s.
   117	func ContainsAny(s, chars string) bool {
   118		return IndexAny(s, chars) >= 0
   119	}
   120	
   121	// ContainsRune returns true if the Unicode code point r is within s.
   122	func ContainsRune(s string, r rune) bool {
   123		return IndexRune(s, r) >= 0
   124	}
   125	
   126	// Index returns the index of the first instance of sep in s, or -1 if sep is not present in s.
   127	func Index(s, sep string) int {
   128		n := len(sep)
   129		switch {
   130		case n == 0:
   131			return 0
   132		case n == 1:
   133			c := sep[0]
   134			// special case worth making fast
   135			for i := 0; i < len(s); i++ {
   136				if s[i] == c {
   137					return i
   138				}
   139			}
   140			return -1
   141		case n == len(s):
   142			if sep == s {
   143				return 0
   144			}
   145			return -1
   146		case n > len(s):
   147			return -1
   148		}
   149		// Hash sep.
   150		hashsep, pow := hashstr(sep)
   151		var h uint32
   152		for i := 0; i < n; i++ {
   153			h = h*primeRK + uint32(s[i])
   154		}
   155		if h == hashsep && s[:n] == sep {
   156			return 0
   157		}
   158		for i := n; i < len(s); {
   159			h *= primeRK
   160			h += uint32(s[i])
   161			h -= pow * uint32(s[i-n])
   162			i++
   163			if h == hashsep && s[i-n:i] == sep {
   164				return i - n
   165			}
   166		}
   167		return -1
   168	}
   169	
   170	// LastIndex returns the index of the last instance of sep in s, or -1 if sep is not present in s.
   171	func LastIndex(s, sep string) int {
   172		n := len(sep)
   173		if n == 0 {
   174			return len(s)
   175		}
   176		c := sep[0]
   177		if n == 1 {
   178			// special case worth making fast
   179			for i := len(s) - 1; i >= 0; i-- {
   180				if s[i] == c {
   181					return i
   182				}
   183			}
   184			return -1
   185		}
   186		// n > 1
   187		for i := len(s) - n; i >= 0; i-- {
   188			if s[i] == c && s[i:i+n] == sep {
   189				return i
   190			}
   191		}
   192		return -1
   193	}
   194	
   195	// IndexRune returns the index of the first instance of the Unicode code point
   196	// r, or -1 if rune is not present in s.
   197	func IndexRune(s string, r rune) int {
   198		switch {
   199		case r < 0x80:
   200			b := byte(r)
   201			for i := 0; i < len(s); i++ {
   202				if s[i] == b {
   203					return i
   204				}
   205			}
   206		default:
   207			for i, c := range s {
   208				if c == r {
   209					return i
   210				}
   211			}
   212		}
   213		return -1
   214	}
   215	
   216	// IndexAny returns the index of the first instance of any Unicode code point
   217	// from chars in s, or -1 if no Unicode code point from chars is present in s.
   218	func IndexAny(s, chars string) int {
   219		if len(chars) > 0 {
   220			for i, c := range s {
   221				for _, m := range chars {
   222					if c == m {
   223						return i
   224					}
   225				}
   226			}
   227		}
   228		return -1
   229	}
   230	
   231	// LastIndexAny returns the index of the last instance of any Unicode code
   232	// point from chars in s, or -1 if no Unicode code point from chars is
   233	// present in s.
   234	func LastIndexAny(s, chars string) int {
   235		if len(chars) > 0 {
   236			for i := len(s); i > 0; {
   237				rune, size := utf8.DecodeLastRuneInString(s[0:i])
   238				i -= size
   239				for _, m := range chars {
   240					if rune == m {
   241						return i
   242					}
   243				}
   244			}
   245		}
   246		return -1
   247	}
   248	
   249	// Generic split: splits after each instance of sep,
   250	// including sepSave bytes of sep in the subarrays.
   251	func genSplit(s, sep string, sepSave, n int) []string {
   252		if n == 0 {
   253			return nil
   254		}
   255		if sep == "" {
   256			return explode(s, n)
   257		}
   258		if n < 0 {
   259			n = Count(s, sep) + 1
   260		}
   261		c := sep[0]
   262		start := 0
   263		a := make([]string, n)
   264		na := 0
   265		for i := 0; i+len(sep) <= len(s) && na+1 < n; i++ {
   266			if s[i] == c && (len(sep) == 1 || s[i:i+len(sep)] == sep) {
   267				a[na] = s[start : i+sepSave]
   268				na++
   269				start = i + len(sep)
   270				i += len(sep) - 1
   271			}
   272		}
   273		a[na] = s[start:]
   274		return a[0 : na+1]
   275	}
   276	
   277	// SplitN slices s into substrings separated by sep and returns a slice of
   278	// the substrings between those separators.
   279	// If sep is empty, SplitN splits after each UTF-8 sequence.
   280	// The count determines the number of substrings to return:
   281	//   n > 0: at most n substrings; the last substring will be the unsplit remainder.
   282	//   n == 0: the result is nil (zero substrings)
   283	//   n < 0: all substrings
   284	func SplitN(s, sep string, n int) []string { return genSplit(s, sep, 0, n) }
   285	
   286	// SplitAfterN slices s into substrings after each instance of sep and
   287	// returns a slice of those substrings.
   288	// If sep is empty, SplitAfterN splits after each UTF-8 sequence.
   289	// The count determines the number of substrings to return:
   290	//   n > 0: at most n substrings; the last substring will be the unsplit remainder.
   291	//   n == 0: the result is nil (zero substrings)
   292	//   n < 0: all substrings
   293	func SplitAfterN(s, sep string, n int) []string {
   294		return genSplit(s, sep, len(sep), n)
   295	}
   296	
   297	// Split slices s into all substrings separated by sep and returns a slice of
   298	// the substrings between those separators.
   299	// If sep is empty, Split splits after each UTF-8 sequence.
   300	// It is equivalent to SplitN with a count of -1.
   301	func Split(s, sep string) []string { return genSplit(s, sep, 0, -1) }
   302	
   303	// SplitAfter slices s into all substrings after each instance of sep and
   304	// returns a slice of those substrings.
   305	// If sep is empty, SplitAfter splits after each UTF-8 sequence.
   306	// It is equivalent to SplitAfterN with a count of -1.
   307	func SplitAfter(s, sep string) []string {
   308		return genSplit(s, sep, len(sep), -1)
   309	}
   310	
   311	// Fields splits the string s around each instance of one or more consecutive white space
   312	// characters, as defined by unicode.IsSpace, returning an array of substrings of s or an
   313	// empty list if s contains only white space.
   314	func Fields(s string) []string {
   315		return FieldsFunc(s, unicode.IsSpace)
   316	}
   317	
   318	// FieldsFunc splits the string s at each run of Unicode code points c satisfying f(c)
   319	// and returns an array of slices of s. If all code points in s satisfy f(c) or the
   320	// string is empty, an empty slice is returned.
   321	func FieldsFunc(s string, f func(rune) bool) []string {
   322		// First count the fields.
   323		n := 0
   324		inField := false
   325		for _, rune := range s {
   326			wasInField := inField
   327			inField = !f(rune)
   328			if inField && !wasInField {
   329				n++
   330			}
   331		}
   332	
   333		// Now create them.
   334		a := make([]string, n)
   335		na := 0
   336		fieldStart := -1 // Set to -1 when looking for start of field.
   337		for i, rune := range s {
   338			if f(rune) {
   339				if fieldStart >= 0 {
   340					a[na] = s[fieldStart:i]
   341					na++
   342					fieldStart = -1
   343				}
   344			} else if fieldStart == -1 {
   345				fieldStart = i
   346			}
   347		}
   348		if fieldStart >= 0 { // Last field might end at EOF.
   349			a[na] = s[fieldStart:]
   350		}
   351		return a
   352	}
   353	
   354	// Join concatenates the elements of a to create a single string.   The separator string
   355	// sep is placed between elements in the resulting string.
   356	func Join(a []string, sep string) string {
   357		if len(a) == 0 {
   358			return ""
   359		}
   360		if len(a) == 1 {
   361			return a[0]
   362		}
   363		n := len(sep) * (len(a) - 1)
   364		for i := 0; i < len(a); i++ {
   365			n += len(a[i])
   366		}
   367	
   368		b := make([]byte, n)
   369		bp := copy(b, a[0])
   370		for _, s := range a[1:] {
   371			bp += copy(b[bp:], sep)
   372			bp += copy(b[bp:], s)
   373		}
   374		return string(b)
   375	}
   376	
   377	// HasPrefix tests whether the string s begins with prefix.
   378	func HasPrefix(s, prefix string) bool {
   379		return len(s) >= len(prefix) && s[0:len(prefix)] == prefix
   380	}
   381	
   382	// HasSuffix tests whether the string s ends with suffix.
   383	func HasSuffix(s, suffix string) bool {
   384		return len(s) >= len(suffix) && s[len(s)-len(suffix):] == suffix
   385	}
   386	
   387	// Map returns a copy of the string s with all its characters modified
   388	// according to the mapping function. If mapping returns a negative value, the character is
   389	// dropped from the string with no replacement.
   390	func Map(mapping func(rune) rune, s string) string {
   391		// In the worst case, the string can grow when mapped, making
   392		// things unpleasant.  But it's so rare we barge in assuming it's
   393		// fine.  It could also shrink but that falls out naturally.
   394		maxbytes := len(s) // length of b
   395		nbytes := 0        // number of bytes encoded in b
   396		// The output buffer b is initialized on demand, the first
   397		// time a character differs.
   398		var b []byte
   399	
   400		for i, c := range s {
   401			r := mapping(c)
   402			if b == nil {
   403				if r == c {
   404					continue
   405				}
   406				b = make([]byte, maxbytes)
   407				nbytes = copy(b, s[:i])
   408			}
   409			if r >= 0 {
   410				wid := 1
   411				if r >= utf8.RuneSelf {
   412					wid = utf8.RuneLen(r)
   413				}
   414				if nbytes+wid > maxbytes {
   415					// Grow the buffer.
   416					maxbytes = maxbytes*2 + utf8.UTFMax
   417					nb := make([]byte, maxbytes)
   418					copy(nb, b[0:nbytes])
   419					b = nb
   420				}
   421				nbytes += utf8.EncodeRune(b[nbytes:maxbytes], r)
   422			}
   423		}
   424		if b == nil {
   425			return s
   426		}
   427		return string(b[0:nbytes])
   428	}
   429	
   430	// Repeat returns a new string consisting of count copies of the string s.
   431	func Repeat(s string, count int) string {
   432		b := make([]byte, len(s)*count)
   433		bp := 0
   434		for i := 0; i < count; i++ {
   435			for j := 0; j < len(s); j++ {
   436				b[bp] = s[j]
   437				bp++
   438			}
   439		}
   440		return string(b)
   441	}
   442	
   443	// ToUpper returns a copy of the string s with all Unicode letters mapped to their upper case.
   444	func ToUpper(s string) string { return Map(unicode.ToUpper, s) }
   445	
   446	// ToLower returns a copy of the string s with all Unicode letters mapped to their lower case.
   447	func ToLower(s string) string { return Map(unicode.ToLower, s) }
   448	
   449	// ToTitle returns a copy of the string s with all Unicode letters mapped to their title case.
   450	func ToTitle(s string) string { return Map(unicode.ToTitle, s) }
   451	
   452	// ToUpperSpecial returns a copy of the string s with all Unicode letters mapped to their
   453	// upper case, giving priority to the special casing rules.
   454	func ToUpperSpecial(_case unicode.SpecialCase, s string) string {
   455		return Map(func(r rune) rune { return _case.ToUpper(r) }, s)
   456	}
   457	
   458	// ToLowerSpecial returns a copy of the string s with all Unicode letters mapped to their
   459	// lower case, giving priority to the special casing rules.
   460	func ToLowerSpecial(_case unicode.SpecialCase, s string) string {
   461		return Map(func(r rune) rune { return _case.ToLower(r) }, s)
   462	}
   463	
   464	// ToTitleSpecial returns a copy of the string s with all Unicode letters mapped to their
   465	// title case, giving priority to the special casing rules.
   466	func ToTitleSpecial(_case unicode.SpecialCase, s string) string {
   467		return Map(func(r rune) rune { return _case.ToTitle(r) }, s)
   468	}
   469	
   470	// isSeparator reports whether the rune could mark a word boundary.
   471	// TODO: update when package unicode captures more of the properties.
   472	func isSeparator(r rune) bool {
   473		// ASCII alphanumerics and underscore are not separators
   474		if r <= 0x7F {
   475			switch {
   476			case '0' <= r && r <= '9':
   477				return false
   478			case 'a' <= r && r <= 'z':
   479				return false
   480			case 'A' <= r && r <= 'Z':
   481				return false
   482			case r == '_':
   483				return false
   484			}
   485			return true
   486		}
   487		// Letters and digits are not separators
   488		if unicode.IsLetter(r) || unicode.IsDigit(r) {
   489			return false
   490		}
   491		// Otherwise, all we can do for now is treat spaces as separators.
   492		return unicode.IsSpace(r)
   493	}
   494	
   495	// Title returns a copy of the string s with all Unicode letters that begin words
   496	// mapped to their title case.
   497	//
   498	// BUG: The rule Title uses for word boundaries does not handle Unicode punctuation properly.
   499	func Title(s string) string {
   500		// Use a closure here to remember state.
   501		// Hackish but effective. Depends on Map scanning in order and calling
   502		// the closure once per rune.
   503		prev := ' '
   504		return Map(
   505			func(r rune) rune {
   506				if isSeparator(prev) {
   507					prev = r
   508					return unicode.ToTitle(r)
   509				}
   510				prev = r
   511				return r
   512			},
   513			s)
   514	}
   515	
   516	// TrimLeftFunc returns a slice of the string s with all leading
   517	// Unicode code points c satisfying f(c) removed.
   518	func TrimLeftFunc(s string, f func(rune) bool) string {
   519		i := indexFunc(s, f, false)
   520		if i == -1 {
   521			return ""
   522		}
   523		return s[i:]
   524	}
   525	
   526	// TrimRightFunc returns a slice of the string s with all trailing
   527	// Unicode code points c satisfying f(c) removed.
   528	func TrimRightFunc(s string, f func(rune) bool) string {
   529		i := lastIndexFunc(s, f, false)
   530		if i >= 0 && s[i] >= utf8.RuneSelf {
   531			_, wid := utf8.DecodeRuneInString(s[i:])
   532			i += wid
   533		} else {
   534			i++
   535		}
   536		return s[0:i]
   537	}
   538	
   539	// TrimFunc returns a slice of the string s with all leading
   540	// and trailing Unicode code points c satisfying f(c) removed.
   541	func TrimFunc(s string, f func(rune) bool) string {
   542		return TrimRightFunc(TrimLeftFunc(s, f), f)
   543	}
   544	
   545	// IndexFunc returns the index into s of the first Unicode
   546	// code point satisfying f(c), or -1 if none do.
   547	func IndexFunc(s string, f func(rune) bool) int {
   548		return indexFunc(s, f, true)
   549	}
   550	
   551	// LastIndexFunc returns the index into s of the last
   552	// Unicode code point satisfying f(c), or -1 if none do.
   553	func LastIndexFunc(s string, f func(rune) bool) int {
   554		return lastIndexFunc(s, f, true)
   555	}
   556	
   557	// indexFunc is the same as IndexFunc except that if
   558	// truth==false, the sense of the predicate function is
   559	// inverted.
   560	func indexFunc(s string, f func(rune) bool, truth bool) int {
   561		start := 0
   562		for start < len(s) {
   563			wid := 1
   564			r := rune(s[start])
   565			if r >= utf8.RuneSelf {
   566				r, wid = utf8.DecodeRuneInString(s[start:])
   567			}
   568			if f(r) == truth {
   569				return start
   570			}
   571			start += wid
   572		}
   573		return -1
   574	}
   575	
   576	// lastIndexFunc is the same as LastIndexFunc except that if
   577	// truth==false, the sense of the predicate function is
   578	// inverted.
   579	func lastIndexFunc(s string, f func(rune) bool, truth bool) int {
   580		for i := len(s); i > 0; {
   581			r, size := utf8.DecodeLastRuneInString(s[0:i])
   582			i -= size
   583			if f(r) == truth {
   584				return i
   585			}
   586		}
   587		return -1
   588	}
   589	
   590	func makeCutsetFunc(cutset string) func(rune) bool {
   591		return func(r rune) bool { return IndexRune(cutset, r) >= 0 }
   592	}
   593	
   594	// Trim returns a slice of the string s with all leading and
   595	// trailing Unicode code points contained in cutset removed.
   596	func Trim(s string, cutset string) string {
   597		if s == "" || cutset == "" {
   598			return s
   599		}
   600		return TrimFunc(s, makeCutsetFunc(cutset))
   601	}
   602	
   603	// TrimLeft returns a slice of the string s with all leading
   604	// Unicode code points contained in cutset removed.
   605	func TrimLeft(s string, cutset string) string {
   606		if s == "" || cutset == "" {
   607			return s
   608		}
   609		return TrimLeftFunc(s, makeCutsetFunc(cutset))
   610	}
   611	
   612	// TrimRight returns a slice of the string s, with all trailing
   613	// Unicode code points contained in cutset removed.
   614	func TrimRight(s string, cutset string) string {
   615		if s == "" || cutset == "" {
   616			return s
   617		}
   618		return TrimRightFunc(s, makeCutsetFunc(cutset))
   619	}
   620	
   621	// TrimSpace returns a slice of the string s, with all leading
   622	// and trailing white space removed, as defined by Unicode.
   623	func TrimSpace(s string) string {
   624		return TrimFunc(s, unicode.IsSpace)
   625	}
   626	
   627	// TrimPrefix returns s without the provided leading prefix string.
   628	// If s doesn't start with prefix, s is returned unchanged.
   629	func TrimPrefix(s, prefix string) string {
   630		if HasPrefix(s, prefix) {
   631			return s[len(prefix):]
   632		}
   633		return s
   634	}
   635	
   636	// TrimSuffix returns s without the provided trailing suffix string.
   637	// If s doesn't end with suffix, s is returned unchanged.
   638	func TrimSuffix(s, suffix string) string {
   639		if HasSuffix(s, suffix) {
   640			return s[:len(s)-len(suffix)]
   641		}
   642		return s
   643	}
   644	
   645	// Replace returns a copy of the string s with the first n
   646	// non-overlapping instances of old replaced by new.
   647	// If n < 0, there is no limit on the number of replacements.
   648	func Replace(s, old, new string, n int) string {
   649		if old == new || n == 0 {
   650			return s // avoid allocation
   651		}
   652	
   653		// Compute number of replacements.
   654		if m := Count(s, old); m == 0 {
   655			return s // avoid allocation
   656		} else if n < 0 || m < n {
   657			n = m
   658		}
   659	
   660		// Apply replacements to buffer.
   661		t := make([]byte, len(s)+n*(len(new)-len(old)))
   662		w := 0
   663		start := 0
   664		for i := 0; i < n; i++ {
   665			j := start
   666			if len(old) == 0 {
   667				if i > 0 {
   668					_, wid := utf8.DecodeRuneInString(s[start:])
   669					j += wid
   670				}
   671			} else {
   672				j += Index(s[start:], old)
   673			}
   674			w += copy(t[w:], s[start:j])
   675			w += copy(t[w:], new)
   676			start = j + len(old)
   677		}
   678		w += copy(t[w:], s[start:])
   679		return string(t[0:w])
   680	}
   681	
   682	// EqualFold reports whether s and t, interpreted as UTF-8 strings,
   683	// are equal under Unicode case-folding.
   684	func EqualFold(s, t string) bool {
   685		for s != "" && t != "" {
   686			// Extract first rune from each string.
   687			var sr, tr rune
   688			if s[0] < utf8.RuneSelf {
   689				sr, s = rune(s[0]), s[1:]
   690			} else {
   691				r, size := utf8.DecodeRuneInString(s)
   692				sr, s = r, s[size:]
   693			}
   694			if t[0] < utf8.RuneSelf {
   695				tr, t = rune(t[0]), t[1:]
   696			} else {
   697				r, size := utf8.DecodeRuneInString(t)
   698				tr, t = r, t[size:]
   699			}
   700	
   701			// If they match, keep going; if not, return false.
   702	
   703			// Easy case.
   704			if tr == sr {
   705				continue
   706			}
   707	
   708			// Make sr < tr to simplify what follows.
   709			if tr < sr {
   710				tr, sr = sr, tr
   711			}
   712			// Fast check for ASCII.
   713			if tr < utf8.RuneSelf && 'A' <= sr && sr <= 'Z' {
   714				// ASCII, and sr is upper case.  tr must be lower case.
   715				if tr == sr+'a'-'A' {
   716					continue
   717				}
   718				return false
   719			}
   720	
   721			// General case.  SimpleFold(x) returns the next equivalent rune > x
   722			// or wraps around to smaller values.
   723			r := unicode.SimpleFold(sr)
   724			for r != sr && r < tr {
   725				r = unicode.SimpleFold(r)
   726			}
   727			if r == tr {
   728				continue
   729			}
   730			return false
   731		}
   732	
   733		// One string is empty.  Are both?
   734		return s == t
   735	}

View as plain text