X-Git-Url: http://nitlanguage.org diff --git a/lib/standard/string_search.nit b/lib/standard/string_search.nit index 846de56..3a885ad 100644 --- a/lib/standard/string_search.nit +++ b/lib/standard/string_search.nit @@ -28,7 +28,7 @@ interface Pattern # This method is usually faster than `search_in` if what is # required is only the index. # Note: in most implementation, `search_in` is implemented with this method. - protected fun search_index_in(s: String, from: Int): Int is abstract + protected fun search_index_in(s: Text, from: Int): Int is abstract # Search `self` into `s` from a certain position. # Return null if not found. @@ -40,7 +40,7 @@ interface Pattern # If only the index of the first character if required, see `search_index_in`. # # Note: Is used by `String::search`, `String::search_from`, and others. - protected fun search_in(s: String, from: Int): nullable Match is abstract + protected fun search_in(s: Text, from: Int): nullable Match is abstract # Search all `self` occurrences into `s`. # @@ -48,7 +48,7 @@ interface Pattern # assert 'z'.search_all_in("hello world").length == 0 # # Note: Is used by `String::search_all`. - protected fun search_all_in(s: String): Array[Match] + protected fun search_all_in(s: Text): Array[Match] do var res = new Array[Match] # Result var match = search_in(s, 0) @@ -68,7 +68,7 @@ interface Pattern # assert 'z'.split_in("hello world").join("|") == "hello world" # # Note: is used by `String::split` - protected fun split_in(s: String): Array[Match] + protected fun split_in(s: Text): Array[Match] do var res = new Array[Match] # Result var i = 0 # Cursor @@ -76,14 +76,17 @@ interface Pattern while match != null do # Compute the splited part length var len = match.from - i - res.add(new Match(s, i, len)) + res.add(new Match(s.to_s, i, len)) i = match.after match = search_in(s, i) end # Add the last part - res.add(new Match(s, i, s.length - i)) + res.add(new Match(s.to_s, i, s.length - i)) return res end + + # Is `self` in `s`? + protected fun is_in(s: Text): Bool do return search_index_in(s, 0) != -1 end # BM_Pattern are pre-compiled string motif for the Boyer-Moore algorithm. @@ -133,26 +136,23 @@ class BM_Pattern if to < 0 then return null else - return new Match(s, to, _length) + return new Match(s.to_s, to, _length) end end - # Compile a new motif - init(motif: String) + init do - _motif = motif - _length = motif.length + _length = _motif.length _gs = new Array[Int].with_capacity(_length) - _bc_table = new ArrayMap[Char, Int] compute_gs compute_bc end # searched motif - var _motif: String + private var motif: String # length of the motif - var _length: Int + private var length: Int is noinit private fun bc(e: Char): Int do @@ -164,10 +164,10 @@ class BM_Pattern end # good shifts - var _gs: Array[Int] + private var gs: Array[Int] is noinit # bad characters - var _bc_table: Map[Char, Int] + private var bc_table = new ArrayMap[Char, Int] private fun compute_bc do @@ -205,7 +205,6 @@ class BM_Pattern private fun compute_gs do - var x = _motif var m = _length var suff = suffixes var i = 0 @@ -235,33 +234,29 @@ class BM_Pattern redef fun ==(o) do return o isa BM_Pattern and o._motif == _motif end -# Matches are a part of a `String` found ba a `Pattern`. +# Matches are a part of a `Text` found by a `Pattern`. class Match # The base string matched - readable var _string: String + var string: String # The starting position in the string - readable var _from: Int + var from: Int # The length of the matching part - readable var _length: Int + var length: Int # The position of the first character just after the matching part. # May be out of the base string - fun after: Int do return _from + _length + fun after: Int do return from + length # The contents of the matching part - redef fun to_s do return _string.substring(_from, _length) + redef fun to_s do return string.substring(from,length) - # Matches `len` characters of `s` from `f`. - init(s: String, f: Int, len: Int) + init do - assert positive_length: len >= 0 - assert valid_from: f >= 0 - assert valid_after: f + len <= s.length - _string = s - _from = f - _length = len + assert positive_length: length >= 0 + assert valid_from: from >= 0 + assert valid_after: from + length <= string.length end end @@ -284,12 +279,12 @@ redef class Char if pos < 0 then return null else - return new Match(s, pos, 1) + return new Match(s.to_s, pos, 1) end end end -redef class String +redef class Text super Pattern redef fun search_index_in(s, from) @@ -313,7 +308,7 @@ redef class String if pos < 0 then return null else - return new Match(s, pos, length) + return new Match(s.to_s, pos, length) end end @@ -332,6 +327,37 @@ redef class String # assert "I say hello to the world!".search_from("hello",7) == null fun search_from(p: Pattern, from: Int): nullable Match do return p.search_in(self, from) + # Search the last occurence of the text `t`. + # + # assert "bob".search_last("b").from == 2 + # assert "bob".search_last("bo").from == 0 + # assert "bob".search_last("ob").from == 1 + # assert "bobob".search_last("ob").from == 3 + # assert "bobbob".search_last("bb").from == 2 + # assert "bobbob".search_last("bob").from == 3 + # assert "bob".search_last("z") == null + # assert "".search_last("b") == null + fun search_last(t: Text): nullable Match do + return search_last_up_to(t, length) + end + + # Search the last occurence of the text `t` before `up_to`. + # + # assert "bobbob".search_last_up_to("b", 3).from == 2 + # assert "bobbob".search_last_up_to("b", 6).from == 5 + # assert "bobbob".search_last_up_to("b", 0) == null + fun search_last_up_to(t: Text, up_to: Int): nullable Match do + var i = up_to - t.length + + while i >= 0 do + if substring(i, t.length) == t then + return new Match(self.to_s, i, t.length) + end + i -= 1 + end + return null + end + # Search all occurrences of p into self. # # var a = new Array[Int] @@ -355,25 +381,30 @@ redef class String # @deprecated alias for `split` fun split_with(p: Pattern): Array[String] do return self.split(p) + # Split `self` on the first `=` + # + # assert "hello".split_once_on('l') == ["he", "lo"] + # assert "a, b, c, d, e".split_once_on(", ") == ["a", "b, c, d, e"] + fun split_once_on(p: Pattern): Array[SELFTYPE] + do + var m = p.search_in(self, 0) + if m == null then return [self] + return new Array[SELFTYPE].with_items(substring(0, m.from), substring_from(m.after)) + end + # Replace all occurences of a pattern with a string # # assert "hlelo".replace("le", "el") == "hello" # assert "hello".replace('l', "") == "heo" - fun replace(p: Pattern, string: String): String + fun replace(p: Pattern, string: SELFTYPE): String do return self.split_with(p).join(string) end - # Escape the four characters `<`, `>`, `&`, and `"` with their html counterpart + # Does `self` contains at least one instance of `pattern`? # - # assert "a&b->\"x\"".html_escape == "a&b->"x"" - fun html_escape: String - do - var ret = self - if ret.chars.has('&') then ret = ret.replace('&', "&") - if ret.chars.has('<') then ret = ret.replace('<', "<") - if ret.chars.has('>') then ret = ret.replace('>', ">") - if ret.chars.has('"') then ret = ret.replace('"', """) - return ret - end + # assert "hello".has('l') + # assert "hello".has("ll") + # assert not "hello".has("lll") + fun has(pattern: Pattern): Bool do return pattern.is_in(self) end