1 # This file is part of NIT ( http://www.nitlanguage.org ).
3 # Copyright 2005-2008 Jean Privat <jean@pryen.org>
5 # This file is free software, which comes along with NIT. This software is
6 # distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY;
7 # without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
8 # PARTICULAR PURPOSE. You can modify it is you want, provided this header
9 # is kept unaltered, and a notification of the changes is added.
10 # You are allowed to redistribute it and sell it, alone or is a part of
13 # Basic string search, match and replace.
18 # Patterns are abstract string motifs (include `String` and `Char`).
20 # Search `self` into `s` from a certain position.
21 # Return the position of the first character of the matching section.
22 # Return -1 if not found.
24 # assert 'l'.search_index_in("hello world", 0) == 2
25 # assert 'l'.search_index_in("hello world", 3) == 3
26 # assert 'z'.search_index_in("hello world", 0) == -1
28 # This method is usually faster than `search_in` if what is
29 # required is only the index.
30 # Note: in most implementation, `search_in` is implemented with this method.
31 protected fun search_index_in
(s
: String, from
: Int): Int is abstract
33 # Search `self` into `s` from a certain position.
34 # Return null if not found.
36 # assert 'l'.search_in("hello world", 0).from == 2
37 # assert 'l'.search_in("hello world", 3).from == 3
38 # assert 'z'.search_in("hello world", 0) == null
40 # If only the index of the first character if required, see `search_index_in`.
42 # Note: Is used by `String::search`, `String::search_from`, and others.
43 protected fun search_in
(s
: String, from
: Int): nullable Match is abstract
45 # Search all `self` occurrences into `s`.
47 # assert 'l'.search_all_in("hello world").length == 3
48 # assert 'z'.search_all_in("hello world").length == 0
50 # Note: Is used by `String::search_all`.
51 protected fun search_all_in
(s
: String): Array[Match]
53 var res
= new Array[Match] # Result
54 var match
= search_in
(s
, 0)
55 while match
!= null do
57 match
= search_in
(s
, match
.after
)
62 # Split `s` using `self` is separator.
64 # Returns an array of matches that are between each occurence of `self`.
65 # If self is not present, an array with a single match on `s` is retunred.
67 # assert 'l'.split_in("hello world").join("|") == "he||o wor|d"
68 # assert 'z'.split_in("hello world").join("|") == "hello world"
70 # Note: is used by `String::split`
71 protected fun split_in
(s
: String): Array[Match]
73 var res
= new Array[Match] # Result
75 var match
= search_in
(s
, 0)
76 while match
!= null do
77 # Compute the splited part length
78 var len
= match
.from
- i
79 res
.add
(new Match(s
, i
, len
))
81 match
= search_in
(s
, i
)
84 res
.add
(new Match(s
, i
, s
.length
- i
))
89 # BM_Pattern are pre-compiled string motif for the Boyer-Moore algorithm.
90 # (cf. A Fast String Searching Algorithm, with R.S. Boyer. Communications
91 # of the Association for Computing Machinery, 20(10), 1977, pp. 762-772.)
92 # http://www.cs.utexas.edu/users/moore/best-ideas/string-searching/index.html
94 # var pat = new BM_Pattern("hello")
95 # assert "I say hello to the world!".search(pat).from == 6
96 # assert "I say goodbye to the world!".search(pat) == null
100 redef fun to_s
do return _motif
102 # boyer-moore search gives the position of the first occurrence of a pattern starting at position `from`
103 redef fun search_index_in
(s
, from
)
110 while j
< n
- m
+ 1 do
111 var i
= m
- 1 # Cursor in the pattern
112 while i
>= 0 and _motif
.chars
[i
] == s
.chars
[i
+ j
] do i
-= 1
116 var gs
= _gs
[i
] # Good shift
117 var bc
= bc
(s
.chars
[i
+j
]) - m
+ 1 + i
# Bad char
118 # Both are true, do move to the best
126 return -1 # found nothing
129 # boyer-moore search. Return null if not found
130 redef fun search_in
(s
, from
)
132 var to
= search_index_in
(s
, from
)
136 return new Match(s
, to
, _length
)
140 # Compile a new motif
144 _length
= motif
.length
145 _gs
= new Array[Int].with_capacity
(_length
)
146 _bc_table
= new ArrayMap[Char, Int]
154 # length of the motif
157 private fun bc
(e
: Char): Int
159 if _bc_table
.has_key
(e
) then
170 var _bc_table
: Map[Char, Int]
172 private fun compute_bc
178 _bc_table
[x
.chars
[i
]] = m
- i
- 1
183 private fun suffixes
: Array[Int]
187 var suff
= new Array[Int].filled_with
(m
, m
)
193 if i
> g
and suff
[i
+m-1-f
] < i
- g
then
194 suff
[i
] = suff
[i
+m-1-f
]
198 while g
>= 0 and x
.chars
[g
] == x
.chars
[g
+ m
- 1 - f
] do g
-= 1
206 private fun compute_gs
219 if i
== -1 or suff
[i
] == i
+ 1 then
220 while j
< m
- 1 - i
do
221 if _gs
[j
] == m
then _gs
[j
] = m
- 1 - i
229 _gs
[m
- 1 - suff
[i
]] = m
- 1 - i
234 redef fun hash
do return _motif
.hash
235 redef fun ==(o
) do return o
isa BM_Pattern and o
._motif
== _motif
238 # Matches are a part of a `String` found ba a `Pattern`.
240 # The base string matched
241 readable var _string
: String
243 # The starting position in the string
244 readable var _from
: Int
246 # The length of the matching part
247 readable var _length
: Int
249 # The position of the first character just after the matching part.
250 # May be out of the base string
251 fun after
: Int do return _from
+ _length
253 # The contents of the matching part
254 redef fun to_s
do return _string
.substring
(_from
, _length
)
256 # Matches `len` characters of `s` from `f`.
257 init(s
: String, f
: Int, len
: Int)
259 assert positive_length
: len
>= 0
260 assert valid_from
: f
>= 0
261 assert valid_after
: f
+ len
<= s
.length
271 redef fun search_index_in
(s
, from
)
275 if s
.chars
[from
] == self then return from
281 redef fun search_in
(s
, from
)
283 var pos
= search_index_in
(s
, from
)
287 return new Match(s
, pos
, 1)
295 redef fun search_index_in
(s
, from
)
298 var stop
= s
.length
- length
+ 1
301 while i
>= 0 and self.chars
[i
] == s
.chars
[i
+ from
] do i
-= 1
303 if i
< 0 then return from
304 # Not found so try next one
310 redef fun search_in
(s
, from
)
312 var pos
= search_index_in
(s
, from
)
316 return new Match(s
, pos
, length
)
320 # Search the first occurence of the pattern `p`.
321 # Return null if not found.
323 # assert "I say hello to the world!".search("hello").from == 6
324 # assert "I say goodbye to the world!".search("hello") == null
325 fun search
(p
: Pattern): nullable Match do return p
.search_in
(self, 0)
327 # Search the first occurence of the pattern `p` after `from`.
328 # The search starts at `from`.
329 # Return null if not found.
331 # assert "I say hello to the world!".search_from("hello",4).from == 6
332 # assert "I say hello to the world!".search_from("hello",7) == null
333 fun search_from
(p
: Pattern, from
: Int): nullable Match do return p
.search_in
(self, from
)
335 # Search all occurrences of p into self.
337 # var a = new Array[Int]
338 # for i in "hello world".search_all('o') do
342 fun search_all
(p
: Pattern): Array[Match] do return p
.search_all_in
(self)
344 # Split `self` using `p` as separator.
346 # assert "hello world".split('o') == ["hell", " w", "rld"]
347 fun split
(p
: Pattern): Array[String]
349 var matches
= p
.split_in
(self)
350 var res
= new Array[String].with_capacity
(matches
.length
)
351 for m
in matches
do res
.add
(m
.to_s
)
355 # @deprecated alias for `split`
356 fun split_with
(p
: Pattern): Array[String] do return self.split
(p
)
358 # Replace all occurences of a pattern with a string
360 # assert "hlelo".replace("le", "el") == "hello"
361 # assert "hello".replace('l', "") == "heo"
362 fun replace
(p
: Pattern, string
: String): String
364 return self.split_with
(p
).join
(string
)
367 # Escape the four characters `<`, `>`, `&`, and `"` with their html counterpart
369 # assert "a&b->\"x\"".html_escape == "a&b->"x""
370 fun html_escape
: String
373 if ret
.chars
.has
('&') then ret
= ret
.replace
('&', "&")
374 if ret
.chars
.has
('<') then ret
= ret
.replace
('<', "<")
375 if ret
.chars
.has
('>') then ret
= ret
.replace
('>', ">")
376 if ret
.chars
.has
('"') then ret
= ret
.replace
('"', """)