Commit | Line | Data |
af1f55d9 |
1 | package Encode::Guess; |
2 | use strict; |
7e19fb92 |
3 | |
af1f55d9 |
4 | use Encode qw(:fallbacks find_encoding); |
e8c86ba6 |
5 | our $VERSION = do { my @r = (q$Revision: 1.4 $ =~ /\d+/g); sprintf "%d."."%02d" x $#r, @r }; |
af1f55d9 |
6 | |
7 | my $Canon = 'Guess'; |
af1f55d9 |
8 | our $DEBUG = 0; |
7e19fb92 |
9 | our %DEF_SUSPECTS = map { $_ => find_encoding($_) } qw(ascii utf8); |
10 | $Encode::Encoding{$Canon} = |
11 | bless { |
12 | Name => $Canon, |
13 | Suspects => { %DEF_SUSPECTS }, |
14 | } => __PACKAGE__; |
15 | |
10c5ecbb |
16 | use base qw(Encode::Encoding); |
7e19fb92 |
17 | sub needs_lines { 1 } |
18 | sub perlio_ok { 0 } |
7e19fb92 |
19 | |
20 | our @EXPORT = qw(guess_encoding); |
21 | |
22 | sub import { # Exporter not used so we do it on our own |
23 | my $callpkg = caller; |
24 | for my $item (@EXPORT){ |
25 | no strict 'refs'; |
26 | *{"$callpkg\::$item"} = \&{"$item"}; |
27 | } |
28 | set_suspects(@_); |
29 | } |
af1f55d9 |
30 | |
7e19fb92 |
31 | sub set_suspects{ |
32 | my $class = shift; |
33 | my $self = ref($class) ? $class : $Encode::Encoding{$Canon}; |
34 | $self->{Suspects} = { %DEF_SUSPECTS }; |
35 | $self->add_suspects(@_); |
36 | } |
af1f55d9 |
37 | |
7e19fb92 |
38 | sub add_suspects{ |
af1f55d9 |
39 | my $class = shift; |
7e19fb92 |
40 | my $self = ref($class) ? $class : $Encode::Encoding{$Canon}; |
af1f55d9 |
41 | for my $c (@_){ |
42 | my $e = find_encoding($c) or die "Unknown encoding: $c"; |
7e19fb92 |
43 | $self->{Suspects}{$e->name} = $e; |
af1f55d9 |
44 | $DEBUG and warn "Added: ", $e->name; |
45 | } |
46 | } |
47 | |
af1f55d9 |
48 | sub decode($$;$){ |
49 | my ($obj, $octet, $chk) = @_; |
7e19fb92 |
50 | my $guessed = guess($obj, $octet); |
10c5ecbb |
51 | unless (ref($guessed)){ |
52 | require Carp; |
53 | Carp::croak($guessed); |
54 | } |
7e19fb92 |
55 | my $utf8 = $guessed->decode($octet, $chk); |
af1f55d9 |
56 | $_[1] = $octet if $chk; |
57 | return $utf8; |
58 | } |
59 | |
7e19fb92 |
60 | sub guess_encoding{ |
61 | guess($Encode::Encoding{$Canon}, @_); |
af1f55d9 |
62 | } |
63 | |
64 | sub guess { |
7e19fb92 |
65 | my $class = shift; |
66 | my $obj = ref($class) ? $class : $Encode::Encoding{$Canon}; |
67 | my $octet = shift; |
68 | # cheat 0: utf8 flag; |
af1f55d9 |
69 | Encode::is_utf8($octet) and return find_encoding('utf8'); |
7e19fb92 |
70 | # cheat 1: BOM |
71 | use Encode::Unicode; |
72 | my $BOM = unpack('n', $octet); |
73 | return find_encoding('UTF-16') |
74 | if ($BOM == 0xFeFF or $BOM == 0xFFFe); |
75 | $BOM = unpack('N', $octet); |
76 | return find_encoding('UTF-32') |
77 | if ($BOM == 0xFeFF or $BOM == 0xFFFe0000); |
78 | |
79 | my %try = %{$obj->{Suspects}}; |
80 | for my $c (@_){ |
81 | my $e = find_encoding($c) or die "Unknown encoding: $c"; |
82 | $try{$e->name} = $e; |
83 | $DEBUG and warn "Added: ", $e->name; |
84 | } |
af1f55d9 |
85 | my $nline = 1; |
86 | for my $line (split /\r|\n|\r\n/, $octet){ |
7e19fb92 |
87 | # cheat 2 -- \e in the string |
af1f55d9 |
88 | if ($line =~ /\e/o){ |
89 | my @keys = keys %try; |
90 | delete @try{qw/utf8 ascii/}; |
91 | for my $k (@keys){ |
92 | ref($try{$k}) eq 'Encode::XS' and delete $try{$k}; |
93 | } |
94 | } |
95 | my %ok = %try; |
96 | # warn join(",", keys %try); |
97 | for my $k (keys %try){ |
98 | my $scratch = $line; |
99 | $try{$k}->decode($scratch, FB_QUIET); |
100 | if ($scratch eq ''){ |
101 | $DEBUG and warn sprintf("%4d:%-24s ok\n", $nline, $k); |
102 | }else{ |
103 | use bytes (); |
104 | $DEBUG and |
105 | warn sprintf("%4d:%-24s not ok; %d bytes left\n", |
106 | $nline, $k, bytes::length($scratch)); |
107 | delete $ok{$k}; |
108 | |
109 | } |
110 | } |
7e19fb92 |
111 | %ok or return "No appropriate encodings found!"; |
af1f55d9 |
112 | if (scalar(keys(%ok)) == 1){ |
113 | my ($retval) = values(%ok); |
114 | return $retval; |
115 | } |
116 | %try = %ok; $nline++; |
117 | } |
7e19fb92 |
118 | $try{ascii} or |
119 | return "Encodings too ambiguous: ", join(" or ", keys %try); |
af1f55d9 |
120 | return $try{ascii}; |
121 | } |
122 | |
123 | |
7e19fb92 |
124 | |
af1f55d9 |
125 | 1; |
126 | __END__ |
127 | |
128 | =head1 NAME |
129 | |
7e19fb92 |
130 | Encode::Guess -- Guesses encoding from data |
131 | |
132 | =head1 SYNOPSIS |
133 | |
134 | # if you are sure $data won't contain anything bogus |
135 | |
e8c86ba6 |
136 | use Encode; |
7e19fb92 |
137 | use Encode::Guess qw/euc-jp shiftjis 7bit-jis/; |
138 | my $utf8 = decode("Guess", $data); |
139 | my $data = encode("Guess", $utf8); # this doesn't work! |
140 | |
141 | # more elaborate way |
142 | use Encode::Guess, |
143 | my $enc = guess_encoding($data, qw/euc-jp shiftjis 7bit-jis/); |
144 | ref($enc) or die "Can't guess: $enc"; # trap error this way |
145 | $utf8 = $enc->decode($data); |
146 | # or |
147 | $utf8 = decode($enc->name, $data) |
148 | |
149 | =head1 ABSTRACT |
150 | |
151 | Encode::Guess enables you to guess in what encoding a given data is |
152 | encoded, or at least tries to. |
153 | |
154 | =head1 DESCRIPTION |
155 | |
156 | By default, it checks only ascii, utf8 and UTF-16/32 with BOM. |
157 | |
158 | use Encode::Guess; # ascii/utf8/BOMed UTF |
159 | |
160 | To use it more practically, you have to give the names of encodings to |
161 | check (I<suspects> as follows). The name of suspects can either be |
162 | canonical names or aliases. |
163 | |
164 | # tries all major Japanese Encodings as well |
165 | use Encode::Guess qw/euc-jp shiftjis 7bit-jis/; |
166 | |
167 | =over 4 |
168 | |
169 | =item Encode::Guess->set_suspects |
170 | |
171 | You can also change the internal suspects list via C<set_suspects> |
172 | method. |
173 | |
174 | use Encode::Guess; |
175 | Encode::Guess->set_suspects(qw/euc-jp shiftjis 7bit-jis/); |
176 | |
177 | =item Encode::Guess->add_suspects |
178 | |
179 | Or you can use C<add_suspects> method. The difference is that |
180 | C<set_suspects> flushes the current suspects list while |
181 | C<add_suspects> adds. |
182 | |
183 | use Encode::Guess; |
184 | Encode::Guess->add_suspects(qw/euc-jp shiftjis 7bit-jis/); |
185 | # now the suspects are euc-jp,shiftjis,7bit-jis, AND |
186 | # euc-kr,euc-cn, and big5-eten |
187 | Encode::Guess->add_suspects(qw/euc-kr euc-cn big5-eten/); |
188 | |
189 | =item Encode::decode("Guess" ...) |
190 | |
191 | When you are content with suspects list, you can now |
192 | |
193 | my $utf8 = Encode::decode("Guess", $data); |
194 | |
195 | =item Encode::Guess->guess($data) |
196 | |
197 | But it will croak if Encode::Guess fails to eliminate all other |
198 | suspects but the right one or no suspect was good. So you should |
199 | instead try this; |
200 | |
201 | my $decoder = Encode::Guess->guess($data); |
202 | |
203 | On success, $decoder is an object that is documented in |
204 | L<Encode::Encoding>. So you can now do this; |
205 | |
206 | my $utf8 = $decoder->decode($data); |
207 | |
208 | On failure, $decoder now contains an error message so the whole thing |
209 | would be as follows; |
210 | |
211 | my $decoder = Encode::Guess->guess($data); |
212 | die $decoder unless ref($decoder); |
213 | my $utf8 = $decoder->decode($data); |
214 | |
215 | =item guess_encoding($data, [, I<list of suspects>]) |
216 | |
217 | You can also try C<guess_encoding> function which is exported by |
218 | default. It takes $data to check and it also takes the list of |
219 | suspects by option. The optional suspect list is I<not reflected> to |
220 | the internal suspects list. |
221 | |
222 | my $decoder = guess_encoding($data, qw/euc-jp euc-kr euc-cn/); |
223 | die $decoder unless ref($decoder); |
224 | my $utf8 = $decoder->decode($data); |
225 | # check only ascii and utf8 |
226 | my $decoder = guess_encoding($data); |
227 | |
228 | =back |
229 | |
230 | =head1 CAVEATS |
231 | |
232 | =over 4 |
233 | |
234 | =item * |
235 | |
236 | Because of the algorithm used, ISO-8859 series and other single-byte |
237 | encodings do not work well unless either one of ISO-8859 is the only |
238 | one suspect (besides ascii and utf8). |
239 | |
240 | use Encode::Guess; |
241 | # perhaps ok |
242 | my $decoder = guess_encoding($data, 'latin1'); |
243 | # definitely NOT ok |
244 | my $decoder = guess_encoding($data, qw/latin1 greek/); |
245 | |
246 | The reason is that Encode::Guess guesses encoding by trial and error. |
247 | It first splits $data into lines and tries to decode the line for each |
248 | suspect. It keeps it going until all but one encoding was eliminated |
249 | out of suspects list. ISO-8859 series is just too successful for most |
250 | cases (because it fills almost all code points in \x00-\xff). |
251 | |
252 | =item * |
253 | |
254 | Do not mix national standard encodings and the corresponding vendor |
255 | encodings. |
256 | |
257 | # a very bad idea |
258 | my $decoder |
259 | = guess_encoding($data, qw/shiftjis MacJapanese cp932/); |
260 | |
261 | The reason is that vendor encoding is usually a superset of national |
262 | standard so it becomes too ambiguous for most cases. |
263 | |
264 | =item * |
265 | |
266 | On the other hand, mixing various national standard encodings |
267 | automagically works unless $data is too short to allow for guessing. |
268 | |
269 | # This is ok if $data is long enough |
270 | my $decoder = |
271 | guess_encoding($data, qw/euc-cn |
272 | euc-jp shiftjis 7bit-jis |
273 | euc-kr |
274 | big5-eten/); |
275 | |
276 | =item * |
277 | |
278 | DO NOT PUT TOO MANY SUSPECTS! Don't you try something like this! |
279 | |
280 | my $decoder = guess_encoding($data, |
281 | Encode->encodings(":all")); |
282 | |
283 | =back |
284 | |
285 | It is, after all, just a guess. You should alway be explicit when it |
286 | comes to encodings. But there are some, especially Japanese, |
287 | environment that guess-coding is a must. Use this module with care. |
288 | |
289 | =head1 SEE ALSO |
290 | |
291 | L<Encode>, L<Encode::Encoding> |
af1f55d9 |
292 | |
293 | =cut |
294 | |