1 /*****************************************************************************
2 * subtitles.c : subtitles detection
3 *****************************************************************************
4 * Copyright (C) 2003-2009 VLC authors and VideoLAN
7 * Authors: Derk-Jan Hartman <hartman at videolan.org>
8 * This is adapted code from the GPL'ed MPlayer (http://mplayerhq.hu)
10 * This program is free software; you can redistribute it and/or modify it
11 * under the terms of the GNU Lesser General Public License as published by
12 * the Free Software Foundation; either version 2.1 of the License, or
13 * (at your option) any later version.
15 * This program is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU Lesser General Public License for more details.
20 * You should have received a copy of the GNU Lesser General Public License
21 * along with this program; if not, write to the Free Software Foundation,
22 * Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
23 *****************************************************************************/
27 * This file contains functions to dectect subtitle files.
34 #include <ctype.h> /* isalnum() */
38 #include <vlc_common.h>
42 #include "input_internal.h"
45 * We are not going to autodetect more subtitle files than this.
47 #define MAX_SUBTITLE_FILES 128
50 * The possible extensions for subtitle files we support
52 static const char sub_exts
[][6] = {
55 "utf", "utf8", "utf-8",
58 "psb", "mpsub","mpl2",
63 static void strcpy_trim( char *d
, const char *s
)
67 /* skip leading whitespace */
68 while( ((c
= *s
) != '\0') && !isalnum(c
) )
75 while( ((c
= *s
) != '\0') && isalnum(c
) )
81 /* trim excess whitespace */
82 while( ((c
= *s
) != '\0') && !isalnum(c
) )
92 static void strcpy_strip_ext( char *d
, const char *s
)
96 const char *tmp
= strrchr(s
, '.');
103 strlcpy(d
, s
, tmp
- s
+ 1 );
104 while( (c
= *d
) != '\0' )
111 static void strcpy_get_ext( char *d
, const char *s
)
113 const char *tmp
= strrchr(s
, '.');
117 strcpy( d
, tmp
+ 1 );
120 static int whiteonly( const char *s
)
124 while( (c
= *s
) != '\0' )
135 SUB_PRIORITY_NONE
= 0,
136 SUB_PRIORITY_MATCH_NONE
= 1,
137 SUB_PRIORITY_MATCH_RIGHT
= 2,
138 SUB_PRIORITY_MATCH_LEFT
= 3,
139 SUB_PRIORITY_MATCH_ALL
= 4,
148 static int compare_sub_priority( const void *a
, const void *b
)
150 const vlc_subfn_t
*p0
= a
;
151 const vlc_subfn_t
*p1
= b
;
153 if( p0
->priority
> p1
->priority
)
156 if( p0
->priority
< p1
->priority
)
160 return strcoll( p0
->psz_fname
, p1
->psz_fname
);
162 return strcmp( p0
->psz_fname
, p1
->psz_fname
);
167 * Check if a file ends with a subtitle extension
169 int subtitles_Filter( const char *psz_dir_content
)
171 const char *tmp
= strrchr( psz_dir_content
, '.');
177 for( int i
= 0; sub_exts
[i
][0]; i
++ )
178 if( strcasecmp( sub_exts
[i
], tmp
) == 0 )
185 * Convert a list of paths separated by ',' to a char**
187 static char **paths_to_list( const char *psz_dir
, char *psz_path
)
189 unsigned int i
, k
, i_nb_subdirs
;
190 char **subdirs
; /* list of subdirectories to look in */
191 char *psz_parser
= psz_path
;
193 if( !psz_dir
|| !psz_path
)
196 for( k
= 0, i_nb_subdirs
= 1; psz_path
[k
] != '\0'; k
++ )
198 if( psz_path
[k
] == ',' )
202 subdirs
= calloc( i_nb_subdirs
+ 1, sizeof(char*) );
206 for( i
= 0; psz_parser
&& *psz_parser
!= '\0' ; )
208 char *psz_subdir
= psz_parser
;
209 psz_parser
= strchr( psz_subdir
, ',' );
212 *psz_parser
++ = '\0';
213 while( *psz_parser
== ' ' )
216 if( *psz_subdir
== '\0' )
219 if( asprintf( &subdirs
[i
++], "%s%s",
220 psz_subdir
[0] == '.' ? psz_dir
: "",
231 * Detect subtitle files.
233 * When called this function will split up the psz_name string into a
234 * directory, filename and extension. It then opens the directory
235 * in which the file resides and tries to find possible matches of
239 * \param p_this the calling \ref input_thread_t
240 * \param psz_path a list of subdirectories (separated by a ',') to look in.
241 * \param psz_name the complete filename to base the search on.
242 * \return a NULL terminated array of filenames with detected possible subtitles.
243 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
245 char **subtitles_Detect( input_thread_t
*p_this
, char *psz_path
,
246 const char *psz_name_org
)
248 int i_fuzzy
= var_GetInteger( p_this
, "sub-autodetect-fuzzy" );
251 int j
, i_result2
, i_sub_count
, i_fname_len
;
252 char *f_fname_noext
= NULL
, *f_fname_trim
= NULL
;
254 char **subdirs
; /* list of subdirectories to look in */
256 vlc_subfn_t
*result
= NULL
; /* unsorted results */
257 char **result2
; /* sorted results */
262 char *psz_fname
= make_path( psz_name_org
);
266 /* extract filename & dirname from psz_fname */
267 char *f_dir
= strdup( psz_fname
);
274 const char *f_fname
= strrchr( psz_fname
, DIR_SEP_CHAR
);
281 f_fname
++; /* Skip the '/' */
282 f_dir
[f_fname
- psz_fname
] = 0; /* keep dir separator in f_dir */
284 i_fname_len
= strlen( f_fname
);
286 f_fname_noext
= malloc(i_fname_len
+ 1);
287 f_fname_trim
= malloc(i_fname_len
+ 1 );
288 if( !f_fname_noext
|| !f_fname_trim
)
291 free( f_fname_noext
);
292 free( f_fname_trim
);
297 strcpy_strip_ext( f_fname_noext
, f_fname
);
298 strcpy_trim( f_fname_trim
, f_fname_noext
);
300 result
= calloc( MAX_SUBTITLE_FILES
+1, sizeof(vlc_subfn_t
) ); /* We check it later (simplify code) */
301 subdirs
= paths_to_list( f_dir
, psz_path
);
302 for( j
= -1, i_sub_count
= 0; (j
== -1) || ( j
>= 0 && subdirs
!= NULL
&& subdirs
[j
] != NULL
); j
++ )
304 const char *psz_dir
= (j
< 0) ? f_dir
: subdirs
[j
];
305 if( psz_dir
== NULL
|| ( j
>= 0 && !strcmp( psz_dir
, f_dir
) ) )
308 /* parse psz_src dir */
309 DIR *dir
= vlc_opendir( psz_dir
);
313 msg_Dbg( p_this
, "looking for a subtitle file in %s", psz_dir
);
315 const char *psz_name
;
316 while( (psz_name
= vlc_readdir( dir
)) && i_sub_count
< MAX_SUBTITLE_FILES
)
318 if( psz_name
[0] == '.' || !subtitles_Filter( psz_name
) )
321 char tmp_fname_noext
[strlen( psz_name
) + 1];
322 char tmp_fname_trim
[strlen( psz_name
) + 1];
323 char tmp_fname_ext
[strlen( psz_name
) + 1];
325 int i_prio
= SUB_PRIORITY_NONE
;
327 /* retrieve various parts of the filename */
328 strcpy_strip_ext( tmp_fname_noext
, psz_name
);
329 strcpy_get_ext( tmp_fname_ext
, psz_name
);
330 strcpy_trim( tmp_fname_trim
, tmp_fname_noext
);
332 if( !strcmp( tmp_fname_trim
, f_fname_trim
) )
334 /* matches the movie name exactly */
335 i_prio
= SUB_PRIORITY_MATCH_ALL
;
337 else if( (tmp
= strstr( tmp_fname_trim
, f_fname_trim
)) )
339 /* contains the movie name */
340 tmp
+= strlen( f_fname_trim
);
341 if( whiteonly( tmp
) )
343 /* chars in front of the movie name */
344 i_prio
= SUB_PRIORITY_MATCH_RIGHT
;
348 /* chars after (and possibly in front of)
350 i_prio
= SUB_PRIORITY_MATCH_LEFT
;
355 /* doesn't contain the movie name, prefer files in f_dir over subdirs */
356 i_prio
= SUB_PRIORITY_MATCH_NONE
;
358 if( i_prio
>= i_fuzzy
)
360 char psz_path
[strlen( psz_dir
) + strlen( psz_name
) + 2];
363 sprintf( psz_path
, "%s"DIR_SEP
"%s", psz_dir
, psz_name
);
364 if( !strcmp( psz_path
, psz_fname
) )
367 if( !vlc_stat( psz_path
, &st
) && S_ISREG( st
.st_mode
) && result
)
370 "autodetected subtitle: %s with priority %d",
372 result
[i_sub_count
].priority
= i_prio
;
373 result
[i_sub_count
].psz_fname
= strdup( psz_path
);
374 result
[i_sub_count
].psz_ext
= strdup(tmp_fname_ext
);
379 msg_Dbg( p_this
, "stat failed (autodetecting subtitle: %s with priority %d)",
388 for( j
= 0; subdirs
[j
]; j
++ )
393 free( f_fname_trim
);
394 free( f_fname_noext
);
400 qsort( result
, i_sub_count
, sizeof(vlc_subfn_t
), compare_sub_priority
);
402 result2
= calloc( i_sub_count
+ 1, sizeof(char*) );
404 for( j
= 0, i_result2
= 0; j
< i_sub_count
&& result2
!= NULL
; j
++ )
406 bool b_reject
= false;
408 if( !result
[j
].psz_fname
|| !result
[j
].psz_ext
) /* memory out */
411 if( !strcasecmp( result
[j
].psz_ext
, "sub" ) )
414 for( i
= 0; i
< i_sub_count
; i
++ )
416 if( result
[i
].psz_fname
&& result
[i
].psz_ext
&&
417 !strncasecmp( result
[j
].psz_fname
, result
[i
].psz_fname
,
418 strlen( result
[j
].psz_fname
) - 3 ) &&
419 !strcasecmp( result
[i
].psz_ext
, "idx" ) )
422 if( i
< i_sub_count
)
425 else if( !strcasecmp( result
[j
].psz_ext
, "cdg" ) )
427 if( result
[j
].priority
< SUB_PRIORITY_MATCH_ALL
)
433 result2
[i_result2
++] = strdup( result
[j
].psz_fname
);
436 for( j
= 0; j
< i_sub_count
; j
++ )
438 free( result
[j
].psz_fname
);
439 free( result
[j
].psz_ext
);