1 /*****************************************************************************
3 *****************************************************************************
4 * Copyright (C) 2003-2006 the VideoLAN team
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
11 * it under the terms of the GNU General Public License as published by
12 * the Free Software Foundation; either version 2 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 General Public License for more details.
20 * You should have received a copy of the GNU General Public License
21 * along with this program; if not, write to the Free Software
22 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
23 *****************************************************************************/
27 * This file contains functions to dectect subtitle files.
32 #include <vlc/input.h>
50 * What's between a directory and a filename?
53 #define DIRECTORY_SEPARATOR '\\'
55 #define DIRECTORY_SEPARATOR '/'
59 * We are not going to autodetect more subtitle files than this.
61 #define MAX_SUBTITLE_FILES 128
65 * The possible extensions for subtitle files we support
67 static const char * sub_exts[] = { "utf", "utf8", "utf-8", "sub", "srt", "smi", "txt", "ssa", "idx", NULL };
68 /* extensions from unsupported types */
69 /* rt, aqt, jss, js, ass */
71 static void strcpy_trim( char *d, char *s )
73 /* skip leading whitespace */
74 while( *s && !isalnum(*s) )
81 while( *s && isalnum(*s) )
87 /* trim excess whitespace */
88 while( *s && !isalnum(*s) )
98 static void strcpy_strip_ext( char *d, char *s )
100 char *tmp = strrchr(s, '.');
107 strlcpy(d, s, tmp - s);
115 static void strcpy_get_ext( char *d, char *s )
117 char *tmp = strrchr(s, '.');
122 } else strcpy( d, tmp + 1 );
125 static int whiteonly( char *s )
129 if( isalnum( *s ) ) return 0;
135 typedef struct _subfn
142 static int compare_sub_priority( const void *a, const void *b )
144 if (((subfn*)a)->priority > ((subfn*)b)->priority)
149 if (((subfn*)a)->priority < ((subfn*)b)->priority)
155 return strcoll(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
157 return strcmp(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
162 * Check if a file ends with a subtitle extension
164 int subtitles_Filter( const char *psz_dir_content )
166 const char *tmp = strrchr( psz_dir_content, '.');
174 for( i = 0; sub_exts[i]; i++ )
175 if( strcmp( sub_exts[i], tmp ) == 0 )
183 * Convert a list of paths separated by ',' to a char**
185 static char **paths_to_list( char *psz_dir, char *psz_path )
187 unsigned int i, k, i_nb_subdirs;
188 char **subdirs; /* list of subdirectories to look in */
190 if( !psz_dir ) return NULL;
191 if( !psz_path ) return NULL;
194 for( k = 0; k < strlen( psz_path ); k++ )
196 if( psz_path[k] == ',' )
202 if( i_nb_subdirs > 0 )
204 char *psz_parser = NULL, *psz_temp = NULL;
206 subdirs = (char**)malloc( sizeof(char*) * ( i_nb_subdirs + 1 ) );
207 memset( subdirs, 0, sizeof(char*) * ( i_nb_subdirs + 1 ) );
209 psz_parser = psz_path;
210 while( psz_parser && *psz_parser )
213 psz_subdir = psz_parser;
214 psz_parser = strchr( psz_subdir, ',' );
219 while( *psz_parser == ' ' )
224 if( strlen( psz_subdir ) > 0 )
226 psz_temp = (char *)malloc( strlen(psz_dir)
227 + strlen(psz_subdir) + 2 );
230 sprintf( psz_temp, "%s%s%c",
231 psz_subdir[0] == '.' ? psz_dir : "",
233 psz_subdir[strlen(psz_subdir) - 1] ==
234 DIRECTORY_SEPARATOR ? '\0' : DIRECTORY_SEPARATOR );
235 subdirs[i] = psz_temp;
251 * Detect subtitle files.
253 * When called this function will split up the psz_name string into a
254 * directory, filename and extension. It then opens the directory
255 * in which the file resides and tries to find possible matches of
259 * \param p_this the calling \ref input_thread_t
260 * \param psz_path a list of subdirectories (separated by a ',') to look in.
261 * \param psz_name the complete filename to base the search on.
262 * \return a NULL terminated array of filenames with detected possible subtitles.
263 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
265 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
269 int j, i_result2, i_sub_count = 0, i_fname_len = 0;
270 char *f_dir = NULL, *f_fname = NULL, *f_fname_noext = NULL, *f_fname_trim = NULL;
273 char **tmp_subdirs, **subdirs; /* list of subdirectories to look in */
275 subfn *result = NULL; /* unsorted results */
276 char **result2; /* sorted results */
278 char *psz_fname_original = strdup( psz_name );
279 char *psz_fname = psz_fname_original;
281 if( psz_fname == NULL ) return NULL;
283 if( !strncmp( psz_fname, "file://", 7 ) )
288 /* extract filename & dirname from psz_fname */
289 tmp = strrchr( psz_fname, DIRECTORY_SEPARATOR );
294 f_fname = malloc( strlen(tmp) );
296 strcpy( f_fname, tmp+1 ); // we skip the separator, so it will still fit in the allocated space
297 dirlen = strlen(psz_fname) - strlen(tmp) + 2; // add the separator
298 f_dir = malloc( dirlen + 1 );
300 strlcpy( f_dir, psz_fname, dirlen );
304 /* Get the current working directory */
307 f_dir = getcwd( NULL, 0 );
311 if( psz_fname_original ) free( psz_fname_original );
314 dirlen = strlen( f_dir );
315 f_dir = (char *)realloc(f_dir, dirlen +2 );
316 f_dir[dirlen] = DIRECTORY_SEPARATOR;
317 f_dir[dirlen+1] = '\0';
318 f_fname = FromLocaleDup( psz_fname );
321 i_fname_len = strlen( f_fname );
322 f_fname_noext = malloc(i_fname_len + 1);
323 f_fname_trim = malloc(i_fname_len + 1 );
325 strcpy_strip_ext( f_fname_noext, f_fname );
326 strcpy_trim( f_fname_trim, f_fname_noext );
328 result = (subfn*)malloc( sizeof(subfn) * MAX_SUBTITLE_FILES );
330 memset( result, 0, sizeof(subfn) * MAX_SUBTITLE_FILES );
332 var_Get( p_this, "sub-autodetect-fuzzy", &fuzzy );
334 tmp_subdirs = paths_to_list( f_dir, psz_path );
335 subdirs = tmp_subdirs;
337 for( j = -1; (j == -1) || ( (j >= 0) && (subdirs != NULL) &&
338 (*subdirs != NULL) ); j++)
340 const char *psz_dir = j < 0 ? f_dir : *subdirs;
341 char **ppsz_dir_content;
344 if( psz_dir == NULL )
347 /* parse psz_src dir */
348 i_dir_content = utf8_scandir( psz_dir, &ppsz_dir_content,
349 subtitles_Filter, NULL );
351 if( i_dir_content != -1 )
355 msg_Dbg( p_this, "looking for a subtitle file in %s", psz_dir );
356 for( a = 0; a < i_dir_content; a++ )
358 char *psz_name = vlc_fix_readdir_charset( p_this,
359 ppsz_dir_content[a] );
360 char tmp_fname_noext[strlen( psz_name ) + 1];
361 char tmp_fname_trim[strlen( psz_name ) + 1];
362 char tmp_fname_ext[strlen( psz_name ) + 1];
366 if( psz_name == NULL )
369 /* retrieve various parts of the filename */
370 strcpy_strip_ext( tmp_fname_noext, psz_name );
371 strcpy_get_ext( tmp_fname_ext, psz_name );
372 strcpy_trim( tmp_fname_trim, tmp_fname_noext );
374 if( !i_prio && !strcmp( tmp_fname_trim, f_fname_trim ) )
376 /* matches the movie name exactly */
380 ( tmp = strstr( tmp_fname_trim, f_fname_trim ) ) )
382 /* contains the movie name */
383 tmp += strlen( f_fname_trim );
384 if( whiteonly( tmp ) )
386 /* chars in front of the movie name */
391 /* chars after (and possibly in front of)
398 /* doesn't contain the movie name */
399 if( j == 0 ) i_prio = 1;
401 if( i_prio >= fuzzy.i_int )
404 char psz_path[strlen( psz_dir ) + strlen( psz_name ) + 1];
406 sprintf( psz_path, "%s%s", psz_dir, psz_name );
408 "autodetected subtitle: %s with priority %d",
410 /* FIXME: a portable wrapper for stat() or access() would be more suited */
411 if( ( f = utf8_fopen( psz_path, "rt" ) ) )
415 "autodetected subtitle: %s with priority %d",
417 result[i_sub_count].priority = i_prio;
418 result[i_sub_count].psz_fname = strdup( psz_path );
419 result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
424 msg_Dbg( p_this, "fopen failed" );
427 if( i_sub_count >= MAX_SUBTITLE_FILES ) break;
430 for( a = 0; a < i_dir_content; a++ )
431 free( ppsz_dir_content[a] );
432 if( ppsz_dir_content ) free( ppsz_dir_content );
434 if( j >= 0 ) if( *subdirs ) free( *subdirs++ );
437 if( tmp_subdirs ) free( tmp_subdirs );
438 if( f_fname_trim ) free( f_fname_trim );
439 if( f_fname_noext ) free( f_fname_noext );
440 if( f_fname ) free( f_fname );
441 if( f_dir ) free( f_dir );
443 qsort( result, i_sub_count, sizeof( subfn ), compare_sub_priority );
445 result2 = (char**)malloc( sizeof(char*) * ( i_sub_count + 1 ) );
447 memset( result2, 0, sizeof(char*) * ( i_sub_count + 1 ) );
450 for( j = 0; j < i_sub_count; j++ )
452 if( result[j].psz_ext && !strcasecmp( result[j].psz_ext, "sub" ) )
455 for( i = 0; i < i_sub_count; i++ )
457 if( result[i].psz_fname && result[j].psz_fname &&
458 !strncasecmp( result[j].psz_fname, result[i].psz_fname,
459 sizeof( result[j].psz_fname) - 4 ) &&
460 !strcasecmp( result[i].psz_ext, "idx" ) )
463 if( i >= i_sub_count )
465 result2[i_result2] = result[j].psz_fname;
471 result2[i_result2] = result[j].psz_fname;
476 if( psz_fname_original ) free( psz_fname_original );
477 if( result ) free( result );