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.
34 #include <vlc_common.h>
35 #include <vlc_input.h>
36 #include <vlc_charset.h>
51 #include "input_internal.h"
54 * We are not going to autodetect more subtitle files than this.
56 #define MAX_SUBTITLE_FILES 128
60 * The possible extensions for subtitle files we support
62 static const char const sub_exts[][6] = {
65 "utf", "utf8", "utf-8",
68 "psb", "mpsub","mpl2",
73 static void strcpy_trim( char *d, const char *s )
75 /* skip leading whitespace */
76 while( *s && !isalnum(*s) )
83 while( *s && isalnum(*s) )
89 /* trim excess whitespace */
90 while( *s && !isalnum(*s) )
100 static void strcpy_strip_ext( char *d, const char *s )
102 const char *tmp = strrchr(s, '.');
109 strlcpy(d, s, tmp - s + 1 );
117 static void strcpy_get_ext( char *d, const char *s )
119 const char *tmp = strrchr(s, '.');
123 strcpy( d, tmp + 1 );
126 static int whiteonly( const char *s )
139 SUB_PRIORITY_NONE = 0,
140 SUB_PRIORITY_MATCH_NONE = 1,
141 SUB_PRIORITY_MATCH_RIGHT = 2,
142 SUB_PRIORITY_MATCH_LEFT = 3,
143 SUB_PRIORITY_MATCH_ALL = 4,
152 static int compare_sub_priority( const void *a, const void *b )
154 const vlc_subfn_t *p0 = a;
155 const vlc_subfn_t *p1 = b;
157 if( p0->priority > p1->priority )
160 if( p0->priority < p1->priority )
164 return strcoll( p0->psz_fname, p1->psz_fname);
166 return strcmp( p0->psz_fname, p1->psz_fname);
171 * Check if a file ends with a subtitle extension
173 int subtitles_Filter( const char *psz_dir_content )
175 const char *tmp = strrchr( psz_dir_content, '.');
182 for( i = 0; sub_exts[i][0]; i++ )
183 if( strcasecmp( sub_exts[i], tmp ) == 0 )
190 * Convert a list of paths separated by ',' to a char**
192 static char **paths_to_list( const char *psz_dir, char *psz_path )
194 unsigned int i, k, i_nb_subdirs;
195 char **subdirs; /* list of subdirectories to look in */
196 char *psz_parser = psz_path;
198 if( !psz_dir || !psz_path )
201 for( k = 0, i_nb_subdirs = 1; psz_path[k] != '\0'; k++ )
203 if( psz_path[k] == ',' )
207 subdirs = calloc( i_nb_subdirs + 1, sizeof(char*) );
211 for( i = 0; psz_parser && *psz_parser != '\0' ; )
213 char *psz_subdir = psz_parser;
214 psz_parser = strchr( psz_subdir, ',' );
217 *psz_parser++ = '\0';
218 while( *psz_parser == ' ' )
221 if( *psz_subdir == '\0' )
224 if( asprintf( &subdirs[i++], "%s%s%c",
225 psz_subdir[0] == '.' ? psz_dir : "",
227 psz_subdir[strlen(psz_subdir) - 1] == DIR_SEP_CHAR ?
228 '\0' : DIR_SEP_CHAR ) == -1 )
238 * Detect subtitle files.
240 * When called this function will split up the psz_name string into a
241 * directory, filename and extension. It then opens the directory
242 * in which the file resides and tries to find possible matches of
246 * \param p_this the calling \ref input_thread_t
247 * \param psz_path a list of subdirectories (separated by a ',') to look in.
248 * \param psz_name the complete filename to base the search on.
249 * \return a NULL terminated array of filenames with detected possible subtitles.
250 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
252 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
253 const char *psz_name_org )
256 int j, i_result2, i_sub_count, i_fname_len;
257 char *f_dir = NULL, *f_fname = NULL, *f_fname_noext = NULL, *f_fname_trim = NULL;
260 char **subdirs; /* list of subdirectories to look in */
262 vlc_subfn_t *result = NULL; /* unsorted results */
263 char **result2; /* sorted results */
268 const char *psz_fname = decode_URI( psz_name_org );
269 if( !strncmp( psz_fname, "file://", 7 ) )
272 if( !strncmp( psz_fname, "localhost", 9 ) )
276 /* extract filename & dirname from psz_fname */
277 tmp = strrchr( psz_fname, DIR_SEP_CHAR );
280 const int i_dirlen = strlen(psz_fname)-strlen(tmp)+1; /* include the separator */
281 f_fname = strdup( &tmp[1] ); /* skip the separator */
282 f_dir = strndup( psz_fname, i_dirlen );
286 #if defined (HAVE_UNISTD_H) && !defined (UNDER_CE)
287 /* Get the current working directory */
288 char *psz_cwd = getcwd( NULL, 0 );
290 char *psz_cwd = NULL;
295 f_fname = strdup( psz_fname );
296 if( asprintf( &f_dir, "%s%c", psz_cwd, DIR_SEP_CHAR ) == -1 )
297 f_dir = NULL; /* Assure that function will return in next test */
300 if( !f_fname || !f_dir )
307 i_fname_len = strlen( f_fname );
309 f_fname_noext = malloc(i_fname_len + 1);
310 f_fname_trim = malloc(i_fname_len + 1 );
311 if( !f_fname_noext || !f_fname_trim )
315 free( f_fname_noext );
316 free( f_fname_trim );
320 strcpy_strip_ext( f_fname_noext, f_fname );
321 strcpy_trim( f_fname_trim, f_fname_noext );
323 i_fuzzy = var_GetInteger( p_this, "sub-autodetect-fuzzy" );
325 result = calloc( MAX_SUBTITLE_FILES+1, sizeof(vlc_subfn_t) ); /* We check it later (simplify code) */
326 subdirs = paths_to_list( f_dir, psz_path );
327 for( j = -1, i_sub_count = 0; (j == -1) || ( j >= 0 && subdirs != NULL && subdirs[j] != NULL ); j++ )
329 const char *psz_dir = j < 0 ? f_dir : subdirs[j];
330 char **ppsz_dir_content;
334 if( psz_dir == NULL || ( j >= 0 && !strcmp( psz_dir, f_dir ) ) )
337 /* parse psz_src dir */
338 i_dir_content = utf8_scandir( psz_dir, &ppsz_dir_content,
339 subtitles_Filter, NULL );
340 if( i_dir_content < 0 )
343 msg_Dbg( p_this, "looking for a subtitle file in %s", psz_dir );
344 for( a = 0; a < i_dir_content && i_sub_count < MAX_SUBTITLE_FILES ; a++ )
346 char *psz_name = ppsz_dir_content[a];
347 char tmp_fname_noext[strlen( psz_name ) + 1];
348 char tmp_fname_trim[strlen( psz_name ) + 1];
349 char tmp_fname_ext[strlen( psz_name ) + 1];
353 if( psz_name == NULL || psz_name[0] == '.' )
356 /* retrieve various parts of the filename */
357 strcpy_strip_ext( tmp_fname_noext, psz_name );
358 strcpy_get_ext( tmp_fname_ext, psz_name );
359 strcpy_trim( tmp_fname_trim, tmp_fname_noext );
361 i_prio = SUB_PRIORITY_NONE;
362 if( i_prio == SUB_PRIORITY_NONE && !strcmp( tmp_fname_trim, f_fname_trim ) )
364 /* matches the movie name exactly */
365 i_prio = SUB_PRIORITY_MATCH_ALL;
367 if( i_prio == SUB_PRIORITY_NONE &&
368 ( tmp = strstr( tmp_fname_trim, f_fname_trim ) ) )
370 /* contains the movie name */
371 tmp += strlen( f_fname_trim );
372 if( whiteonly( tmp ) )
374 /* chars in front of the movie name */
375 i_prio = SUB_PRIORITY_MATCH_RIGHT;
379 /* chars after (and possibly in front of)
381 i_prio = SUB_PRIORITY_MATCH_LEFT;
384 if( i_prio == SUB_PRIORITY_NONE &&
387 /* doesn't contain the movie name, prefer files in f_dir over subdirs */
388 i_prio = SUB_PRIORITY_MATCH_NONE;
390 if( i_prio >= i_fuzzy )
392 char psz_path[strlen( psz_dir ) + strlen( psz_name ) + 1];
395 sprintf( psz_path, "%s%s", psz_dir, psz_name );
396 if( !strcmp( psz_path, psz_fname ) )
399 if( !utf8_stat( psz_path, &st ) && S_ISREG( st.st_mode ) && result )
402 "autodetected subtitle: %s with priority %d",
404 result[i_sub_count].priority = i_prio;
405 result[i_sub_count].psz_fname = strdup( psz_path );
406 result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
411 msg_Dbg( p_this, "stat failed (autodetecting subtitle: %s with priority %d)",
416 if( ppsz_dir_content )
418 for( a = 0; a < i_dir_content; a++ )
419 free( ppsz_dir_content[a] );
420 free( ppsz_dir_content );
425 for( j = 0; subdirs[j]; j++ )
431 free( f_fname_trim );
432 free( f_fname_noext );
437 qsort( result, i_sub_count, sizeof(vlc_subfn_t), compare_sub_priority );
439 result2 = calloc( i_sub_count + 1, sizeof(char*) );
441 for( j = 0, i_result2 = 0; j < i_sub_count && result2 != NULL; j++ )
443 bool b_reject = false;
445 if( !result[j].psz_fname || !result[j].psz_ext ) /* memory out */
448 if( !strcasecmp( result[j].psz_ext, "sub" ) )
451 for( i = 0; i < i_sub_count; i++ )
453 if( result[i].psz_fname && result[i].psz_ext &&
454 !strncasecmp( result[j].psz_fname, result[i].psz_fname,
455 strlen( result[j].psz_fname) - 3 ) &&
456 !strcasecmp( result[i].psz_ext, "idx" ) )
459 if( i < i_sub_count )
462 else if( !strcasecmp( result[j].psz_ext, "cdg" ) )
464 if( result[j].priority < SUB_PRIORITY_MATCH_ALL )
470 result2[i_result2++] = strdup( result[j].psz_fname );
473 for( j = 0; j < i_sub_count; j++ )
475 free( result[j].psz_fname );
476 free( result[j].psz_ext );