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.
31 #include <vlc_input.h>
32 #include <vlc_charset.h>
48 #include "input_internal.h"
51 * We are not going to autodetect more subtitle files than this.
53 #define MAX_SUBTITLE_FILES 128
57 * The possible extensions for subtitle files we support
59 static const char const sub_exts[][6] = {
60 "utf", "utf8", "utf-8",
69 /* extensions from unsupported types */
70 /* rt, aqt, jss, js, ass */
72 static void strcpy_trim( char *d, char *s )
74 /* skip leading whitespace */
75 while( *s && !isalnum(*s) )
82 while( *s && isalnum(*s) )
88 /* trim excess whitespace */
89 while( *s && !isalnum(*s) )
99 static void strcpy_strip_ext( char *d, char *s )
101 char *tmp = strrchr(s, '.');
108 strlcpy(d, s, tmp - s + 1 );
116 static void strcpy_get_ext( char *d, char *s )
118 char *tmp = strrchr(s, '.');
123 } else strcpy( d, tmp + 1 );
126 static int whiteonly( char *s )
130 if( isalnum( *s ) ) return 0;
136 typedef struct _subfn
143 static int compare_sub_priority( const void *a, const void *b )
145 if (((subfn*)a)->priority > ((subfn*)b)->priority)
150 if (((subfn*)a)->priority < ((subfn*)b)->priority)
156 return strcoll(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
158 return strcmp(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
163 * Check if a file ends with a subtitle extension
165 int subtitles_Filter( const char *psz_dir_content )
167 const char *tmp = strrchr( psz_dir_content, '.');
175 for( i = 0; sub_exts[i][0]; i++ )
176 if( strcasecmp( sub_exts[i], tmp ) == 0 )
184 * Convert a list of paths separated by ',' to a char**
186 static char **paths_to_list( const char *psz_dir, char *psz_path )
188 unsigned int i, k, i_nb_subdirs;
189 char **subdirs; /* list of subdirectories to look in */
191 if( !psz_dir ) return NULL;
192 if( !psz_path ) return NULL;
195 for( k = 0; k < strlen( psz_path ); k++ )
197 if( psz_path[k] == ',' )
203 if( i_nb_subdirs > 0 )
205 char *psz_parser = NULL, *psz_temp = NULL;
207 subdirs = (char**)malloc( sizeof(char*) * ( i_nb_subdirs + 1 ) );
208 memset( subdirs, 0, sizeof(char*) * ( i_nb_subdirs + 1 ) );
210 psz_parser = psz_path;
211 while( psz_parser && *psz_parser )
214 psz_subdir = psz_parser;
215 psz_parser = strchr( psz_subdir, ',' );
220 while( *psz_parser == ' ' )
225 if( strlen( psz_subdir ) > 0 )
227 psz_temp = (char *)malloc( strlen(psz_dir)
228 + strlen(psz_subdir) + 2 );
231 sprintf( psz_temp, "%s%s%c",
232 psz_subdir[0] == '.' ? psz_dir : "",
234 psz_subdir[strlen(psz_subdir) - 1] ==
235 DIR_SEP_CHAR ? '\0' : DIR_SEP_CHAR );
236 subdirs[i] = psz_temp;
252 * Detect subtitle files.
254 * When called this function will split up the psz_name string into a
255 * directory, filename and extension. It then opens the directory
256 * in which the file resides and tries to find possible matches of
260 * \param p_this the calling \ref input_thread_t
261 * \param psz_path a list of subdirectories (separated by a ',') to look in.
262 * \param psz_name the complete filename to base the search on.
263 * \return a NULL terminated array of filenames with detected possible subtitles.
264 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
266 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
267 const char *psz_name )
270 int j, i_result2, i_sub_count = 0, i_fname_len = 0;
271 char *f_dir = NULL, *f_fname = NULL, *f_fname_noext = NULL, *f_fname_trim = NULL;
274 char **tmp_subdirs, **subdirs; /* list of subdirectories to look in */
276 subfn *result = NULL; /* unsorted results */
277 char **result2; /* sorted results */
279 char *psz_fname_original = strdup( psz_name );
280 char *psz_fname = psz_fname_original;
282 if( psz_fname == NULL ) return NULL;
284 if( !strncmp( psz_fname, "file://", 7 ) )
289 /* extract filename & dirname from psz_fname */
290 tmp = strrchr( psz_fname, DIR_SEP_CHAR );
295 f_fname = malloc( strlen(tmp) );
297 strcpy( f_fname, tmp+1 ); // we skip the separator, so it will still fit in the allocated space
298 dirlen = strlen(psz_fname) - strlen(tmp) + 2; // add the separator
299 f_dir = malloc( dirlen + 1 );
301 strlcpy( f_dir, psz_fname, dirlen );
305 /* Get the current working directory */
308 f_dir = getcwd( NULL, 0 );
312 if( psz_fname_original ) free( psz_fname_original );
315 dirlen = strlen( f_dir );
316 f_dir = (char *)realloc(f_dir, dirlen +2 );
317 f_dir[dirlen] = DIR_SEP_CHAR;
318 f_dir[dirlen+1] = '\0';
319 f_fname = strdup( psz_fname );
322 i_fname_len = strlen( f_fname );
323 f_fname_noext = malloc(i_fname_len + 1);
324 f_fname_trim = malloc(i_fname_len + 1 );
326 strcpy_strip_ext( f_fname_noext, f_fname );
327 strcpy_trim( f_fname_trim, f_fname_noext );
329 result = (subfn*)malloc( sizeof(subfn) * MAX_SUBTITLE_FILES );
331 memset( result, 0, sizeof(subfn) * MAX_SUBTITLE_FILES );
333 var_Get( p_this, "sub-autodetect-fuzzy", &fuzzy );
335 tmp_subdirs = paths_to_list( f_dir, psz_path );
336 subdirs = tmp_subdirs;
338 for( j = -1; (j == -1) || ( (j >= 0) && (subdirs != NULL) &&
339 (*subdirs != NULL) ); j++)
341 const char *psz_dir = j < 0 ? f_dir : *subdirs;
342 char **ppsz_dir_content;
345 if( psz_dir == NULL )
348 /* parse psz_src dir */
349 i_dir_content = utf8_scandir( psz_dir, &ppsz_dir_content,
350 subtitles_Filter, NULL );
352 if( i_dir_content != -1 )
356 msg_Dbg( p_this, "looking for a subtitle file in %s", psz_dir );
357 for( a = 0; a < i_dir_content; a++ )
359 char *psz_name = 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 )
403 char psz_path[strlen( psz_dir ) + strlen( psz_name ) + 1];
406 sprintf( psz_path, "%s%s", psz_dir, psz_name );
407 if( !strcmp( psz_path, psz_fname ) )
411 "autodetected subtitle: %s with priority %d",
414 if( !utf8_stat( psz_path, &st ) && S_ISREG( st.st_mode ) )
417 "autodetected subtitle: %s with priority %d",
419 result[i_sub_count].priority = i_prio;
420 result[i_sub_count].psz_fname = strdup( psz_path );
421 result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
426 msg_Dbg( p_this, "stat failed" );
429 if( i_sub_count >= MAX_SUBTITLE_FILES ) break;
431 for( a = 0; a < i_dir_content; a++ )
432 free( ppsz_dir_content[a] );
433 if( ppsz_dir_content ) free( ppsz_dir_content );
435 if( j >= 0 ) if( *subdirs ) free( *subdirs++ );
438 if( tmp_subdirs ) free( tmp_subdirs );
439 if( f_fname_trim ) free( f_fname_trim );
440 if( f_fname_noext ) free( f_fname_noext );
441 if( f_fname ) free( f_fname );
442 if( f_dir ) free( f_dir );
444 qsort( result, i_sub_count, sizeof( subfn ), compare_sub_priority );
446 result2 = (char**)malloc( sizeof(char*) * ( i_sub_count + 1 ) );
448 memset( result2, 0, sizeof(char*) * ( i_sub_count + 1 ) );
451 for( j = 0; j < i_sub_count; j++ )
453 if( result[j].psz_ext && !strcasecmp( result[j].psz_ext, "sub" ) )
456 for( i = 0; i < i_sub_count; i++ )
458 if( result[i].psz_fname && result[j].psz_fname &&
459 !strncasecmp( result[j].psz_fname, result[i].psz_fname,
460 sizeof( result[j].psz_fname) - 4 ) &&
461 !strcasecmp( result[i].psz_ext, "idx" ) )
464 if( i >= i_sub_count )
466 result2[i_result2] = result[j].psz_fname;
470 else if( result[j].psz_ext && !strcasecmp( result[j].psz_ext, "cdg" ) )
472 if( result[j].priority == 4 )
474 result2[i_result2] = result[j].psz_fname;
480 result2[i_result2] = result[j].psz_fname;
485 if( psz_fname_original ) free( psz_fname_original );
486 if( result ) free( result );