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 * sub_exts[] = { "utf", "utf8", "utf-8", "sub", "srt", "smi", "txt", "ssa", "idx", NULL };
60 /* extensions from unsupported types */
61 /* rt, aqt, jss, js, ass */
63 static void strcpy_trim( char *d, char *s )
65 /* skip leading whitespace */
66 while( *s && !isalnum(*s) )
73 while( *s && isalnum(*s) )
79 /* trim excess whitespace */
80 while( *s && !isalnum(*s) )
90 static void strcpy_strip_ext( char *d, char *s )
92 char *tmp = strrchr(s, '.');
99 strlcpy(d, s, tmp - s + 1 );
107 static void strcpy_get_ext( char *d, char *s )
109 char *tmp = strrchr(s, '.');
114 } else strcpy( d, tmp + 1 );
117 static int whiteonly( char *s )
121 if( isalnum( *s ) ) return 0;
127 typedef struct _subfn
134 static int compare_sub_priority( const void *a, const void *b )
136 if (((subfn*)a)->priority > ((subfn*)b)->priority)
141 if (((subfn*)a)->priority < ((subfn*)b)->priority)
147 return strcoll(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
149 return strcmp(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
154 * Check if a file ends with a subtitle extension
156 int subtitles_Filter( const char *psz_dir_content )
158 const char *tmp = strrchr( psz_dir_content, '.');
166 for( i = 0; sub_exts[i]; i++ )
167 if( strcasecmp( sub_exts[i], tmp ) == 0 )
175 * Convert a list of paths separated by ',' to a char**
177 static char **paths_to_list( const char *psz_dir, char *psz_path )
179 unsigned int i, k, i_nb_subdirs;
180 char **subdirs; /* list of subdirectories to look in */
182 if( !psz_dir ) return NULL;
183 if( !psz_path ) return NULL;
186 for( k = 0; k < strlen( psz_path ); k++ )
188 if( psz_path[k] == ',' )
194 if( i_nb_subdirs > 0 )
196 char *psz_parser = NULL, *psz_temp = NULL;
198 subdirs = (char**)malloc( sizeof(char*) * ( i_nb_subdirs + 1 ) );
199 memset( subdirs, 0, sizeof(char*) * ( i_nb_subdirs + 1 ) );
201 psz_parser = psz_path;
202 while( psz_parser && *psz_parser )
205 psz_subdir = psz_parser;
206 psz_parser = strchr( psz_subdir, ',' );
211 while( *psz_parser == ' ' )
216 if( strlen( psz_subdir ) > 0 )
218 psz_temp = (char *)malloc( strlen(psz_dir)
219 + strlen(psz_subdir) + 2 );
222 sprintf( psz_temp, "%s%s%c",
223 psz_subdir[0] == '.' ? psz_dir : "",
225 psz_subdir[strlen(psz_subdir) - 1] ==
226 DIR_SEP_CHAR ? '\0' : DIR_SEP_CHAR );
227 subdirs[i] = psz_temp;
243 * Detect subtitle files.
245 * When called this function will split up the psz_name string into a
246 * directory, filename and extension. It then opens the directory
247 * in which the file resides and tries to find possible matches of
251 * \param p_this the calling \ref input_thread_t
252 * \param psz_path a list of subdirectories (separated by a ',') to look in.
253 * \param psz_name the complete filename to base the search on.
254 * \return a NULL terminated array of filenames with detected possible subtitles.
255 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
257 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
258 const char *psz_name )
261 int j, i_result2, i_sub_count = 0, i_fname_len = 0;
262 char *f_dir = NULL, *f_fname = NULL, *f_fname_noext = NULL, *f_fname_trim = NULL;
265 char **tmp_subdirs, **subdirs; /* list of subdirectories to look in */
267 subfn *result = NULL; /* unsorted results */
268 char **result2; /* sorted results */
270 char *psz_fname_original = strdup( psz_name );
271 char *psz_fname = psz_fname_original;
273 if( psz_fname == NULL ) return NULL;
275 if( !strncmp( psz_fname, "file://", 7 ) )
280 /* extract filename & dirname from psz_fname */
281 tmp = strrchr( psz_fname, DIR_SEP_CHAR );
286 f_fname = malloc( strlen(tmp) );
288 strcpy( f_fname, tmp+1 ); // we skip the separator, so it will still fit in the allocated space
289 dirlen = strlen(psz_fname) - strlen(tmp) + 2; // add the separator
290 f_dir = malloc( dirlen + 1 );
292 strlcpy( f_dir, psz_fname, dirlen );
296 /* Get the current working directory */
299 f_dir = getcwd( NULL, 0 );
303 if( psz_fname_original ) free( psz_fname_original );
306 dirlen = strlen( f_dir );
307 f_dir = (char *)realloc(f_dir, dirlen +2 );
308 f_dir[dirlen] = DIR_SEP_CHAR;
309 f_dir[dirlen+1] = '\0';
310 f_fname = strdup( psz_fname );
313 i_fname_len = strlen( f_fname );
314 f_fname_noext = malloc(i_fname_len + 1);
315 f_fname_trim = malloc(i_fname_len + 1 );
317 strcpy_strip_ext( f_fname_noext, f_fname );
318 strcpy_trim( f_fname_trim, f_fname_noext );
320 result = (subfn*)malloc( sizeof(subfn) * MAX_SUBTITLE_FILES );
322 memset( result, 0, sizeof(subfn) * MAX_SUBTITLE_FILES );
324 var_Get( p_this, "sub-autodetect-fuzzy", &fuzzy );
326 tmp_subdirs = paths_to_list( f_dir, psz_path );
327 subdirs = tmp_subdirs;
329 for( j = -1; (j == -1) || ( (j >= 0) && (subdirs != NULL) &&
330 (*subdirs != NULL) ); j++)
332 const char *psz_dir = j < 0 ? f_dir : *subdirs;
333 char **ppsz_dir_content;
336 if( psz_dir == NULL )
339 /* parse psz_src dir */
340 i_dir_content = utf8_scandir( psz_dir, &ppsz_dir_content,
341 subtitles_Filter, NULL );
343 if( i_dir_content != -1 )
347 msg_Dbg( p_this, "looking for a subtitle file in %s", psz_dir );
348 for( a = 0; a < i_dir_content; a++ )
350 char *psz_name = ppsz_dir_content[a];
351 char tmp_fname_noext[strlen( psz_name ) + 1];
352 char tmp_fname_trim[strlen( psz_name ) + 1];
353 char tmp_fname_ext[strlen( psz_name ) + 1];
357 if( psz_name == NULL )
360 /* retrieve various parts of the filename */
361 strcpy_strip_ext( tmp_fname_noext, psz_name );
362 strcpy_get_ext( tmp_fname_ext, psz_name );
363 strcpy_trim( tmp_fname_trim, tmp_fname_noext );
365 if( !i_prio && !strcmp( tmp_fname_trim, f_fname_trim ) )
367 /* matches the movie name exactly */
371 ( tmp = strstr( tmp_fname_trim, f_fname_trim ) ) )
373 /* contains the movie name */
374 tmp += strlen( f_fname_trim );
375 if( whiteonly( tmp ) )
377 /* chars in front of the movie name */
382 /* chars after (and possibly in front of)
389 /* doesn't contain the movie name */
390 if( j == 0 ) i_prio = 1;
392 if( i_prio >= fuzzy.i_int )
394 char psz_path[strlen( psz_dir ) + strlen( psz_name ) + 1];
397 sprintf( psz_path, "%s%s", psz_dir, psz_name );
399 "autodetected subtitle: %s with priority %d",
402 if( !utf8_stat( psz_path, &st ) && S_ISREG( st.st_mode ) )
405 "autodetected subtitle: %s with priority %d",
407 result[i_sub_count].priority = i_prio;
408 result[i_sub_count].psz_fname = strdup( psz_path );
409 result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
414 msg_Dbg( p_this, "stat failed" );
417 if( i_sub_count >= MAX_SUBTITLE_FILES ) break;
419 for( a = 0; a < i_dir_content; a++ )
420 free( ppsz_dir_content[a] );
421 if( ppsz_dir_content ) free( ppsz_dir_content );
423 if( j >= 0 ) if( *subdirs ) free( *subdirs++ );
426 if( tmp_subdirs ) free( tmp_subdirs );
427 if( f_fname_trim ) free( f_fname_trim );
428 if( f_fname_noext ) free( f_fname_noext );
429 if( f_fname ) free( f_fname );
430 if( f_dir ) free( f_dir );
432 qsort( result, i_sub_count, sizeof( subfn ), compare_sub_priority );
434 result2 = (char**)malloc( sizeof(char*) * ( i_sub_count + 1 ) );
436 memset( result2, 0, sizeof(char*) * ( i_sub_count + 1 ) );
439 for( j = 0; j < i_sub_count; j++ )
441 if( result[j].psz_ext && !strcasecmp( result[j].psz_ext, "sub" ) )
444 for( i = 0; i < i_sub_count; i++ )
446 if( result[i].psz_fname && result[j].psz_fname &&
447 !strncasecmp( result[j].psz_fname, result[i].psz_fname,
448 sizeof( result[j].psz_fname) - 4 ) &&
449 !strcasecmp( result[i].psz_ext, "idx" ) )
452 if( i >= i_sub_count )
454 result2[i_result2] = result[j].psz_fname;
460 result2[i_result2] = result[j].psz_fname;
465 if( psz_fname_original ) free( psz_fname_original );
466 if( result ) free( result );