1 /*****************************************************************************
3 *****************************************************************************
4 * Copyright (C) 2003-2004 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., 59 Temple Place - Suite 330, Boston, MA 02111, USA.
23 *****************************************************************************/
27 * This file contains functions to dectect subtitle files.
32 #include <vlc/input.h>
46 * What's between a directory and a filename?
49 #define DIRECTORY_SEPARATOR '\\'
51 #define DIRECTORY_SEPARATOR '/'
55 * We are not going to autodetect more subtitle files than this.
57 #define MAX_SUBTITLE_FILES 128
61 * The possible extentions for subtitle files we support
63 static const char * sub_exts[] = { "utf", "utf8", "utf-8", "sub", "srt", "smi", "txt", "ssa", "idx", NULL};
64 /* extensions from unsupported types */
65 /* rt, aqt, jss, js, ass */
67 static void strcpy_trim( char *d, char *s )
69 /* skip leading whitespace */
70 while( *s && !isalnum(*s) )
77 while( *s && isalnum(*s) )
83 /* trim excess whitespace */
84 while( *s && !isalnum(*s) )
94 static void strcpy_strip_ext( char *d, char *s )
96 char *tmp = strrchr(s, '.');
104 strncpy(d, s, tmp - s);
114 static void strcpy_get_ext( char *d, char *s )
116 char *tmp = strrchr(s, '.');
121 } else strcpy( d, tmp + 1 );
124 static int whiteonly( char *s )
128 if( isalnum( *s ) ) return 0;
134 typedef struct _subfn
141 static int compare_sub_priority( const void *a, const void *b )
143 if (((subfn*)a)->priority > ((subfn*)b)->priority)
148 if (((subfn*)a)->priority < ((subfn*)b)->priority)
154 return strcoll(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
156 return strcmp(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
160 /* Utility function for scandir */
161 static int Filter( const struct dirent *p_dir_content )
166 if( p_dir_content == NULL || p_dir_content->d_name == NULL ) return VLC_FALSE;
167 /* does it end with a subtitle extension? */
168 tmp = strrchr( p_dir_content->d_name, '.');
175 for( i = 0; sub_exts[i]; i++ )
177 if( strcmp( sub_exts[i], tmp+1 ) == 0 )
188 * Convert a list of paths separated by ',' to a char**
190 static char **paths_to_list( char *psz_dir, char *psz_path )
192 unsigned int i, k, i_nb_subdirs;
193 char **subdirs; /* list of subdirectories to look in */
195 if( !psz_dir ) return NULL;
196 if( !psz_path ) return NULL;
199 for( k = 0; k < strlen( psz_path ); k++ )
201 if( psz_path[k] == ',' )
207 if( i_nb_subdirs > 0 )
209 char *psz_parser = NULL, *psz_temp = NULL;
211 subdirs = (char**)malloc( sizeof(char*) * ( i_nb_subdirs + 1 ) );
212 memset( subdirs, 0, sizeof(char*) * ( i_nb_subdirs + 1 ) );
214 psz_parser = psz_path;
215 while( psz_parser && *psz_parser )
218 psz_subdir = psz_parser;
219 psz_parser = strchr( psz_subdir, ',' );
224 while( *psz_parser == ' ' )
229 if( strlen( psz_subdir ) > 0 )
231 psz_temp = (char *)malloc( strlen(psz_dir)
232 + strlen(psz_subdir) + 2 );
235 sprintf( psz_temp, "%s%s%c",
236 psz_subdir[0] == '.' ? psz_dir : "",
238 psz_subdir[strlen(psz_subdir) - 1] ==
239 DIRECTORY_SEPARATOR ? '\0' : DIRECTORY_SEPARATOR );
240 subdirs[i] = psz_temp;
256 * Detect subtitle files.
258 * When called this function will split up the psz_name string into a
259 * directory, filename and extension. It then opens the directory
260 * in which the file resides and tries to find possible matches of
264 * \param p_this the calling \ref input_thread_t
265 * \param psz_path a list of subdirectories (separated by a ',') to look in.
266 * \param psz_name the complete filename to base the search on.
267 * \return a NULL terminated array of filenames with detected possible subtitles.
268 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
270 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
274 int j, i_result2, i_dir_content, i_sub_count = 0, i_fname_len = 0;
275 char *f_dir = NULL, *f_fname = NULL, *f_fname_noext = NULL, *f_fname_trim = NULL;
278 char tmp_fname_noext[PATH_MAX];
279 char tmp_fname_trim[PATH_MAX];
280 char tmp_fname_ext[PATH_MAX];
282 struct dirent **pp_dir_content;
283 char **tmp_subdirs, **subdirs; /* list of subdirectories to look in */
285 subfn *result = NULL; /* unsorted results */
286 char **result2; /* sorted results */
288 char *psz_fname_original = strdup( psz_name );
289 char *psz_fname = psz_fname_original;
291 if( !strncmp( psz_fname, "file://", 7 ) )
296 /* extract filename & dirname from psz_fname */
297 tmp = strrchr( psz_fname, DIRECTORY_SEPARATOR );
302 f_fname = malloc( strlen(tmp) );
304 strcpy( f_fname, tmp+1 ); // we skip the seperator, so it will still fit in the allocated space
305 dirlen = strlen(psz_fname) - strlen(tmp) + 1; // add the seperator
306 f_dir = malloc( dirlen + 1 );
309 strncpy( f_dir, psz_fname, dirlen );
315 f_fname = strdup( psz_fname );
318 i_fname_len = strlen( f_fname );
319 f_fname_noext = malloc(i_fname_len + 1);
320 f_fname_trim = malloc(i_fname_len + 1 );
322 strcpy_strip_ext( f_fname_noext, f_fname );
323 strcpy_trim( f_fname_trim, f_fname_noext );
325 result = (subfn*)malloc( sizeof(subfn) * MAX_SUBTITLE_FILES );
327 memset( result, 0, sizeof(subfn) * MAX_SUBTITLE_FILES );
329 var_Get( p_this, "sub-autodetect-fuzzy", &fuzzy );
331 tmp_subdirs = paths_to_list( f_dir, psz_path );
332 subdirs = tmp_subdirs;
334 for( j = -1; (j == -1) || ( (j >= 0) && (subdirs != NULL) && (*subdirs != NULL) );
337 pp_dir_content = NULL;
340 /* parse psz_src dir */
341 if( ( i_dir_content = scandir( j < 0 ? f_dir : *subdirs, &pp_dir_content, Filter,
346 msg_Dbg( p_this, "looking for a subtitle file in %s", j < 0 ? f_dir : *subdirs );
347 for( a = 0; a < i_dir_content; a++ )
350 struct dirent *p_dir_content = pp_dir_content[a];
351 char *psz_inUTF8 = FromLocale( p_dir_content->d_name );
352 char *p_fixed_name = vlc_fix_readdir_charset( p_this, psz_inUTF8 );
354 LocaleFree( psz_inUTF8 );
356 /* retrieve various parts of the filename */
357 strcpy_strip_ext( tmp_fname_noext, p_fixed_name );
358 strcpy_get_ext( tmp_fname_ext, p_fixed_name );
359 strcpy_trim( tmp_fname_trim, tmp_fname_noext );
361 if( !i_prio && !strcmp( tmp_fname_trim, f_fname_trim ) )
363 /* matches the movie name exactly */
367 ( tmp = strstr( tmp_fname_trim, f_fname_trim ) ) )
369 /* contains the movie name */
370 tmp += strlen( f_fname_trim );
371 if( whiteonly( tmp ) )
373 /* chars in front of the movie name */
378 /* chars after (and possibly in front of)
385 /* doesn't contain the movie name */
386 if( j == 0 ) i_prio = 1;
388 if( i_prio >= fuzzy.i_int )
393 asprintf( &tmpresult, "%s%s", j < 0 ? f_dir : *subdirs, p_fixed_name );
394 msg_Dbg( p_this, "autodetected subtitle: %s with priority %d", p_fixed_name, i_prio );
395 if( ( f = fopen( tmpresult, "rt" ) ) )
398 result[i_sub_count].priority = i_prio;
399 result[i_sub_count].psz_fname = tmpresult;
400 result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
402 } else free( tmpresult );
404 if( i_sub_count >= MAX_SUBTITLE_FILES ) break;
405 free( p_fixed_name );
408 if( j >= 0 ) free( *subdirs++ );
411 if( tmp_subdirs ) free( tmp_subdirs );
412 if( f_fname_trim ) free( f_fname_trim );
413 if( f_fname_noext ) free( f_fname_noext );
414 if( f_fname ) free( f_fname );
415 if( f_dir ) free( f_dir );
417 qsort( result, i_sub_count, sizeof( subfn ), compare_sub_priority );
419 result2 = (char**)malloc( sizeof(char*) * ( i_sub_count + 1 ) );
421 memset( result2, 0, sizeof(char*) * ( i_sub_count + 1 ) );
424 for( j = 0; j < i_sub_count; j++ )
426 if( result[j].psz_ext && !strcasecmp( result[j].psz_ext, "sub" ) )
429 for( i = 0; i < i_sub_count; i++ )
431 if( result[i].psz_fname && result[j].psz_fname &&
432 !strncasecmp( result[j].psz_fname, result[i].psz_fname, sizeof( result[j].psz_fname) - 4 ) &&
433 !strcasecmp( result[i].psz_ext, "idx" ) )
436 if( i >= i_sub_count )
438 result2[i_result2] = result[j].psz_fname;
444 result2[i_result2] = result[j].psz_fname;
449 if( psz_fname_original ) free( psz_fname_original );
450 if( result ) free( result );