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.
33 #include <vlc/input.h>
43 * What's between a directory and a filename?
46 #define DIRECTORY_SEPARATOR '\\'
48 #define DIRECTORY_SEPARATOR '/'
52 * We are not going to autodetect more subtitle files than this.
54 #define MAX_SUBTITLE_FILES 128
58 * The possible extentions for subtitle files we support
60 static const char * sub_exts[] = { "utf", "utf8", "utf-8", "sub", "srt", "smi", "txt", "ssa", "idx", NULL};
61 /* extensions from unsupported types */
62 /* rt, aqt, jss, js, ass */
65 static void strcpy_trim( char *d, char *s )
67 /* skip leading whitespace */
68 while( *s && !isalnum(*s) )
75 while( *s && isalnum(*s) )
81 /* trim excess whitespace */
82 while( *s && !isalnum(*s) )
92 static void strcpy_strip_ext( char *d, char *s )
94 char *tmp = strrchr(s, '.');
101 strncpy(d, s, tmp - s);
111 static void strcpy_get_ext( char *d, char *s )
113 char *tmp = strrchr(s, '.');
118 } else strcpy( d, tmp + 1 );
121 static int whiteonly( char *s )
125 if( isalnum( *s ) ) return 0;
131 typedef struct _subfn
138 static int compare_sub_priority( const void *a, const void *b )
140 if (((subfn*)a)->priority > ((subfn*)b)->priority)
145 if (((subfn*)a)->priority < ((subfn*)b)->priority)
151 return strcoll(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
153 return strcmp(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
157 /* Utility function for scandir */
158 static int Filter( const struct dirent *p_dir_content )
163 if( p_dir_content == NULL || p_dir_content->d_name == NULL ) return VLC_FALSE;
164 /* does it end with a subtitle extension? */
165 tmp = strrchr( p_dir_content->d_name, '.');
172 for( i = 0; sub_exts[i]; i++ )
174 if( strcmp( sub_exts[i], tmp+1 ) == 0 )
185 * Convert a list of paths separated by ',' to a char**
187 static char **paths_to_list( char *psz_dir, char *psz_path )
189 unsigned int i, k, i_nb_subdirs;
190 char **subdirs; /* list of subdirectories to look in */
193 for( k = 0; k < strlen( psz_path ); k++ )
195 if( psz_path[k] == ',' )
201 if( i_nb_subdirs > 0 )
203 char *psz_parser, *psz_temp;
205 subdirs = (char**)malloc( sizeof(char*) * ( i_nb_subdirs + 1 ) );
206 memset( subdirs, 0, sizeof(char*) * ( i_nb_subdirs + 1 ) );
208 psz_parser = psz_path;
209 while( psz_parser && *psz_parser )
212 psz_subdir = psz_parser;
213 psz_parser = strchr( psz_subdir, ',' );
218 while( *psz_parser == ' ' )
223 if( strlen( psz_subdir ) > 0 )
225 psz_temp = (char *)malloc( strlen(psz_dir)
226 + strlen(psz_subdir) + 2 );
229 sprintf( psz_temp, "%s%s%c",
230 psz_subdir[0] == '.' ? psz_dir : "",
232 psz_subdir[strlen(psz_subdir) - 1] ==
233 DIRECTORY_SEPARATOR ? '\0' : DIRECTORY_SEPARATOR );
234 subdirs[i] = psz_temp;
250 * Detect subtitle files.
252 * When called this function will split up the psz_name string into a
253 * directory, filename and extension. It then opens the directory
254 * in which the file resides and tries to find possible matches of
258 * \param p_this the calling \ref input_thread_t
259 * \param psz_path a list of subdirectories (separated by a ',') to look in.
260 * \param psz_name the complete filename to base the search on.
261 * \return a NULL terminated array of filenames with detected possible subtitles.
262 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
264 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
268 int j, i_result2, i_dir_content, i_sub_count = 0, i_fname_len = 0;
269 char *f_dir, *f_fname, *f_fname_noext, *f_fname_trim;
272 char tmp_fname_noext[PATH_MAX];
273 char tmp_fname_trim[PATH_MAX];
274 char tmp_fname_ext[PATH_MAX];
276 struct dirent **pp_dir_content;
277 char **tmp_subdirs, **subdirs; /* list of subdirectories to look in */
279 subfn *result; /* unsorted results */
280 char **result2; /* sorted results */
282 char *psz_fname_original = strdup( psz_name );
283 char *psz_fname = psz_fname_original;
285 if( !strncmp( psz_fname, "file://", 7 ) )
290 /* extract filename & dirname from psz_fname */
291 tmp = strrchr( psz_fname, DIRECTORY_SEPARATOR );
295 f_fname = malloc( strlen(tmp) );
296 strcpy( f_fname, tmp+1 ); // we skip the seperator, so it will still fit in the allocated space
297 dirlen = strlen(psz_fname) - strlen(tmp);
298 f_dir = malloc( dirlen + 1 );
299 strncpy( f_dir, psz_fname, dirlen + 1 );
300 f_dir[dirlen + 1] = 0;
304 f_fname = strdup( psz_fname );
307 i_fname_len = strlen( f_fname );
308 f_fname_noext = malloc(i_fname_len + 1);
309 f_fname_trim = malloc(i_fname_len + 1 );
311 strcpy_strip_ext( f_fname_noext, f_fname );
312 strcpy_trim( f_fname_trim, f_fname_noext );
314 result = (subfn*)malloc( sizeof(subfn) * MAX_SUBTITLE_FILES );
315 memset( result, 0, sizeof(subfn) * MAX_SUBTITLE_FILES );
317 var_Get( p_this, "sub-autodetect-fuzzy", &fuzzy );
319 tmp_subdirs = paths_to_list( f_dir, psz_path );
320 subdirs = tmp_subdirs;
322 for( j = -1; j == -1 || ( j >= 0 && subdirs != NULL && *subdirs != NULL );
325 pp_dir_content = NULL;
328 /* parse psz_src dir */
329 if( ( i_dir_content = scandir( j < 0 ? f_dir : *subdirs, &pp_dir_content, Filter,
333 msg_Dbg( p_this, "looking for a subtitle file in %s", j < 0 ? f_dir : *subdirs );
335 for( a = 0; a < i_dir_content; a++ )
338 struct dirent *p_dir_content = pp_dir_content[a];
339 char *psz_inUTF8 = FromLocale( p_dir_content->d_name );
340 char *p_fixed_name = vlc_fix_readdir_charset( p_this, psz_inUTF8 );
342 LocaleFree( psz_inUTF8 );
344 /* retrieve various parts of the filename */
345 strcpy_strip_ext( tmp_fname_noext, p_fixed_name );
346 strcpy_get_ext( tmp_fname_ext, p_fixed_name );
347 strcpy_trim( tmp_fname_trim, tmp_fname_noext );
349 if( !i_prio && !strcmp( tmp_fname_trim, f_fname_trim ) )
351 /* matches the movie name exactly */
355 ( tmp = strstr( tmp_fname_trim, f_fname_trim ) ) )
357 /* contains the movie name */
358 tmp += strlen( f_fname_trim );
359 if( whiteonly( tmp ) )
361 /* chars in front of the movie name */
366 /* chars after (and possibly in front of)
373 /* doesn't contain the movie name */
374 if( j == 0 ) i_prio = 1;
376 if( i_prio >= fuzzy.i_int )
381 asprintf( &tmpresult, "%s%s", j < 0 ? f_dir : *subdirs, p_fixed_name );
382 msg_Dbg( p_this, "autodetected subtitle: %s with priority %d", p_fixed_name, i_prio );
383 if( ( f = fopen( tmpresult, "rt" ) ) )
386 result[i_sub_count].priority = i_prio;
387 result[i_sub_count].psz_fname = tmpresult;
388 result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
390 } else free( tmpresult );
392 if( i_sub_count >= MAX_SUBTITLE_FILES ) break;
393 free( p_fixed_name );
396 if( j >= 0 ) free( *subdirs++ );
399 if(tmp_subdirs) free(tmp_subdirs);
400 if(f_fname_trim) free(f_fname_trim);
401 if(f_fname_noext) free(f_fname_noext);
402 if(f_fname) free(f_fname);
403 if(f_dir) free(f_dir);
405 qsort( result, i_sub_count, sizeof( subfn ), compare_sub_priority );
407 result2 = (char**)malloc( sizeof(char*) * ( i_sub_count + 1 ) );
408 memset( result2, 0, sizeof(char*) * ( i_sub_count + 1 ) );
411 for( j = 0; j < i_sub_count; j++ )
413 if( result[j].psz_ext && !strcasecmp( result[j].psz_ext, "sub" ) )
416 for( i = 0; i < i_sub_count; i++ )
418 if( result[i].psz_fname && result[j].psz_fname &&
419 !strncasecmp( result[j].psz_fname, result[i].psz_fname, sizeof( result[j].psz_fname) - 4 ) &&
420 !strcasecmp( result[i].psz_ext, "idx" ) )
424 if( i >= i_sub_count )
426 result2[i_result2] = result[j].psz_fname;
432 result2[i_result2] = result[j].psz_fname;
437 free( psz_fname_original );