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>
47 * What's between a directory and a filename?
50 #define DIRECTORY_SEPARATOR '\\'
52 #define DIRECTORY_SEPARATOR '/'
56 * We are not going to autodetect more subtitle files than this.
58 #define MAX_SUBTITLE_FILES 128
62 * The possible extentions for subtitle files we support
64 static const char * sub_exts[] = { "utf", "utf8", "utf-8", "sub", "srt", "smi", "txt", "ssa", "idx", NULL};
65 /* extensions from unsupported types */
66 /* rt, aqt, jss, js, ass */
69 static void strcpy_trim( char *d, char *s )
71 /* skip leading whitespace */
72 while( *s && !isalnum(*s) )
79 while( *s && isalnum(*s) )
85 /* trim excess whitespace */
86 while( *s && !isalnum(*s) )
96 static void strcpy_strip_ext( char *d, char *s )
98 char *tmp = strrchr(s, '.');
105 strncpy(d, s, tmp - s);
115 static void strcpy_get_ext( char *d, char *s )
117 char *tmp = strrchr(s, '.');
122 } else strcpy( d, tmp + 1 );
125 static int whiteonly( char *s )
129 if( isalnum( *s ) ) return 0;
135 typedef struct _subfn
142 static int compare_sub_priority( const void *a, const void *b )
144 if (((subfn*)a)->priority > ((subfn*)b)->priority)
149 if (((subfn*)a)->priority < ((subfn*)b)->priority)
155 return strcoll(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
157 return strcmp(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
161 /* Utility function for scandir */
162 static int Filter( const struct dirent *p_dir_content )
167 if( p_dir_content == NULL || p_dir_content->d_name == NULL ) return VLC_FALSE;
168 /* does it end with a subtitle extension? */
169 tmp = strrchr( p_dir_content->d_name, '.');
176 for( i = 0; sub_exts[i]; i++ )
178 if( strcmp( sub_exts[i], tmp+1 ) == 0 )
189 * Convert a list of paths separated by ',' to a char**
191 static char **paths_to_list( char *psz_dir, char *psz_path )
193 unsigned int i, k, i_nb_subdirs;
194 char **subdirs; /* list of subdirectories to look in */
197 for( k = 0; k < strlen( psz_path ); k++ )
199 if( psz_path[k] == ',' )
205 if( i_nb_subdirs > 0 )
207 char *psz_parser, *psz_temp;
209 subdirs = (char**)malloc( sizeof(char*) * ( i_nb_subdirs + 1 ) );
210 memset( subdirs, 0, sizeof(char*) * ( i_nb_subdirs + 1 ) );
212 psz_parser = psz_path;
213 while( psz_parser && *psz_parser )
216 psz_subdir = psz_parser;
217 psz_parser = strchr( psz_subdir, ',' );
222 while( *psz_parser == ' ' )
227 if( strlen( psz_subdir ) > 0 )
229 psz_temp = (char *)malloc( strlen(psz_dir)
230 + strlen(psz_subdir) + 2 );
233 sprintf( psz_temp, "%s%s%c",
234 psz_subdir[0] == '.' ? psz_dir : "",
236 psz_subdir[strlen(psz_subdir) - 1] ==
237 DIRECTORY_SEPARATOR ? '\0' : DIRECTORY_SEPARATOR );
238 subdirs[i] = psz_temp;
254 * Detect subtitle files.
256 * When called this function will split up the psz_name string into a
257 * directory, filename and extension. It then opens the directory
258 * in which the file resides and tries to find possible matches of
262 * \param p_this the calling \ref input_thread_t
263 * \param psz_path a list of subdirectories (separated by a ',') to look in.
264 * \param psz_name the complete filename to base the search on.
265 * \return a NULL terminated array of filenames with detected possible subtitles.
266 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
268 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
272 int j, i_result2, i_dir_content, i_sub_count = 0, i_fname_len = 0;
273 char *f_dir, *f_fname, *f_fname_noext, *f_fname_trim;
276 char tmp_fname_noext[PATH_MAX];
277 char tmp_fname_trim[PATH_MAX];
278 char tmp_fname_ext[PATH_MAX];
280 struct dirent **pp_dir_content;
281 char **tmp_subdirs, **subdirs; /* list of subdirectories to look in */
283 subfn *result; /* unsorted results */
284 char **result2; /* sorted results */
286 char *psz_fname_original = strdup( psz_name );
287 char *psz_fname = psz_fname_original;
289 if( !strncmp( psz_fname, "file://", 7 ) )
294 /* extract filename & dirname from psz_fname */
295 tmp = strrchr( psz_fname, DIRECTORY_SEPARATOR );
299 f_fname = malloc( strlen(tmp) );
300 strcpy( f_fname, tmp+1 ); // we skip the seperator, so it will still fit in the allocated space
301 dirlen = strlen(psz_fname) - strlen(tmp);
302 f_dir = malloc( dirlen + 1 );
303 strncpy( f_dir, psz_fname, dirlen + 1 );
304 f_dir[dirlen + 1] = 0;
308 f_fname = strdup( psz_fname );
311 i_fname_len = strlen( f_fname );
312 f_fname_noext = malloc(i_fname_len + 1);
313 f_fname_trim = malloc(i_fname_len + 1 );
315 strcpy_strip_ext( f_fname_noext, f_fname );
316 strcpy_trim( f_fname_trim, f_fname_noext );
318 result = (subfn*)malloc( sizeof(subfn) * MAX_SUBTITLE_FILES );
319 memset( result, 0, sizeof(subfn) * MAX_SUBTITLE_FILES );
321 var_Get( p_this, "sub-autodetect-fuzzy", &fuzzy );
323 tmp_subdirs = paths_to_list( f_dir, psz_path );
324 subdirs = tmp_subdirs;
326 for( j = -1; j == -1 || ( j >= 0 && subdirs != NULL && *subdirs != NULL );
329 pp_dir_content = NULL;
332 /* parse psz_src dir */
333 if( ( i_dir_content = scandir( j < 0 ? f_dir : *subdirs, &pp_dir_content, Filter,
337 msg_Dbg( p_this, "looking for a subtitle file in %s", j < 0 ? f_dir : *subdirs );
339 for( a = 0; a < i_dir_content; a++ )
342 struct dirent *p_dir_content = pp_dir_content[a];
343 char *psz_inUTF8 = FromLocale( p_dir_content->d_name );
344 char *p_fixed_name = vlc_fix_readdir_charset( p_this, psz_inUTF8 );
346 LocaleFree( psz_inUTF8 );
348 /* retrieve various parts of the filename */
349 strcpy_strip_ext( tmp_fname_noext, p_fixed_name );
350 strcpy_get_ext( tmp_fname_ext, p_fixed_name );
351 strcpy_trim( tmp_fname_trim, tmp_fname_noext );
353 if( !i_prio && !strcmp( tmp_fname_trim, f_fname_trim ) )
355 /* matches the movie name exactly */
359 ( tmp = strstr( tmp_fname_trim, f_fname_trim ) ) )
361 /* contains the movie name */
362 tmp += strlen( f_fname_trim );
363 if( whiteonly( tmp ) )
365 /* chars in front of the movie name */
370 /* chars after (and possibly in front of)
377 /* doesn't contain the movie name */
378 if( j == 0 ) i_prio = 1;
380 if( i_prio >= fuzzy.i_int )
385 asprintf( &tmpresult, "%s%s", j < 0 ? f_dir : *subdirs, p_fixed_name );
386 msg_Dbg( p_this, "autodetected subtitle: %s with priority %d", p_fixed_name, i_prio );
387 if( ( f = fopen( tmpresult, "rt" ) ) )
390 result[i_sub_count].priority = i_prio;
391 result[i_sub_count].psz_fname = tmpresult;
392 result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
394 } else free( tmpresult );
396 if( i_sub_count >= MAX_SUBTITLE_FILES ) break;
397 free( p_fixed_name );
400 if( j >= 0 ) free( *subdirs++ );
403 if(tmp_subdirs) free(tmp_subdirs);
404 if(f_fname_trim) free(f_fname_trim);
405 if(f_fname_noext) free(f_fname_noext);
406 if(f_fname) free(f_fname);
407 if(f_dir) free(f_dir);
409 qsort( result, i_sub_count, sizeof( subfn ), compare_sub_priority );
411 result2 = (char**)malloc( sizeof(char*) * ( i_sub_count + 1 ) );
412 memset( result2, 0, sizeof(char*) * ( i_sub_count + 1 ) );
415 for( j = 0; j < i_sub_count; j++ )
417 if( result[j].psz_ext && !strcasecmp( result[j].psz_ext, "sub" ) )
420 for( i = 0; i < i_sub_count; i++ )
422 if( result[i].psz_fname && result[j].psz_fname &&
423 !strncasecmp( result[j].psz_fname, result[i].psz_fname, sizeof( result[j].psz_fname) - 4 ) &&
424 !strcasecmp( result[i].psz_ext, "idx" ) )
428 if( i >= i_sub_count )
430 result2[i_result2] = result[j].psz_fname;
436 result2[i_result2] = result[j].psz_fname;
441 free( psz_fname_original );