1 /*****************************************************************************
3 *****************************************************************************
4 * Copyright (C) 2003-2004 VideoLAN (Centrale Réseaux) and its contributors
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>
41 * What's between a directory and a filename?
44 #define DIRECTORY_SEPARATOR '\\'
46 #define DIRECTORY_SEPARATOR '/'
50 * We are not going to autodetect more subtitle files than this.
52 #define MAX_SUBTITLE_FILES 128
56 * The possible extentions for subtitle files we support
58 static const char * sub_exts[] = { "utf", "utf8", "utf-8", "sub", "srt", "smi", "txt", "ssa", "idx", NULL};
59 /* extensions from unsupported types */
60 /* rt, aqt, jss, js, ass */
62 static void strcpy_trim( char *d, char *s )
64 /* skip leading whitespace */
65 while( *s && !isalnum(*s) )
72 while( *s && isalnum(*s) )
78 /* trim excess whitespace */
79 while( *s && !isalnum(*s) )
89 static void strcpy_strip_ext( char *d, char *s )
91 char *tmp = strrchr(s, '.');
98 strncpy(d, s, tmp - s);
108 static void strcpy_get_ext( char *d, char *s )
110 char *tmp = strrchr(s, '.');
115 } else strcpy( d, tmp + 1 );
118 static int whiteonly( char *s )
122 if( isalnum( *s ) ) return 0;
128 typedef struct _subfn
135 static int compare_sub_priority( const void *a, const void *b )
137 if (((subfn*)a)->priority > ((subfn*)b)->priority)
142 if (((subfn*)a)->priority < ((subfn*)b)->priority)
148 return strcoll(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
150 return strcmp(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
155 * Convert a list of paths separated by ',' to a char**
157 static char **paths_to_list( char *psz_dir, char *psz_path )
159 unsigned int i, k, i_nb_subdirs;
160 char **subdirs; /* list of subdirectories to look in */
163 for( k = 0; k < strlen( psz_path ); k++ )
165 if( psz_path[k] == ',' )
171 if( i_nb_subdirs > 0 )
173 char *psz_parser, *psz_temp;
175 subdirs = (char**)malloc( sizeof(char*) * ( i_nb_subdirs + 1 ) );
176 memset( subdirs, 0, sizeof(char*) * ( i_nb_subdirs + 1 ) );
178 psz_parser = psz_path;
179 while( psz_parser && *psz_parser )
182 psz_subdir = psz_parser;
183 psz_parser = strchr( psz_subdir, ',' );
188 while( *psz_parser == ' ' )
193 if( strlen( psz_subdir ) > 0 )
195 psz_temp = (char *)malloc( strlen(psz_dir)
196 + strlen(psz_subdir) + 2 );
199 sprintf( psz_temp, "%s%s%c",
200 psz_subdir[0] == '.' ? psz_dir : "",
202 psz_subdir[strlen(psz_subdir) - 1] ==
203 DIRECTORY_SEPARATOR ? '\0' : DIRECTORY_SEPARATOR );
204 subdirs[i] = psz_temp;
219 * Detect subtitle files.
221 * When called this function will split up the psz_fname string into a
222 * directory, filename and extension. It then opens the directory
223 * in which the file resides and tries to find possible matches of
227 * \param p_this the calling \ref input_thread_t
228 * \param psz_path a list of subdirectories (separated by a ',') to look in.
229 * \param psz_fname the complete filename to base the search on.
230 * \return a NULL terminated array of filenames with detected possible subtitles.
231 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
233 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
236 /* variables to be used for derivatives of psz_fname */
237 char *f_dir, *f_fname, *f_fname_noext, *f_fname_trim, *tmp;
238 /* variables to be used for derivatives FILE *f */
239 char *tmp_fname_noext, *tmp_fname_trim, *tmp_fname_ext, *tmpresult;
241 int len, i, j, i_sub_count, i_result2;
242 subfn *result; /* unsorted results */
243 char **result2; /* sorted results */
244 char **tmp_subdirs, **subdirs; /* list of subdirectories to look in */
250 if( !strncmp( psz_fname, "file://", 7 ) )
256 len = strlen( psz_fname ) > 256 ? strlen( psz_fname ) : 256;
258 f_dir = (char*)malloc(len);
259 f_fname = (char*)malloc(len);
260 f_fname_noext = (char*)malloc(len);
261 f_fname_trim = (char*)malloc(len);
263 tmp_fname_noext = (char*)malloc(len);
264 tmp_fname_trim = (char*)malloc(len);
265 tmp_fname_ext = (char*)malloc(len);
267 tmpresult = (char*)malloc(len);
269 result = (subfn*)malloc( sizeof(subfn) * MAX_SUBTITLE_FILES );
270 memset( result, 0, sizeof(subfn) * MAX_SUBTITLE_FILES );
272 /* extract filename & dirname from psz_fname */
273 tmp = strrchr( psz_fname, DIRECTORY_SEPARATOR );
277 strncpy( f_fname, tmp + 1, len - 1 );
278 f_fname[len - 1] = 0;
279 pos = tmp - psz_fname + 1;
280 strncpy( f_dir, psz_fname, __MIN(pos,len-1) );
281 f_dir[__MIN(pos,len-1)] = 0;
285 strncpy( f_fname, psz_fname, len - 1 );
286 f_fname[len - 1] = 0;
290 strcpy_strip_ext( f_fname_noext, f_fname );
291 strcpy_trim( f_fname_trim, f_fname_noext );
292 var_Get( p_this, "sub-autodetect-fuzzy", &fuzzy );
294 tmp_subdirs = paths_to_list( f_dir, psz_path );
295 subdirs = tmp_subdirs;
297 for( j = -1; j == -1 || ( j >= 0 && subdirs != NULL && *subdirs != NULL );
300 d = opendir( j < 0 ? f_dir : *subdirs );
304 msg_Dbg( p_this, "looking for a subtitle file in %s",
305 j < 0 ? f_dir : *subdirs );
306 while( ( de = readdir( d ) ) )
308 /* retrieve various parts of the filename */
309 strcpy_strip_ext( tmp_fname_noext, de->d_name );
310 strcpy_get_ext( tmp_fname_ext, de->d_name );
311 strcpy_trim( tmp_fname_trim, tmp_fname_noext );
313 /* does it end with a subtitle extension? */
315 for( i = 0; sub_exts[i]; i++ )
317 if( strcmp( sub_exts[i], tmp_fname_ext ) == 0 )
320 msg_Dbg( p_this, "found a possible subtitle: %s",
326 /* we have a (likely) subtitle file */
330 if( !i_prio && !strcmp( tmp_fname_trim, f_fname_trim ) )
332 /* matches the movie name exactly */
336 ( tmp = strstr( tmp_fname_trim, f_fname_trim ) ) )
338 /* contains the movie name */
339 tmp += strlen( f_fname_trim );
340 if( whiteonly( tmp ) )
342 /* chars in front of the movie name */
347 /* chars after (and possibly in front of)
354 /* doesn't contain the movie name */
355 if( j == 0 ) i_prio = 1;
358 if( i_prio >= fuzzy.i_int )
360 sprintf( tmpresult, "%s%s", j == -1 ? f_dir : *subdirs,
362 msg_Dbg( p_this, "autodetected subtitle: %s with "
363 "priority %d", de->d_name, i_prio );
364 if( ( f = fopen( tmpresult, "rt" ) ) )
367 result[i_sub_count].priority = i_prio;
368 result[i_sub_count].psz_fname = strdup(tmpresult);
369 result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
374 if( i_sub_count >= MAX_SUBTITLE_FILES ) break;
378 if( j >= 0 ) free( *subdirs++ );
381 if( tmp_subdirs ) free( tmp_subdirs );
385 free( f_fname_noext );
386 free( f_fname_trim );
388 free( tmp_fname_noext );
389 free( tmp_fname_trim );
390 free( tmp_fname_ext );
394 qsort( result, i_sub_count, sizeof( subfn ), compare_sub_priority );
396 result2 = (char**)malloc( sizeof(char*) * ( i_sub_count + 1 ) );
397 memset( result2, 0, sizeof(char*) * ( i_sub_count + 1 ) );
400 for( i = 0; i < i_sub_count; i++ )
402 if( result[i].psz_ext && !strcasecmp( result[i].psz_ext, "sub" ) )
405 for( j = 0; j < i_sub_count; j++ )
407 if( result[j].psz_fname && result[i].psz_fname &&
408 !strncasecmp( result[i].psz_fname, result[j].psz_fname, sizeof( result[i].psz_fname) - 4 ) &&
409 !strcasecmp( result[j].psz_ext, "idx" ) )
413 if( j >= i_sub_count )
415 result2[i_result2] = result[i].psz_fname;
421 result2[i_result2] = result[i].psz_fname;