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 const sub_exts[][6] = {
60 "utf", "utf8", "utf-8",
62 "txt", "ssa", "idx", ""
65 /* extensions from unsupported types */
66 /* rt, aqt, jss, js, ass */
68 static void strcpy_trim( char *d, char *s )
70 /* skip leading whitespace */
71 while( *s && !isalnum(*s) )
78 while( *s && isalnum(*s) )
84 /* trim excess whitespace */
85 while( *s && !isalnum(*s) )
95 static void strcpy_strip_ext( char *d, char *s )
97 char *tmp = strrchr(s, '.');
104 strlcpy(d, s, tmp - s + 1 );
112 static void strcpy_get_ext( char *d, char *s )
114 char *tmp = strrchr(s, '.');
119 } else strcpy( d, tmp + 1 );
122 static int whiteonly( char *s )
126 if( isalnum( *s ) ) return 0;
132 typedef struct _subfn
139 static int compare_sub_priority( const void *a, const void *b )
141 if (((subfn*)a)->priority > ((subfn*)b)->priority)
146 if (((subfn*)a)->priority < ((subfn*)b)->priority)
152 return strcoll(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
154 return strcmp(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
159 * Check if a file ends with a subtitle extension
161 int subtitles_Filter( const char *psz_dir_content )
163 const char *tmp = strrchr( psz_dir_content, '.');
171 for( i = 0; sub_exts[i][0]; i++ )
172 if( strcasecmp( sub_exts[i], tmp ) == 0 )
180 * Convert a list of paths separated by ',' to a char**
182 static char **paths_to_list( const char *psz_dir, char *psz_path )
184 unsigned int i, k, i_nb_subdirs;
185 char **subdirs; /* list of subdirectories to look in */
187 if( !psz_dir ) return NULL;
188 if( !psz_path ) return NULL;
191 for( k = 0; k < strlen( psz_path ); k++ )
193 if( psz_path[k] == ',' )
199 if( i_nb_subdirs > 0 )
201 char *psz_parser = NULL, *psz_temp = NULL;
203 subdirs = (char**)malloc( sizeof(char*) * ( i_nb_subdirs + 1 ) );
204 memset( subdirs, 0, sizeof(char*) * ( i_nb_subdirs + 1 ) );
206 psz_parser = psz_path;
207 while( psz_parser && *psz_parser )
210 psz_subdir = psz_parser;
211 psz_parser = strchr( psz_subdir, ',' );
216 while( *psz_parser == ' ' )
221 if( strlen( psz_subdir ) > 0 )
223 psz_temp = (char *)malloc( strlen(psz_dir)
224 + strlen(psz_subdir) + 2 );
227 sprintf( psz_temp, "%s%s%c",
228 psz_subdir[0] == '.' ? psz_dir : "",
230 psz_subdir[strlen(psz_subdir) - 1] ==
231 DIR_SEP_CHAR ? '\0' : DIR_SEP_CHAR );
232 subdirs[i] = psz_temp;
248 * Detect subtitle files.
250 * When called this function will split up the psz_name string into a
251 * directory, filename and extension. It then opens the directory
252 * in which the file resides and tries to find possible matches of
256 * \param p_this the calling \ref input_thread_t
257 * \param psz_path a list of subdirectories (separated by a ',') to look in.
258 * \param psz_name the complete filename to base the search on.
259 * \return a NULL terminated array of filenames with detected possible subtitles.
260 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
262 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
263 const char *psz_name )
266 int j, i_result2, i_sub_count = 0, i_fname_len = 0;
267 char *f_dir = NULL, *f_fname = NULL, *f_fname_noext = NULL, *f_fname_trim = NULL;
270 char **tmp_subdirs, **subdirs; /* list of subdirectories to look in */
272 subfn *result = NULL; /* unsorted results */
273 char **result2; /* sorted results */
275 char *psz_fname_original = strdup( psz_name );
276 char *psz_fname = psz_fname_original;
278 if( psz_fname == NULL ) return NULL;
280 if( !strncmp( psz_fname, "file://", 7 ) )
285 /* extract filename & dirname from psz_fname */
286 tmp = strrchr( psz_fname, DIR_SEP_CHAR );
291 f_fname = malloc( strlen(tmp) );
293 strcpy( f_fname, tmp+1 ); // we skip the separator, so it will still fit in the allocated space
294 dirlen = strlen(psz_fname) - strlen(tmp) + 2; // add the separator
295 f_dir = malloc( dirlen + 1 );
297 strlcpy( f_dir, psz_fname, dirlen );
301 /* Get the current working directory */
304 f_dir = getcwd( NULL, 0 );
308 if( psz_fname_original ) free( psz_fname_original );
311 dirlen = strlen( f_dir );
312 f_dir = (char *)realloc(f_dir, dirlen +2 );
313 f_dir[dirlen] = DIR_SEP_CHAR;
314 f_dir[dirlen+1] = '\0';
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) &&
335 (*subdirs != NULL) ); j++)
337 const char *psz_dir = j < 0 ? f_dir : *subdirs;
338 char **ppsz_dir_content;
341 if( psz_dir == NULL )
344 /* parse psz_src dir */
345 i_dir_content = utf8_scandir( psz_dir, &ppsz_dir_content,
346 subtitles_Filter, NULL );
348 if( i_dir_content != -1 )
352 msg_Dbg( p_this, "looking for a subtitle file in %s", psz_dir );
353 for( a = 0; a < i_dir_content; a++ )
355 char *psz_name = ppsz_dir_content[a];
356 char tmp_fname_noext[strlen( psz_name ) + 1];
357 char tmp_fname_trim[strlen( psz_name ) + 1];
358 char tmp_fname_ext[strlen( psz_name ) + 1];
362 if( psz_name == NULL )
365 /* retrieve various parts of the filename */
366 strcpy_strip_ext( tmp_fname_noext, psz_name );
367 strcpy_get_ext( tmp_fname_ext, psz_name );
368 strcpy_trim( tmp_fname_trim, tmp_fname_noext );
370 if( !i_prio && !strcmp( tmp_fname_trim, f_fname_trim ) )
372 /* matches the movie name exactly */
376 ( tmp = strstr( tmp_fname_trim, f_fname_trim ) ) )
378 /* contains the movie name */
379 tmp += strlen( f_fname_trim );
380 if( whiteonly( tmp ) )
382 /* chars in front of the movie name */
387 /* chars after (and possibly in front of)
394 /* doesn't contain the movie name */
395 if( j == 0 ) i_prio = 1;
397 if( i_prio >= fuzzy.i_int )
399 char psz_path[strlen( psz_dir ) + strlen( psz_name ) + 1];
402 sprintf( psz_path, "%s%s", psz_dir, psz_name );
404 "autodetected subtitle: %s with priority %d",
407 if( !utf8_stat( psz_path, &st ) && S_ISREG( st.st_mode ) )
410 "autodetected subtitle: %s with priority %d",
412 result[i_sub_count].priority = i_prio;
413 result[i_sub_count].psz_fname = strdup( psz_path );
414 result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
419 msg_Dbg( p_this, "stat failed" );
422 if( i_sub_count >= MAX_SUBTITLE_FILES ) break;
424 for( a = 0; a < i_dir_content; a++ )
425 free( ppsz_dir_content[a] );
426 if( ppsz_dir_content ) free( ppsz_dir_content );
428 if( j >= 0 ) if( *subdirs ) free( *subdirs++ );
431 if( tmp_subdirs ) free( tmp_subdirs );
432 if( f_fname_trim ) free( f_fname_trim );
433 if( f_fname_noext ) free( f_fname_noext );
434 if( f_fname ) free( f_fname );
435 if( f_dir ) free( f_dir );
437 qsort( result, i_sub_count, sizeof( subfn ), compare_sub_priority );
439 result2 = (char**)malloc( sizeof(char*) * ( i_sub_count + 1 ) );
441 memset( result2, 0, sizeof(char*) * ( i_sub_count + 1 ) );
444 for( j = 0; j < i_sub_count; j++ )
446 if( result[j].psz_ext && !strcasecmp( result[j].psz_ext, "sub" ) )
449 for( i = 0; i < i_sub_count; i++ )
451 if( result[i].psz_fname && result[j].psz_fname &&
452 !strncasecmp( result[j].psz_fname, result[i].psz_fname,
453 sizeof( result[j].psz_fname) - 4 ) &&
454 !strcasecmp( result[i].psz_ext, "idx" ) )
457 if( i >= i_sub_count )
459 result2[i_result2] = result[j].psz_fname;
465 result2[i_result2] = result[j].psz_fname;
470 if( psz_fname_original ) free( psz_fname_original );
471 if( result ) free( result );