1 /*****************************************************************************
2 * subtitles.c : subtitles detection
3 *****************************************************************************
4 * Copyright (C) 2003-2009 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.
34 #include <vlc_common.h>
44 #include <ctype.h> /* isalnum */
46 #include "input_internal.h"
49 * We are not going to autodetect more subtitle files than this.
51 #define MAX_SUBTITLE_FILES 128
54 * The possible extensions for subtitle files we support
56 static const char const sub_exts[][6] = {
59 "utf", "utf8", "utf-8",
62 "psb", "mpsub","mpl2",
67 static void strcpy_trim( char *d, const 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, const char *s )
96 const char *tmp = strrchr(s, '.');
103 strlcpy(d, s, tmp - s + 1 );
111 static void strcpy_get_ext( char *d, const char *s )
113 const char *tmp = strrchr(s, '.');
117 strcpy( d, tmp + 1 );
120 static int whiteonly( const char *s )
133 SUB_PRIORITY_NONE = 0,
134 SUB_PRIORITY_MATCH_NONE = 1,
135 SUB_PRIORITY_MATCH_RIGHT = 2,
136 SUB_PRIORITY_MATCH_LEFT = 3,
137 SUB_PRIORITY_MATCH_ALL = 4,
146 static int compare_sub_priority( const void *a, const void *b )
148 const vlc_subfn_t *p0 = a;
149 const vlc_subfn_t *p1 = b;
151 if( p0->priority > p1->priority )
154 if( p0->priority < p1->priority )
158 return strcoll( p0->psz_fname, p1->psz_fname);
160 return strcmp( p0->psz_fname, p1->psz_fname);
165 * Check if a file ends with a subtitle extension
167 int subtitles_Filter( const char *psz_dir_content )
169 const char *tmp = strrchr( psz_dir_content, '.');
175 for( int i = 0; sub_exts[i][0]; i++ )
176 if( strcasecmp( sub_exts[i], tmp ) == 0 )
183 * Convert a list of paths separated by ',' to a char**
185 static char **paths_to_list( const char *psz_dir, char *psz_path )
187 unsigned int i, k, i_nb_subdirs;
188 char **subdirs; /* list of subdirectories to look in */
189 char *psz_parser = psz_path;
191 if( !psz_dir || !psz_path )
194 for( k = 0, i_nb_subdirs = 1; psz_path[k] != '\0'; k++ )
196 if( psz_path[k] == ',' )
200 subdirs = calloc( i_nb_subdirs + 1, sizeof(char*) );
204 for( i = 0; psz_parser && *psz_parser != '\0' ; )
206 char *psz_subdir = psz_parser;
207 psz_parser = strchr( psz_subdir, ',' );
210 *psz_parser++ = '\0';
211 while( *psz_parser == ' ' )
214 if( *psz_subdir == '\0' )
217 if( asprintf( &subdirs[i++], "%s%s",
218 psz_subdir[0] == '.' ? psz_dir : "",
229 * Detect subtitle files.
231 * When called this function will split up the psz_name string into a
232 * directory, filename and extension. It then opens the directory
233 * in which the file resides and tries to find possible matches of
237 * \param p_this the calling \ref input_thread_t
238 * \param psz_path a list of subdirectories (separated by a ',') to look in.
239 * \param psz_name the complete filename to base the search on.
240 * \return a NULL terminated array of filenames with detected possible subtitles.
241 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
243 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
244 const char *psz_name_org )
247 int j, i_result2, i_sub_count, i_fname_len;
248 char *f_fname_noext = NULL, *f_fname_trim = NULL;
250 char **subdirs; /* list of subdirectories to look in */
252 vlc_subfn_t *result = NULL; /* unsorted results */
253 char **result2; /* sorted results */
258 char *psz_fname = make_path( psz_name_org );
262 /* extract filename & dirname from psz_fname */
263 char *f_dir = strdup( psz_fname );
270 char *f_fname = strrchr( f_dir, DIR_SEP_CHAR );
277 *(f_fname++) = 0; /* skip dir separator */
279 i_fname_len = strlen( f_fname );
281 f_fname_noext = malloc(i_fname_len + 1);
282 f_fname_trim = malloc(i_fname_len + 1 );
283 if( !f_fname_noext || !f_fname_trim )
286 free( f_fname_noext );
287 free( f_fname_trim );
292 strcpy_strip_ext( f_fname_noext, f_fname );
293 strcpy_trim( f_fname_trim, f_fname_noext );
295 i_fuzzy = var_GetInteger( p_this, "sub-autodetect-fuzzy" );
297 result = calloc( MAX_SUBTITLE_FILES+1, sizeof(vlc_subfn_t) ); /* We check it later (simplify code) */
298 subdirs = paths_to_list( f_dir, psz_path );
299 for( j = -1, i_sub_count = 0; (j == -1) || ( j >= 0 && subdirs != NULL && subdirs[j] != NULL ); j++ )
301 const char *psz_dir = (j < 0) ? f_dir : subdirs[j];
302 if( psz_dir == NULL || ( j >= 0 && !strcmp( psz_dir, f_dir ) ) )
305 /* parse psz_src dir */
306 DIR *dir = vlc_opendir( psz_dir );
310 msg_Dbg( p_this, "looking for a subtitle file in %s", psz_dir );
313 while( (psz_name = vlc_readdir( dir )) && i_sub_count < MAX_SUBTITLE_FILES )
315 if( psz_name[0] == '.' || !subtitles_Filter( psz_name ) )
321 char tmp_fname_noext[strlen( psz_name ) + 1];
322 char tmp_fname_trim[strlen( psz_name ) + 1];
323 char tmp_fname_ext[strlen( psz_name ) + 1];
328 /* retrieve various parts of the filename */
329 strcpy_strip_ext( tmp_fname_noext, psz_name );
330 strcpy_get_ext( tmp_fname_ext, psz_name );
331 strcpy_trim( tmp_fname_trim, tmp_fname_noext );
333 i_prio = SUB_PRIORITY_NONE;
334 if( i_prio == SUB_PRIORITY_NONE && !strcmp( tmp_fname_trim, f_fname_trim ) )
336 /* matches the movie name exactly */
337 i_prio = SUB_PRIORITY_MATCH_ALL;
339 if( i_prio == SUB_PRIORITY_NONE &&
340 ( tmp = strstr( tmp_fname_trim, f_fname_trim ) ) )
342 /* contains the movie name */
343 tmp += strlen( f_fname_trim );
344 if( whiteonly( tmp ) )
346 /* chars in front of the movie name */
347 i_prio = SUB_PRIORITY_MATCH_RIGHT;
351 /* chars after (and possibly in front of)
353 i_prio = SUB_PRIORITY_MATCH_LEFT;
356 if( i_prio == SUB_PRIORITY_NONE &&
359 /* doesn't contain the movie name, prefer files in f_dir over subdirs */
360 i_prio = SUB_PRIORITY_MATCH_NONE;
362 if( i_prio >= i_fuzzy )
364 char psz_path[strlen( psz_dir ) + strlen( psz_name ) + 2];
367 sprintf( psz_path, "%s"DIR_SEP"%s", psz_dir, psz_name );
368 if( !strcmp( psz_path, psz_fname ) )
374 if( !vlc_stat( psz_path, &st ) && S_ISREG( st.st_mode ) && result )
377 "autodetected subtitle: %s with priority %d",
379 result[i_sub_count].priority = i_prio;
380 result[i_sub_count].psz_fname = strdup( psz_path );
381 result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
386 msg_Dbg( p_this, "stat failed (autodetecting subtitle: %s with priority %d)",
396 for( j = 0; subdirs[j]; j++ )
401 free( f_fname_trim );
402 free( f_fname_noext );
408 qsort( result, i_sub_count, sizeof(vlc_subfn_t), compare_sub_priority );
410 result2 = calloc( i_sub_count + 1, sizeof(char*) );
412 for( j = 0, i_result2 = 0; j < i_sub_count && result2 != NULL; j++ )
414 bool b_reject = false;
416 if( !result[j].psz_fname || !result[j].psz_ext ) /* memory out */
419 if( !strcasecmp( result[j].psz_ext, "sub" ) )
422 for( i = 0; i < i_sub_count; i++ )
424 if( result[i].psz_fname && result[i].psz_ext &&
425 !strncasecmp( result[j].psz_fname, result[i].psz_fname,
426 strlen( result[j].psz_fname) - 3 ) &&
427 !strcasecmp( result[i].psz_ext, "idx" ) )
430 if( i < i_sub_count )
433 else if( !strcasecmp( result[j].psz_ext, "cdg" ) )
435 if( result[j].priority < SUB_PRIORITY_MATCH_ALL )
441 result2[i_result2++] = strdup( result[j].psz_fname );
444 for( j = 0; j < i_sub_count; j++ )
446 free( result[j].psz_fname );
447 free( result[j].psz_ext );