src/input/subtitles.c

   1 /*****************************************************************************
   2  * subtitles.c
   3  *****************************************************************************
   4  * Copyright (C) 2003-2006 the VideoLAN team
   5  * $Id$
   6  *
   7  * Authors: Derk-Jan Hartman <hartman at videolan.org>
   8  * This is adapted code from the GPL'ed MPlayer (http://mplayerhq.hu)
   9  *
  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.
  14  *
  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.
  19  *
  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  *****************************************************************************/
  24
  25 /**
  26  *  \file
  27  *  This file contains functions to dectect subtitle files.
  28  */
  29
  30 #ifdef HAVE_CONFIG_H
  31 # include "config.h"
  32 #endif
  33
  34 #include <vlc_common.h>
  35 #include <vlc_input.h>
  36 #include <vlc_charset.h>
  37
  38 #ifdef HAVE_DIRENT_H
  39 #   include <dirent.h>
  40 #endif
  41
  42 #include <limits.h>
  43
  44 #ifdef HAVE_UNISTD_H
  45 #   include <unistd.h>
  46 #endif
  47 #include <sys/stat.h>
  48
  49 #include <ctype.h>
  50 #include "input_internal.h"
  51
  52 /**
  53  * We are not going to autodetect more subtitle files than this.
  54  */
  55 #define MAX_SUBTITLE_FILES 128
  56
  57
  58 /**
  59  * The possible extensions for subtitle files we support
  60  */
  61 static const char const sub_exts[][6] = {
  62     "utf", "utf8", "utf-8",
  63     "sub", "srt", "smi",
  64     "txt", "ssa", "idx",
  65
  66     "cdg",
  67
  68     ""
  69 };
  70
  71 /* extensions from unsupported types */
  72 /* rt, aqt, jss, js, ass */
  73
  74 static void strcpy_trim( char *d, const char *s )
  75 {
  76     /* skip leading whitespace */
  77     while( *s && !isalnum(*s) )
  78     {
  79         s++;
  80     }
  81     for(;;)
  82     {
  83         /* copy word */
  84         while( *s && isalnum(*s) )
  85         {
  86             *d = tolower(*s);
  87             s++; d++;
  88         }
  89         if( *s == 0 ) break;
  90         /* trim excess whitespace */
  91         while( *s && !isalnum(*s) )
  92         {
  93             s++;
  94         }
  95         if( *s == 0 ) break;
  96         *d++ = ' ';
  97     }
  98     *d = 0;
  99 }
 100
 101 static void strcpy_strip_ext( char *d, const char *s )
 102 {
 103     const char *tmp = strrchr(s, '.');
 104     if( !tmp )
 105     {
 106         strcpy(d, s);
 107         return;
 108     }
 109     else
 110         strlcpy(d, s, tmp - s + 1 );
 111     while( *d )
 112     {
 113         *d = tolower(*d);
 114         d++;
 115     }
 116 }
 117
 118 static void strcpy_get_ext( char *d, const char *s )
 119 {
 120     const char *tmp = strrchr(s, '.');
 121     if( !tmp )
 122         strcpy(d, "");
 123     else
 124         strcpy( d, tmp + 1 );
 125 }
 126
 127 static int whiteonly( const char *s )
 128 {
 129     while( *s )
 130     {
 131         if( isalnum( *s ) )
 132             return 0;
 133         s++;
 134     }
 135     return 1;
 136 }
 137
 138 enum
 139 {
 140     SUB_PRIORITY_NONE = 0,
 141     SUB_PRIORITY_MATCH_NONE = 1,
 142     SUB_PRIORITY_MATCH_RIGHT = 2,
 143     SUB_PRIORITY_MATCH_LEFT = 3,
 144     SUB_PRIORITY_MATCH_ALL = 4,
 145 };
 146 typedef struct
 147 {
 148     int priority;
 149     char *psz_fname;
 150     char *psz_ext;
 151 } vlc_subfn_t;
 152
 153 static int compare_sub_priority( const void *a, const void *b )
 154 {
 155     const vlc_subfn_t *p0 = a;
 156     const vlc_subfn_t *p1 = b;
 157
 158     if( p0->priority > p1->priority )
 159         return -1;
 160
 161     if( p0->priority < p1->priority )
 162         return 1;
 163
 164 #ifndef UNDER_CE
 165     return strcoll( p0->psz_fname, p1->psz_fname);
 166 #else
 167     return strcmp( p0->psz_fname, p1->psz_fname);
 168 #endif
 169 }
 170
 171 /*
 172  * Check if a file ends with a subtitle extension
 173  */
 174 int subtitles_Filter( const char *psz_dir_content )
 175 {
 176     const char *tmp = strrchr( psz_dir_content, '.');
 177     int i;
 178
 179     if( !tmp )
 180         return 0;
 181     tmp++;
 182
 183     for( i = 0; sub_exts[i][0]; i++ )
 184         if( strcasecmp( sub_exts[i], tmp ) == 0 )
 185             return 1;
 186     return 0;
 187 }
 188
 189
 190 /**
 191  * Convert a list of paths separated by ',' to a char**
 192  */
 193 static char **paths_to_list( const char *psz_dir, char *psz_path )
 194 {
 195     unsigned int i, k, i_nb_subdirs;
 196     char **subdirs; /* list of subdirectories to look in */
 197     char *psz_parser = psz_path;
 198
 199     if( !psz_dir || !psz_path )
 200         return NULL;
 201
 202     for( k = 0, i_nb_subdirs = 1; psz_path[k] != '\0'; k++ )
 203     {
 204         if( psz_path[k] == ',' )
 205             i_nb_subdirs++;
 206     }
 207
 208     subdirs = calloc( i_nb_subdirs + 1, sizeof(char*) );
 209     if( !subdirs )
 210         return NULL;
 211
 212     for( i = 0; psz_parser && *psz_parser != '\0' ; )
 213     {
 214         char *psz_subdir = psz_parser;
 215         psz_parser = strchr( psz_subdir, ',' );
 216         if( psz_parser )
 217         {
 218             *psz_parser++ = '\0';
 219             while( *psz_parser == ' ' )
 220                 psz_parser++;
 221         }
 222         if( *psz_subdir == '\0' )
 223             continue;
 224
 225         asprintf( &subdirs[i++], "%s%s%c",
 226                   psz_subdir[0] == '.' ? psz_dir : "",
 227                   psz_subdir,
 228                   psz_subdir[strlen(psz_subdir) - 1] == DIR_SEP_CHAR ? '\0' : DIR_SEP_CHAR );
 229     }
 230     subdirs[i] = NULL;
 231
 232     return subdirs;
 233 }
 234
 235
 236 /**
 237  * Detect subtitle files.
 238  *
 239  * When called this function will split up the psz_name string into a
 240  * directory, filename and extension. It then opens the directory
 241  * in which the file resides and tries to find possible matches of
 242  * subtitles files.
 243  *
 244  * \ingroup Demux
 245  * \param p_this the calling \ref input_thread_t
 246  * \param psz_path a list of subdirectories (separated by a ',') to look in.
 247  * \param psz_name the complete filename to base the search on.
 248  * \return a NULL terminated array of filenames with detected possible subtitles.
 249  * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
 250  */
 251 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
 252                          const char *psz_name_org )
 253 {
 254     vlc_value_t fuzzy;
 255     int j, i_result2, i_sub_count, i_fname_len;
 256     char *f_dir = NULL, *f_fname = NULL, *f_fname_noext = NULL, *f_fname_trim = NULL;
 257     char *tmp = NULL;
 258
 259     char **subdirs; /* list of subdirectories to look in */
 260
 261     vlc_subfn_t *result = NULL; /* unsorted results */
 262     char **result2; /* sorted results */
 263     const char *psz_fname = psz_name_org;
 264
 265     if( !psz_fname )
 266         return NULL;
 267
 268     if( !strncmp( psz_fname, "file://", 7 ) )
 269         psz_fname += 7;
 270
 271     /* extract filename & dirname from psz_fname */
 272     tmp = strrchr( psz_fname, DIR_SEP_CHAR );
 273     if( tmp )
 274     {
 275         const int i_dirlen = strlen(psz_fname)-strlen(tmp)+1; /* include the separator */
 276         f_fname = strdup( &tmp[1] );    /* skip the separator */
 277         f_dir = strndup( psz_fname, i_dirlen );
 278     }
 279     else
 280     {
 281 #ifdef HAVE_UNISTD_H
 282         /* Get the current working directory */
 283         char *psz_cwd = getcwd( NULL, 0 );
 284 #else
 285         char *psz_cwd = NULL;
 286 #endif
 287         if( !psz_cwd )
 288             return NULL;
 289
 290         f_fname = strdup( psz_fname );
 291         asprintf( &f_dir, "%s%c", psz_cwd, DIR_SEP_CHAR );
 292         free( psz_cwd );
 293     }
 294     if( !f_fname || !f_dir )
 295     {
 296         free( f_fname );
 297         free( f_dir );
 298         return NULL;
 299     }
 300
 301     i_fname_len = strlen( f_fname );
 302
 303     f_fname_noext = malloc(i_fname_len + 1);
 304     f_fname_trim = malloc(i_fname_len + 1 );
 305     if( !f_fname_noext || !f_fname_trim )
 306     {
 307         free( f_fname );
 308         free( f_dir );
 309         free( f_fname_noext );
 310         free( f_fname_trim );
 311         return NULL;
 312     }
 313
 314     strcpy_strip_ext( f_fname_noext, f_fname );
 315     strcpy_trim( f_fname_trim, f_fname_noext );
 316
 317     var_Get( p_this, "sub-autodetect-fuzzy", &fuzzy );
 318
 319     result = calloc( MAX_SUBTITLE_FILES+1, sizeof(vlc_subfn_t) ); /* We check it later (simplify code) */
 320     subdirs = paths_to_list( f_dir, psz_path );
 321     for( j = -1, i_sub_count = 0; (j == -1) || ( j >= 0 && subdirs != NULL && subdirs[j] != NULL ); j++ )
 322     {
 323         const char *psz_dir = j < 0 ? f_dir : subdirs[j];
 324         char **ppsz_dir_content;
 325         int i_dir_content;
 326         int a;
 327
 328         if( psz_dir == NULL || ( j >= 0 && !strcmp( psz_dir, f_dir ) ) )
 329             continue;
 330
 331         /* parse psz_src dir */
 332         i_dir_content = utf8_scandir( psz_dir, &ppsz_dir_content,
 333                                       subtitles_Filter, NULL );
 334         if( i_dir_content < 0 )
 335             continue;
 336
 337         msg_Dbg( p_this, "looking for a subtitle file in %s", psz_dir );
 338         for( a = 0; a < i_dir_content && i_sub_count < MAX_SUBTITLE_FILES ; a++ )
 339         {
 340             char *psz_name = ppsz_dir_content[a];
 341             char tmp_fname_noext[strlen( psz_name ) + 1];
 342             char tmp_fname_trim[strlen( psz_name ) + 1];
 343             char tmp_fname_ext[strlen( psz_name ) + 1];
 344
 345             int i_prio;
 346
 347             if( psz_name == NULL )
 348                 continue;
 349
 350             /* retrieve various parts of the filename */
 351             strcpy_strip_ext( tmp_fname_noext, psz_name );
 352             strcpy_get_ext( tmp_fname_ext, psz_name );
 353             strcpy_trim( tmp_fname_trim, tmp_fname_noext );
 354
 355             i_prio = SUB_PRIORITY_NONE;
 356             if( i_prio == SUB_PRIORITY_NONE && !strcmp( tmp_fname_trim, f_fname_trim ) )
 357             {
 358                 /* matches the movie name exactly */
 359                 i_prio = SUB_PRIORITY_MATCH_ALL;
 360             }
 361             if( i_prio == SUB_PRIORITY_NONE &&
 362                 ( tmp = strstr( tmp_fname_trim, f_fname_trim ) ) )
 363             {
 364                 /* contains the movie name */
 365                 tmp += strlen( f_fname_trim );
 366                 if( whiteonly( tmp ) )
 367                 {
 368                     /* chars in front of the movie name */
 369                     i_prio = SUB_PRIORITY_MATCH_RIGHT;
 370                 }
 371                 else
 372                 {
 373                     /* chars after (and possibly in front of)
 374                      * the movie name */
 375                     i_prio = SUB_PRIORITY_MATCH_LEFT;
 376                 }
 377             }
 378             if( i_prio == SUB_PRIORITY_NONE &&
 379                 j == 0 )
 380             {
 381                 /* doesn't contain the movie name, prefer files in f_dir over subdirs */
 382                 i_prio = SUB_PRIORITY_MATCH_NONE;
 383             }
 384             if( i_prio >= fuzzy.i_int )
 385             {
 386                 char psz_path[strlen( psz_dir ) + strlen( psz_name ) + 1];
 387                 struct stat st;
 388
 389                 sprintf( psz_path, "%s%s", psz_dir, psz_name );
 390                 if( !strcmp( psz_path, psz_fname ) )
 391                     continue;
 392
 393                 if( !utf8_stat( psz_path, &st ) && S_ISREG( st.st_mode ) && result )
 394                 {
 395                     msg_Dbg( p_this,
 396                             "autodetected subtitle: %s with priority %d",
 397                             psz_path, i_prio );
 398                     result[i_sub_count].priority = i_prio;
 399                     result[i_sub_count].psz_fname = strdup( psz_path );
 400                     result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
 401                     i_sub_count++;
 402                 }
 403                 else
 404                 {
 405                     msg_Dbg( p_this, "stat failed (autodetecting subtitle: %s with priority %d)",
 406                              psz_path, i_prio );
 407                 }
 408             }
 409         }
 410         if( ppsz_dir_content )
 411         {
 412             for( a = 0; a < i_dir_content; a++ )
 413                 free( ppsz_dir_content[a] );
 414             free( ppsz_dir_content );
 415         }
 416     }
 417     if( subdirs )
 418     {
 419         for( j = 0; subdirs[j]; j++ )
 420             free( subdirs[j] );
 421         free( subdirs );
 422     }
 423     free( f_fname );
 424     free( f_dir );
 425     free( f_fname_trim );
 426     free( f_fname_noext );
 427
 428     if( !result )
 429         return NULL;
 430
 431     qsort( result, i_sub_count, sizeof(vlc_subfn_t), compare_sub_priority );
 432
 433     result2 = calloc( i_sub_count + 1, sizeof(char*) );
 434
 435     for( j = 0, i_result2 = 0; j < i_sub_count && result2 != NULL; j++ )
 436     {
 437         bool b_reject = false;
 438
 439         if( !result[j].psz_fname || !result[j].psz_ext ) /* memory out */
 440             break;
 441
 442         if( !strcasecmp( result[j].psz_ext, "sub" ) )
 443         {
 444             int i;
 445             for( i = 0; i < i_sub_count; i++ )
 446             {
 447                 if( result[i].psz_fname && result[i].psz_ext &&
 448                     !strncasecmp( result[j].psz_fname, result[i].psz_fname,
 449                                   strlen( result[j].psz_fname) - 3 ) &&
 450                     !strcasecmp( result[i].psz_ext, "idx" ) )
 451                     break;
 452             }
 453             if( i < i_sub_count )
 454                 b_reject = true;
 455         }
 456         else if( !strcasecmp( result[j].psz_ext, "cdg" ) )
 457         {
 458             if( result[j].priority < SUB_PRIORITY_MATCH_ALL )
 459                 b_reject = true;
 460         }
 461
 462         /* */
 463         if( !b_reject )
 464             result2[i_result2++] = strdup( result[j].psz_fname );
 465     }
 466
 467     for( j = 0; j < i_sub_count; j++ )
 468     {
 469         free( result[j].psz_fname );
 470         free( result[j].psz_ext );
 471     }
 472     free( result );
 473
 474     return result2;
 475 }
 476