git.sesse.net Git - ffmpeg/blob - libavcodec/htmlsubtitles.c

   1 /*
   2  * Copyright (c) 2010  Aurelien Jacobs <aurel@gnuage.org>
   3  *
   4  * This file is part of FFmpeg.
   5  *
   6  * FFmpeg is free software; you can redistribute it and/or
   7  * modify it under the terms of the GNU Lesser General Public
   8  * License as published by the Free Software Foundation; either
   9  * version 2.1 of the License, or (at your option) any later version.
  10  *
  11  * FFmpeg is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14  * Lesser General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU Lesser General Public
  17  * License along with FFmpeg; if not, write to the Free Software
  18  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  19  */
  20
  21 #include "libavutil/avstring.h"
  22 #include "libavutil/common.h"
  23 #include "libavutil/parseutils.h"
  24 #include "htmlsubtitles.h"
  25
  26 static int html_color_parse(void *log_ctx, const char *str)
  27 {
  28     uint8_t rgba[4];
  29     if (av_parse_color(rgba, str, strcspn(str, "\" >"), log_ctx) < 0)
  30         return -1;
  31     return rgba[0] | rgba[1] << 8 | rgba[2] << 16;
  32 }
  33
  34 enum {
  35     PARAM_UNKNOWN = -1,
  36     PARAM_SIZE,
  37     PARAM_COLOR,
  38     PARAM_FACE,
  39     PARAM_NUMBER
  40 };
  41
  42 typedef struct SrtStack {
  43     char tag[128];
  44     char param[PARAM_NUMBER][128];
  45 } SrtStack;
  46
  47 static void rstrip_spaces_buf(AVBPrint *buf)
  48 {
  49     if (av_bprint_is_complete(buf))
  50         while (buf->len > 0 && buf->str[buf->len - 1] == ' ')
  51             buf->str[--buf->len] = 0;
  52 }
  53
  54 /* skip all {\xxx} substrings except for {\an%d}
  55    and all microdvd like styles such as {Y:xxx} */
  56 static void handle_open_brace(AVBPrint *dst, const char **inp, int *an, int *closing_brace_missing)
  57 {
  58     int len = 0;
  59     const char *in = *inp;
  60
  61     *an += sscanf(in, "{\\an%*1u}%n", &len) >= 0 && len > 0;
  62
  63     if (!*closing_brace_missing) {
  64         if (   (*an != 1 && in[1] == '\\')
  65             || (in[1] && strchr("CcFfoPSsYy", in[1]) && in[2] == ':')) {
  66             char *bracep = strchr(in+2, '}');
  67             if (bracep) {
  68                 *inp = bracep;
  69                 return;
  70             } else
  71                 *closing_brace_missing = 1;
  72         }
  73     }
  74
  75     av_bprint_chars(dst, *in, 1);
  76 }
  77
  78 int ff_htmlmarkup_to_ass(void *log_ctx, AVBPrint *dst, const char *in)
  79 {
  80     char *param, buffer[128], tmp[128];
  81     int len, tag_close, sptr = 1, line_start = 1, an = 0, end = 0;
  82     SrtStack stack[16];
  83     int closing_brace_missing = 0;
  84
  85     stack[0].tag[0] = 0;
  86     strcpy(stack[0].param[PARAM_SIZE],  "{\\fs}");
  87     strcpy(stack[0].param[PARAM_COLOR], "{\\c}");
  88     strcpy(stack[0].param[PARAM_FACE],  "{\\fn}");
  89
  90     for (; !end && *in; in++) {
  91         switch (*in) {
  92         case '\r':
  93             break;
  94         case '\n':
  95             if (line_start) {
  96                 end = 1;
  97                 break;
  98             }
  99             rstrip_spaces_buf(dst);
 100             av_bprintf(dst, "\\N");
 101             line_start = 1;
 102             break;
 103         case ' ':
 104             if (!line_start)
 105                 av_bprint_chars(dst, *in, 1);
 106             break;
 107         case '{':
 108             handle_open_brace(dst, &in, &an, &closing_brace_missing);
 109             break;
 110         case '<':
 111             tag_close = in[1] == '/';
 112             len = 0;
 113             if (sscanf(in+tag_close+1, "%127[^<>]>%n", buffer, &len) >= 1 && len > 0) {
 114                 const char *tagname = buffer;
 115                 while (*tagname == ' ')
 116                     tagname++;
 117                 if ((param = strchr(tagname, ' ')))
 118                     *param++ = 0;
 119                 if ((!tag_close && sptr < FF_ARRAY_ELEMS(stack) && *tagname) ||
 120                     ( tag_close && sptr > 0 && !av_strcasecmp(stack[sptr-1].tag, tagname))) {
 121                     int i, j, unknown = 0;
 122                     in += len + tag_close;
 123                     if (!tag_close)
 124                         memset(stack+sptr, 0, sizeof(*stack));
 125                     if (!av_strcasecmp(tagname, "font")) {
 126                         if (tag_close) {
 127                             for (i=PARAM_NUMBER-1; i>=0; i--)
 128                                 if (stack[sptr-1].param[i][0])
 129                                     for (j=sptr-2; j>=0; j--)
 130                                         if (stack[j].param[i][0]) {
 131                                             av_bprintf(dst, "%s", stack[j].param[i]);
 132                                             break;
 133                                         }
 134                         } else {
 135                             while (param) {
 136                                 if (!av_strncasecmp(param, "size=", 5)) {
 137                                     unsigned font_size;
 138                                     param += 5 + (param[5] == '"');
 139                                     if (sscanf(param, "%u", &font_size) == 1) {
 140                                         snprintf(stack[sptr].param[PARAM_SIZE],
 141                                              sizeof(stack[0].param[PARAM_SIZE]),
 142                                              "{\\fs%u}", font_size);
 143                                     }
 144                                 } else if (!av_strncasecmp(param, "color=", 6)) {
 145                                     param += 6 + (param[6] == '"');
 146                                     snprintf(stack[sptr].param[PARAM_COLOR],
 147                                          sizeof(stack[0].param[PARAM_COLOR]),
 148                                          "{\\c&H%X&}",
 149                                          html_color_parse(log_ctx, param));
 150                                 } else if (!av_strncasecmp(param, "face=", 5)) {
 151                                     param += 5 + (param[5] == '"');
 152                                     len = strcspn(param,
 153                                                   param[-1] == '"' ? "\"" :" ");
 154                                     av_strlcpy(tmp, param,
 155                                                FFMIN(sizeof(tmp), len+1));
 156                                     param += len;
 157                                     snprintf(stack[sptr].param[PARAM_FACE],
 158                                              sizeof(stack[0].param[PARAM_FACE]),
 159                                              "{\\fn%s}", tmp);
 160                                 }
 161                                 if ((param = strchr(param, ' ')))
 162                                     param++;
 163                             }
 164                             for (i=0; i<PARAM_NUMBER; i++)
 165                                 if (stack[sptr].param[i][0])
 166                                     av_bprintf(dst, "%s", stack[sptr].param[i]);
 167                         }
 168                     } else if (tagname[0] && !tagname[1] && av_stristr("bisu", tagname)) {
 169                         av_bprintf(dst, "{\\%c%d}", (char)av_tolower(tagname[0]), !tag_close);
 170                     } else if (!av_strcasecmp(tagname, "br")) {
 171                         av_bprintf(dst, "\\N");
 172                     } else {
 173                         unknown = 1;
 174                         snprintf(tmp, sizeof(tmp), "</%s>", tagname);
 175                     }
 176                     if (tag_close) {
 177                         sptr--;
 178                     } else if (unknown && !strstr(in, tmp)) {
 179                         in -= len + tag_close;
 180                         av_bprint_chars(dst, *in, 1);
 181                     } else
 182                         av_strlcpy(stack[sptr++].tag, tagname,
 183                                    sizeof(stack[0].tag));
 184                     break;
 185                 }
 186             }
 187         default:
 188             av_bprint_chars(dst, *in, 1);
 189             break;
 190         }
 191         if (*in != ' ' && *in != '\r' && *in != '\n')
 192             line_start = 0;
 193     }
 194
 195     if (!av_bprint_is_complete(dst))
 196         return AVERROR(ENOMEM);
 197
 198     while (dst->len >= 2 && !strncmp(&dst->str[dst->len - 2], "\\N", 2))
 199         dst->len -= 2;
 200     dst->str[dst->len] = 0;
 201     rstrip_spaces_buf(dst);
 202
 203     return 0;
 204 }