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 int ff_htmlmarkup_to_ass(void *log_ctx, AVBPrint *dst, const char *in)
  55 {
  56     char *param, buffer[128], tmp[128];
  57     int len, tag_close, sptr = 1, line_start = 1, an = 0, end = 0;
  58     SrtStack stack[16];
  59
  60     stack[0].tag[0] = 0;
  61     strcpy(stack[0].param[PARAM_SIZE],  "{\\fs}");
  62     strcpy(stack[0].param[PARAM_COLOR], "{\\c}");
  63     strcpy(stack[0].param[PARAM_FACE],  "{\\fn}");
  64
  65     for (; !end && *in; in++) {
  66         switch (*in) {
  67         case '\r':
  68             break;
  69         case '\n':
  70             if (line_start) {
  71                 end = 1;
  72                 break;
  73             }
  74             rstrip_spaces_buf(dst);
  75             av_bprintf(dst, "\\N");
  76             line_start = 1;
  77             break;
  78         case ' ':
  79             if (!line_start)
  80                 av_bprint_chars(dst, *in, 1);
  81             break;
  82         case '{':    /* skip all {\xxx} substrings except for {\an%d}
  83                         and all microdvd like styles such as {Y:xxx} */
  84             len = 0;
  85             an += sscanf(in, "{\\an%*1u}%n", &len) >= 0 && len > 0;
  86             if ((an != 1 && (len = 0, sscanf(in, "{\\%*[^}]}%n", &len) >= 0 && len > 0)) ||
  87                 (len = 0, sscanf(in, "{%*1[CcFfoPSsYy]:%*[^}]}%n", &len) >= 0 && len > 0)) {
  88                 in += len - 1;
  89             } else
  90                 av_bprint_chars(dst, *in, 1);
  91             break;
  92         case '<':
  93             tag_close = in[1] == '/';
  94             len = 0;
  95             if (sscanf(in+tag_close+1, "%127[^>]>%n", buffer, &len) >= 1 && len > 0) {
  96                 const char *tagname = buffer;
  97                 while (*tagname == ' ')
  98                     tagname++;
  99                 if ((param = strchr(tagname, ' ')))
 100                     *param++ = 0;
 101                 if ((!tag_close && sptr < FF_ARRAY_ELEMS(stack)) ||
 102                     ( tag_close && sptr > 0 && !strcmp(stack[sptr-1].tag, tagname))) {
 103                     int i, j, unknown = 0;
 104                     in += len + tag_close;
 105                     if (!tag_close)
 106                         memset(stack+sptr, 0, sizeof(*stack));
 107                     if (!strcmp(tagname, "font")) {
 108                         if (tag_close) {
 109                             for (i=PARAM_NUMBER-1; i>=0; i--)
 110                                 if (stack[sptr-1].param[i][0])
 111                                     for (j=sptr-2; j>=0; j--)
 112                                         if (stack[j].param[i][0]) {
 113                                             av_bprintf(dst, "%s", stack[j].param[i]);
 114                                             break;
 115                                         }
 116                         } else {
 117                             while (param) {
 118                                 if (!strncmp(param, "size=", 5)) {
 119                                     unsigned font_size;
 120                                     param += 5 + (param[5] == '"');
 121                                     if (sscanf(param, "%u", &font_size) == 1) {
 122                                         snprintf(stack[sptr].param[PARAM_SIZE],
 123                                              sizeof(stack[0].param[PARAM_SIZE]),
 124                                              "{\\fs%u}", font_size);
 125                                     }
 126                                 } else if (!strncmp(param, "color=", 6)) {
 127                                     param += 6 + (param[6] == '"');
 128                                     snprintf(stack[sptr].param[PARAM_COLOR],
 129                                          sizeof(stack[0].param[PARAM_COLOR]),
 130                                          "{\\c&H%X&}",
 131                                          html_color_parse(log_ctx, param));
 132                                 } else if (!strncmp(param, "face=", 5)) {
 133                                     param += 5 + (param[5] == '"');
 134                                     len = strcspn(param,
 135                                                   param[-1] == '"' ? "\"" :" ");
 136                                     av_strlcpy(tmp, param,
 137                                                FFMIN(sizeof(tmp), len+1));
 138                                     param += len;
 139                                     snprintf(stack[sptr].param[PARAM_FACE],
 140                                              sizeof(stack[0].param[PARAM_FACE]),
 141                                              "{\\fn%s}", tmp);
 142                                 }
 143                                 if ((param = strchr(param, ' ')))
 144                                     param++;
 145                             }
 146                             for (i=0; i<PARAM_NUMBER; i++)
 147                                 if (stack[sptr].param[i][0])
 148                                     av_bprintf(dst, "%s", stack[sptr].param[i]);
 149                         }
 150                     } else if (tagname[0] && !tagname[1] && strspn(tagname, "bisu") == 1) {
 151                         av_bprintf(dst, "{\\%c%d}", tagname[0], !tag_close);
 152                     } else {
 153                         unknown = 1;
 154                         snprintf(tmp, sizeof(tmp), "</%s>", tagname);
 155                     }
 156                     if (tag_close) {
 157                         sptr--;
 158                     } else if (unknown && !strstr(in, tmp)) {
 159                         in -= len + tag_close;
 160                         av_bprint_chars(dst, *in, 1);
 161                     } else
 162                         av_strlcpy(stack[sptr++].tag, tagname,
 163                                    sizeof(stack[0].tag));
 164                     break;
 165                 }
 166             }
 167         default:
 168             av_bprint_chars(dst, *in, 1);
 169             break;
 170         }
 171         if (*in != ' ' && *in != '\r' && *in != '\n')
 172             line_start = 0;
 173     }
 174
 175     if (!av_bprint_is_complete(dst))
 176         return AVERROR(ENOMEM);
 177
 178     while (dst->len >= 2 && !strncmp(&dst->str[dst->len - 2], "\\N", 2))
 179         dst->len -= 2;
 180     dst->str[dst->len] = 0;
 181     rstrip_spaces_buf(dst);
 182
 183     return 0;
 184 }