Mercurial > mplayer.hg
annotate sub/subassconvert.c @ 34912:3f8886645ebd
Order colour names alphabetically / better match W3 list.
Patch by Federico Kereki, fkereki gmail
author | cehoyos |
---|---|
date | Sun, 24 Jun 2012 12:17:03 +0000 |
parents | e301516e3c83 |
children | 69fe59abe044 |
rev | line source |
---|---|
32462 | 1 /* |
2 * Subtitles converter to SSA/ASS in order to allow special formatting | |
3 * | |
4 * This file is part of MPlayer. | |
5 * | |
6 * MPlayer is free software; you can redistribute it and/or modify | |
7 * it under the terms of the GNU General Public License as published by | |
8 * the Free Software Foundation; either version 2 of the License, or | |
9 * (at your option) any later version. | |
10 * | |
11 * MPlayer 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 | |
14 * GNU General Public License for more details. | |
15 * | |
16 * You should have received a copy of the GNU General Public License along | |
17 * with MPlayer; if not, write to the Free Software Foundation, Inc., | |
18 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. | |
19 */ | |
20 | |
21 #include <string.h> | |
22 #include <stdint.h> | |
23 #include <stdlib.h> | |
24 #include <stdio.h> | |
25 #include <stdarg.h> | |
26 | |
27 #include "mp_msg.h" | |
28 #include "help_mp.h" | |
29 #include "bstr.h" | |
32464
22888a8cb312
Do not use a path for including files in the same directory.
reimar
parents:
32462
diff
changeset
|
30 #include "subassconvert.h" |
32462 | 31 #include "libavutil/common.h" |
32 | |
33 struct line { | |
34 char *buf; | |
35 size_t bufsize; | |
36 size_t len; | |
37 }; | |
38 | |
39 #ifdef __GNUC__ | |
40 static void append_text(struct line *dst, char *fmt, ...) __attribute__ ((format(printf, 2, 3))); | |
41 #endif | |
42 | |
43 static void append_text(struct line *dst, char *fmt, ...) | |
44 { | |
45 va_list va; | |
46 int ret; | |
47 | |
48 va_start(va, fmt); | |
49 ret = vsnprintf(dst->buf + dst->len, dst->bufsize - dst->len, fmt, va); | |
50 if (ret >= 0) { | |
51 dst->len += ret; | |
52 if (dst->len > dst->bufsize) | |
53 dst->len = dst->bufsize; | |
54 } | |
55 va_end(va); | |
56 } | |
57 | |
58 static int indexof(const char *s, int c) | |
59 { | |
60 char *f = strchr(s, c); | |
61 return f ? (f - s) : -1; | |
62 } | |
63 | |
64 | |
65 | |
66 /* | |
67 * SubRip | |
68 * | |
69 * Support basic tags (italic, bold, underline, strike-through) | |
70 * and font tag with size, color and face attributes. | |
71 * | |
72 */ | |
73 | |
74 struct font_tag { | |
75 struct bstr face; | |
76 int size; | |
77 uint32_t color; | |
78 }; | |
79 | |
80 static const struct tag_conv { | |
81 const char *from; | |
82 const char *to; | |
83 } subrip_basic_tags[] = { | |
84 {"<i>", "{\\i1}"}, {"</i>", "{\\i0}"}, | |
85 {"<b>", "{\\b1}"}, {"</b>", "{\\b0}"}, | |
86 {"<u>", "{\\u1}"}, {"</u>", "{\\u0}"}, | |
87 {"<s>", "{\\s1}"}, {"</s>", "{\\s0}"}, | |
88 {"{", "\\{"}, {"}", "\\}"}, | |
34757 | 89 {"\r\n", "\\N"}, {"\n", "\\N"}, {"\r", "\\N"}, |
32462 | 90 }; |
91 | |
92 static const struct { | |
93 const char *s; | |
94 uint32_t v; | |
95 } subrip_web_colors[] = { | |
34912
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
96 /* Named CSS3 colors in BGR format; a subset of those |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
97 at http://www.w3.org/TR/css3-color/#svg-color */ |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
98 {"aqua", 0xffff00}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
99 {"black", 0x000000}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
100 {"blue", 0xff0000}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
101 {"fuchsia", 0xff00ff}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
102 {"gray", 0x808080}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
103 {"green", 0x008000}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
104 {"lime", 0x00ff00}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
105 {"maroon", 0x000080}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
106 {"navy", 0x800000}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
107 {"olive", 0x008080}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
108 {"purple", 0x800080}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
109 {"red", 0x0000ff}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
110 {"silver", 0xc0c0c0}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
111 {"teal", 0x808000}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
112 {"white", 0xffffff}, |
3f8886645ebd
Order colour names alphabetically / better match W3 list.
cehoyos
parents:
34758
diff
changeset
|
113 {"yellow", 0x00ffff}, |
32462 | 114 }; |
115 | |
116 #define SUBRIP_MAX_STACKED_FONT_TAGS 16 | |
117 #define SUBRIP_FLAG_COLOR 0x01000000 | |
118 | |
119 /** | |
120 * \brief Convert SubRip lines into ASS markup | |
121 * \param orig original SubRip lines. The content will remain untouched. | |
122 * \param dest ASS markup destination buffer. | |
123 * \param dest_buffer_size maximum size for the destination buffer. | |
124 */ | |
125 void subassconvert_subrip(const char *orig, char *dest, size_t dest_buffer_size) | |
126 { | |
127 /* line is not const to avoid warnings with strtol, etc. | |
128 * orig content won't be changed */ | |
129 char *line = (char *)orig; | |
130 struct line new_line = { | |
131 .buf = dest, | |
132 .bufsize = dest_buffer_size, | |
133 }; | |
134 struct font_tag font_stack[SUBRIP_MAX_STACKED_FONT_TAGS]; | |
135 int sp = 0; | |
136 | |
137 font_stack[0] = (struct font_tag){}; // type with all defaults | |
138 while (*line && new_line.len < new_line.bufsize - 1) { | |
139 char *orig_line = line; | |
140 int i; | |
141 | |
142 for (i = 0; i < FF_ARRAY_ELEMS(subrip_basic_tags); i++) { | |
143 const struct tag_conv *tag = &subrip_basic_tags[i]; | |
144 int from_len = strlen(tag->from); | |
145 if (strncmp(line, tag->from, from_len) == 0) { | |
146 append_text(&new_line, "%s", tag->to); | |
147 line += from_len; | |
148 } | |
149 } | |
150 | |
151 if (strncmp(line, "</font>", 7) == 0) { | |
152 /* Closing font tag */ | |
153 line += 7; | |
154 | |
155 if (sp > 0) { | |
156 struct font_tag *tag = &font_stack[sp]; | |
157 struct font_tag *last_tag = &tag[-1]; | |
158 sp--; | |
159 | |
160 if (tag->size) { | |
161 if (!last_tag->size) | |
162 append_text(&new_line, "{\\fs}"); | |
163 else if (last_tag->size != tag->size) | |
164 append_text(&new_line, "{\\fs%d}", last_tag->size); | |
165 } | |
166 | |
167 if (tag->color & SUBRIP_FLAG_COLOR) { | |
168 if (!(last_tag->color & SUBRIP_FLAG_COLOR)) | |
169 append_text(&new_line, "{\\c}"); | |
170 else if (last_tag->color != tag->color) | |
171 append_text(&new_line, "{\\c&H%06X&}", | |
172 last_tag->color & 0xffffff); | |
173 } | |
174 | |
175 if (tag->face.len) { | |
176 if (!last_tag->face.len) | |
177 append_text(&new_line, "{\\fn}"); | |
178 else if (bstrcmp(last_tag->face, tag->face) != 0) | |
179 append_text(&new_line, "{\\fn%.*s}", | |
180 BSTR_P(last_tag->face)); | |
181 } | |
182 } | |
183 } else if (strncmp(line, "<font ", 6) == 0 | |
184 && sp + 1 < FF_ARRAY_ELEMS(font_stack)) { | |
185 /* Opening font tag */ | |
186 char *potential_font_tag_start = line; | |
187 int len_backup = new_line.len; | |
188 struct font_tag *tag = &font_stack[sp + 1]; | |
189 int has_valid_attr = 0; | |
190 | |
191 *tag = tag[-1]; // keep values from previous tag | |
192 line += 6; | |
193 | |
194 while (*line && *line != '>') { | |
195 if (strncmp(line, "size=\"", 6) == 0) { | |
196 line += 6; | |
197 tag->size = strtol(line, &line, 10); | |
198 if (*line != '"' || !tag->size) | |
199 break; | |
200 append_text(&new_line, "{\\fs%d}", tag->size); | |
201 has_valid_attr = 1; | |
202 } else if (strncmp(line, "color=\"", 7) == 0) { | |
203 line += 7; | |
204 if (*line == '#') { | |
205 // #RRGGBB format | |
206 line++; | |
207 tag->color = strtol(line, &line, 16) & 0x00ffffff; | |
208 if (*line != '"') | |
209 break; | |
210 tag->color = ((tag->color & 0xff) << 16) | | |
211 (tag->color & 0xff00) | | |
212 ((tag->color & 0xff0000) >> 16) | | |
213 SUBRIP_FLAG_COLOR; | |
214 } else { | |
215 // Standard web colors | |
216 int i, len = indexof(line, '"'); | |
217 if (len <= 0) | |
218 break; | |
219 for (i = 0; i < FF_ARRAY_ELEMS(subrip_web_colors); i++) { | |
220 const char *color = subrip_web_colors[i].s; | |
221 if (strlen(color) == len | |
222 && strncasecmp(line, color, len) == 0) { | |
223 tag->color = SUBRIP_FLAG_COLOR | subrip_web_colors[i].v; | |
224 break; | |
225 } | |
226 } | |
227 | |
228 if (i == FF_ARRAY_ELEMS(subrip_web_colors)) { | |
229 /* We didn't find any matching color */ | |
230 line = strchr(line, '"'); // can't be NULL, see above | |
231 mp_msg(MSGT_SUBREADER, MSGL_WARN, | |
232 MSGTR_SUBTITLES_SubRip_UnknownFontColor, orig); | |
233 append_text(&new_line, "{\\c}"); | |
234 line += 2; | |
235 continue; | |
236 } | |
237 | |
238 line += len; | |
239 } | |
240 append_text(&new_line, "{\\c&H%06X&}", tag->color & 0xffffff); | |
241 has_valid_attr = 1; | |
242 } else if (strncmp(line, "face=\"", 6) == 0) { | |
243 /* Font face attribute */ | |
244 int len; | |
245 line += 6; | |
246 len = indexof(line, '"'); | |
247 if (len <= 0) | |
248 break; | |
249 tag->face.start = line; | |
250 tag->face.len = len; | |
251 line += len; | |
252 append_text(&new_line, "{\\fn%.*s}", BSTR_P(tag->face)); | |
253 has_valid_attr = 1; | |
254 } | |
255 line++; | |
256 } | |
257 | |
258 if (!has_valid_attr || *line != '>') { /* Not valid font tag */ | |
259 line = potential_font_tag_start; | |
260 new_line.len = len_backup; | |
261 } else { | |
262 sp++; | |
263 line++; | |
264 } | |
265 } | |
266 | |
267 /* Tag conversion code didn't match */ | |
268 if (line == orig_line) | |
269 new_line.buf[new_line.len++] = *line++; | |
270 } | |
271 new_line.buf[new_line.len] = 0; | |
272 } | |
273 | |
274 | |
275 /* | |
276 * MicroDVD | |
277 * | |
278 * Based on the specifications found here: | |
279 * https://trac.videolan.org/vlc/ticket/1825#comment:6 | |
280 */ | |
281 | |
282 struct microdvd_tag { | |
283 char key; | |
284 int persistent; | |
285 uint32_t data1; | |
286 uint32_t data2; | |
287 struct bstr data_string; | |
288 }; | |
289 | |
290 #define MICRODVD_PERSISTENT_OFF 0 | |
291 #define MICRODVD_PERSISTENT_ON 1 | |
292 #define MICRODVD_PERSISTENT_OPENED 2 | |
293 | |
294 // Color, Font, Size, cHarset, stYle, Position, cOordinate | |
295 #define MICRODVD_TAGS "cfshyYpo" | |
296 | |
297 static void microdvd_set_tag(struct microdvd_tag *tags, struct microdvd_tag tag) | |
298 { | |
299 int tag_index = indexof(MICRODVD_TAGS, tag.key); | |
300 | |
301 if (tag_index < 0) | |
302 return; | |
303 memcpy(&tags[tag_index], &tag, sizeof(tag)); | |
304 } | |
305 | |
306 // italic, bold, underline, strike-through | |
307 #define MICRODVD_STYLES "ibus" | |
308 | |
309 static char *microdvd_load_tags(struct microdvd_tag *tags, char *s) | |
310 { | |
311 while (*s == '{') { | |
312 char *start = s; | |
313 char tag_char = *(s + 1); | |
34758 | 314 struct microdvd_tag tag = {0}; |
32462 | 315 |
316 if (!tag_char || *(s + 2) != ':') | |
317 break; | |
318 s += 3; | |
319 | |
320 switch (tag_char) { | |
321 | |
322 /* Style */ | |
323 case 'Y': | |
324 tag.persistent = MICRODVD_PERSISTENT_ON; | |
325 case 'y': | |
326 while (*s && *s != '}') { | |
327 int style_index = indexof(MICRODVD_STYLES, *s); | |
328 | |
329 if (style_index >= 0) | |
330 tag.data1 |= (1 << style_index); | |
331 s++; | |
332 } | |
333 if (*s != '}') | |
334 break; | |
335 /* We must distinguish persistent and non-persistent styles | |
336 * to handle this kind of style tags: {y:ib}{Y:us} */ | |
337 tag.key = tag_char; | |
338 break; | |
339 | |
340 /* Color */ | |
341 case 'C': | |
342 tag.persistent = MICRODVD_PERSISTENT_ON; | |
343 case 'c': | |
344 tag.data1 = strtol(s, &s, 16) & 0x00ffffff; | |
345 if (*s != '}') | |
346 break; | |
347 tag.key = 'c'; | |
348 break; | |
349 | |
350 /* Font name */ | |
351 case 'F': | |
352 tag.persistent = MICRODVD_PERSISTENT_ON; | |
353 case 'f': | |
354 { | |
355 int len = indexof(s, '}'); | |
356 if (len < 0) | |
357 break; | |
358 tag.data_string.start = s; | |
359 tag.data_string.len = len; | |
360 s += len; | |
361 tag.key = 'f'; | |
362 break; | |
363 } | |
364 | |
365 /* Font size */ | |
366 case 'S': | |
367 tag.persistent = MICRODVD_PERSISTENT_ON; | |
368 case 's': | |
369 tag.data1 = strtol(s, &s, 10); | |
370 if (*s != '}') | |
371 break; | |
372 tag.key = 's'; | |
373 break; | |
374 | |
375 /* Charset */ | |
376 case 'H': | |
377 { | |
378 //TODO: not yet handled, just parsed. | |
379 int len = indexof(s, '}'); | |
380 if (len < 0) | |
381 break; | |
382 tag.data_string.start = s; | |
383 tag.data_string.len = len; | |
384 s += len; | |
385 tag.key = 'h'; | |
386 break; | |
387 } | |
388 | |
389 /* Position */ | |
390 case 'P': | |
391 tag.persistent = MICRODVD_PERSISTENT_ON; | |
392 tag.data1 = (*s++ == '1'); | |
393 if (*s != '}') | |
394 break; | |
395 tag.key = 'p'; | |
396 break; | |
397 | |
398 /* Coordinates */ | |
399 case 'o': | |
400 tag.persistent = MICRODVD_PERSISTENT_ON; | |
401 tag.data1 = strtol(s, &s, 10); | |
402 if (*s != ',') | |
403 break; | |
404 s++; | |
405 tag.data2 = strtol(s, &s, 10); | |
406 if (*s != '}') | |
407 break; | |
408 tag.key = 'o'; | |
409 break; | |
410 | |
411 default: /* Unknown tag, we consider it's text */ | |
412 break; | |
413 } | |
414 | |
415 if (tag.key == 0) | |
416 return start; | |
417 | |
418 microdvd_set_tag(tags, tag); | |
419 s++; | |
420 } | |
421 return s; | |
422 } | |
423 | |
424 static void microdvd_open_tags(struct line *new_line, struct microdvd_tag *tags) | |
425 { | |
426 int i, sidx; | |
427 for (i = 0; i < sizeof(MICRODVD_TAGS) - 1; i++) { | |
428 if (tags[i].persistent == MICRODVD_PERSISTENT_OPENED) | |
429 continue; | |
430 switch (tags[i].key) { | |
431 case 'Y': | |
432 case 'y': | |
433 for (sidx = 0; sidx < sizeof(MICRODVD_STYLES) - 1; sidx++) | |
434 if (tags[i].data1 & (1 << sidx)) | |
435 append_text(new_line, "{\\%c1}", MICRODVD_STYLES[sidx]); | |
436 break; | |
437 | |
438 case 'c': | |
439 append_text(new_line, "{\\c&H%06X&}", tags[i].data1); | |
440 break; | |
441 | |
442 case 'f': | |
443 append_text(new_line, "{\\fn%.*s}", BSTR_P(tags[i].data_string)); | |
444 break; | |
445 | |
446 case 's': | |
447 append_text(new_line, "{\\fs%d}", tags[i].data1); | |
448 break; | |
449 | |
450 case 'p': | |
451 if (tags[i].data1 == 0) | |
452 append_text(new_line, "{\\an8}"); | |
453 break; | |
454 | |
455 case 'o': | |
456 append_text(new_line, "{\\pos(%d,%d)}", | |
457 tags[i].data1, tags[i].data2); | |
458 break; | |
459 } | |
460 if (tags[i].persistent == MICRODVD_PERSISTENT_ON) | |
461 tags[i].persistent = MICRODVD_PERSISTENT_OPENED; | |
462 } | |
463 } | |
464 | |
465 static void microdvd_close_no_persistent_tags(struct line *new_line, | |
466 struct microdvd_tag *tags) | |
467 { | |
468 int i, sidx; | |
469 | |
470 for (i = sizeof(MICRODVD_TAGS) - 2; i; i--) { | |
471 if (tags[i].persistent != MICRODVD_PERSISTENT_OFF) | |
472 continue; | |
473 switch (tags[i].key) { | |
474 | |
475 case 'y': | |
476 for (sidx = sizeof(MICRODVD_STYLES) - 2; sidx >= 0; sidx--) | |
477 if (tags[i].data1 & (1 << sidx)) | |
478 append_text(new_line, "{\\%c0}", MICRODVD_STYLES[sidx]); | |
479 break; | |
480 | |
481 case 'c': | |
482 append_text(new_line, "{\\c}"); | |
483 break; | |
484 | |
485 case 'f': | |
486 append_text(new_line, "{\\fn}"); | |
487 break; | |
488 | |
489 case 's': | |
490 append_text(new_line, "{\\fs}"); | |
491 break; | |
492 } | |
493 tags[i].key = 0; | |
494 } | |
495 } | |
496 | |
497 /** | |
498 * \brief Convert MicroDVD lines into ASS markup | |
499 * \param orig original MicroDVD line. The content will remain untouched. | |
500 * \param dest ASS markup destination buffer. | |
501 * \param dest_buffer_size maximum size for the destination buffer. | |
502 */ | |
503 void subassconvert_microdvd(const char *orig, char *dest, size_t dest_buffer_size) | |
504 { | |
505 /* line is not const to avoid warnings with strtol, etc. | |
506 * orig content won't be changed */ | |
507 char *line = (char *)orig; | |
508 struct line new_line = { | |
509 .buf = dest, | |
510 .bufsize = dest_buffer_size, | |
511 }; | |
34758 | 512 struct microdvd_tag tags[sizeof(MICRODVD_TAGS) - 1] = {{0}}; |
32462 | 513 |
514 while (*line) { | |
515 line = microdvd_load_tags(tags, line); | |
516 microdvd_open_tags(&new_line, tags); | |
517 | |
518 while (*line && *line != '|') | |
519 new_line.buf[new_line.len++] = *line++; | |
520 | |
521 if (*line == '|') { | |
522 microdvd_close_no_persistent_tags(&new_line, tags); | |
523 append_text(&new_line, "\\N"); | |
524 line++; | |
525 } | |
526 } | |
527 new_line.buf[new_line.len] = 0; | |
528 } |