diff options
Diffstat (limited to 'src/plugins/thumbnailffmpeg_extractor.c')
-rw-r--r-- | src/plugins/thumbnailffmpeg_extractor.c | 536 |
1 files changed, 536 insertions, 0 deletions
diff --git a/src/plugins/thumbnailffmpeg_extractor.c b/src/plugins/thumbnailffmpeg_extractor.c new file mode 100644 index 0000000..7fbde2d --- /dev/null +++ b/src/plugins/thumbnailffmpeg_extractor.c | |||
@@ -0,0 +1,536 @@ | |||
1 | /* | ||
2 | This file is part of libextractor. | ||
3 | Copyright (C) 2008 Heikki Lindholm | ||
4 | |||
5 | libextractor is free software; you can redistribute it and/or modify | ||
6 | it under the terms of the GNU General Public License as published | ||
7 | by the Free Software Foundation; either version 2, or (at your | ||
8 | option) any later version. | ||
9 | |||
10 | libextractor is distributed in the hope that it will be useful, but | ||
11 | WITHOUT ANY WARRANTY; without even the implied warranty of | ||
12 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | ||
13 | General Public License for more details. | ||
14 | |||
15 | You should have received a copy of the GNU General Public License | ||
16 | along with libextractor; see the file COPYING. If not, write to the | ||
17 | Free Software Foundation, Inc., 59 Temple Place - Suite 330, | ||
18 | Boston, MA 02111-1307, USA. | ||
19 | */ | ||
20 | |||
21 | /** | ||
22 | * @file thumbnailffmpeg_extractor.c | ||
23 | * @author Heikki Lindholm | ||
24 | * @brief this extractor produces a binary encoded | ||
25 | * thumbnail of images and videos using the ffmpeg libs. | ||
26 | */ | ||
27 | |||
28 | /* This is a thumbnail extractor using the ffmpeg libraries that will eventually | ||
29 | support extracting thumbnails from both image and video files. | ||
30 | |||
31 | Note that ffmpeg has a few issues: | ||
32 | (1) there are no recent official releases of the ffmpeg libs | ||
33 | (2) ffmpeg has a history of having security issues (parser is not robust) | ||
34 | |||
35 | So this plugin cannot be recommended for system with high security | ||
36 | requirements. | ||
37 | */ | ||
38 | |||
39 | #include "platform.h" | ||
40 | #include "extractor.h" | ||
41 | #include <libavformat/avformat.h> | ||
42 | #include <libavcodec/avcodec.h> | ||
43 | #include <libswscale/swscale.h> | ||
44 | |||
45 | #define DEBUG 0 | ||
46 | |||
47 | struct StreamDescriptor | ||
48 | { | ||
49 | const uint8_t *data; | ||
50 | size_t offset; | ||
51 | size_t size; | ||
52 | }; | ||
53 | |||
54 | |||
55 | void __attribute__ ((constructor)) ffmpeg_lib_init (void) | ||
56 | { | ||
57 | #if DEBUG | ||
58 | printf ("av_register_all()\n"); | ||
59 | #endif | ||
60 | av_register_all (); | ||
61 | } | ||
62 | |||
63 | static int | ||
64 | stream_read (void *opaque, uint8_t * buf, int buf_size) | ||
65 | { | ||
66 | struct StreamDescriptor *rs = (struct StreamDescriptor *) opaque; | ||
67 | size_t len; | ||
68 | #if DEBUG | ||
69 | printf ("read_packet: %zu\n", buf_size); | ||
70 | #endif | ||
71 | if (rs) | ||
72 | { | ||
73 | if (rs->data == NULL) | ||
74 | return -1; | ||
75 | if (rs->offset >= rs->size) | ||
76 | return 0; | ||
77 | len = buf_size; | ||
78 | if (rs->offset + len > rs->size) | ||
79 | len = rs->size - rs->offset; | ||
80 | |||
81 | memcpy (buf, rs->data + rs->offset, len); | ||
82 | rs->offset += len; | ||
83 | #if DEBUG | ||
84 | printf ("read_packet: len: %zu\n", len); | ||
85 | #endif | ||
86 | return len; | ||
87 | } | ||
88 | return -1; | ||
89 | } | ||
90 | |||
91 | static offset_t | ||
92 | stream_seek (void *opaque, offset_t offset, int whence) | ||
93 | { | ||
94 | struct StreamDescriptor *rs = (struct StreamDescriptor *) opaque; | ||
95 | offset_t off_abs; | ||
96 | #if DEBUG | ||
97 | printf ("my_seek: %lld %d\n", offset, whence); | ||
98 | #endif | ||
99 | if (rs) | ||
100 | { | ||
101 | if (whence == AVSEEK_SIZE) | ||
102 | return (offset_t) rs->size; | ||
103 | else if (whence == SEEK_CUR) | ||
104 | off_abs = (offset_t) rs->offset + offset; | ||
105 | else if (whence == SEEK_SET) | ||
106 | off_abs = offset; | ||
107 | else if (whence == SEEK_END) | ||
108 | off_abs = (offset_t) rs->size + offset; | ||
109 | else | ||
110 | { | ||
111 | printf ("whence error %d\n", whence); | ||
112 | abort (); | ||
113 | return AVERROR (EINVAL); | ||
114 | } | ||
115 | if (off_abs >= 0 && off_abs < (offset_t) rs->size) | ||
116 | rs->offset = (size_t) off_abs; | ||
117 | else | ||
118 | off_abs = AVERROR (EINVAL); | ||
119 | return off_abs; | ||
120 | } | ||
121 | return -1; | ||
122 | } | ||
123 | |||
124 | static EXTRACTOR_KeywordList * | ||
125 | addKeyword (EXTRACTOR_KeywordType type, | ||
126 | char *keyword, EXTRACTOR_KeywordList * next) | ||
127 | { | ||
128 | EXTRACTOR_KeywordList *result; | ||
129 | |||
130 | if (keyword == NULL) | ||
131 | return next; | ||
132 | result = malloc (sizeof (EXTRACTOR_KeywordList)); | ||
133 | result->next = next; | ||
134 | result->keyword = keyword; | ||
135 | result->keywordType = type; | ||
136 | return result; | ||
137 | } | ||
138 | |||
139 | |||
140 | struct MimeToDecoderMapping | ||
141 | { | ||
142 | const char *mime_type; | ||
143 | enum CodecID codec_id; | ||
144 | }; | ||
145 | |||
146 | /* map mime image types to a decoder */ | ||
147 | static const struct MimeToDecoderMapping m2d_map[] = { | ||
148 | {"image/x-bmp", CODEC_ID_BMP}, | ||
149 | {"image/gif", CODEC_ID_GIF}, | ||
150 | {"image/jpeg", CODEC_ID_MJPEG}, | ||
151 | {"image/png", CODEC_ID_PNG}, | ||
152 | {"image/x-portable-pixmap", CODEC_ID_PPM}, | ||
153 | {NULL, CODEC_ID_NONE} | ||
154 | }; | ||
155 | |||
156 | #define PROBE_MAX (1<<20) | ||
157 | #define BIOBUF_SIZE (64*1024) | ||
158 | #define THUMBSIZE 128 /* max dimension in pixels */ | ||
159 | #define MAX_THUMB_SIZE (100*1024) /* in bytes */ | ||
160 | |||
161 | int | ||
162 | EXTRACTOR_thumbnailffmpeg_extract (const unsigned char *data, | ||
163 | size_t size, | ||
164 | EXTRACTOR_MetaDataProcessor proc, | ||
165 | void *proc_cls, | ||
166 | const char *options) | ||
167 | { | ||
168 | int score; | ||
169 | AVInputFormat *fmt; | ||
170 | AVProbeData pdat; | ||
171 | ByteIOContext *bio_ctx = NULL; | ||
172 | uint8_t *bio_buffer; | ||
173 | struct StreamDescriptor reader_state; | ||
174 | AVFormatContext *format_ctx = NULL; | ||
175 | AVCodecContext *codec_ctx = NULL; | ||
176 | AVPacket packet; | ||
177 | int video_stream_index; | ||
178 | AVCodec *codec; | ||
179 | AVFrame *frame = NULL; | ||
180 | AVFrame *thumb_frame = NULL; | ||
181 | int64_t ts; | ||
182 | |||
183 | struct SwsContext *scaler_ctx; | ||
184 | int sws_flags = SWS_BILINEAR; | ||
185 | uint8_t *thumb_buffer; | ||
186 | int thumb_width, thumb_height; | ||
187 | int sar_num, sar_den; | ||
188 | |||
189 | uint8_t *encoder_output_buffer; | ||
190 | size_t encoder_output_buffer_size; | ||
191 | AVCodecContext *enc_codec_ctx; | ||
192 | AVCodec *enc_codec; | ||
193 | |||
194 | int i; | ||
195 | int err; | ||
196 | int frame_finished; | ||
197 | |||
198 | char *binary; | ||
199 | const char *mime; | ||
200 | int is_image; | ||
201 | enum CodecID image_codec_id; | ||
202 | |||
203 | bio_ctx = NULL; | ||
204 | bio_buffer = NULL; | ||
205 | format_ctx = NULL; | ||
206 | codec = NULL; | ||
207 | frame = NULL; | ||
208 | thumb_frame = NULL; | ||
209 | thumb_buffer = NULL; | ||
210 | scaler_ctx = NULL; | ||
211 | encoder_output_buffer = NULL; | ||
212 | enc_codec = NULL; | ||
213 | enc_codec_ctx = NULL; | ||
214 | |||
215 | is_image = 0; | ||
216 | |||
217 | mime = EXTRACTOR_extractLast (EXTRACTOR_MIMETYPE, prev); | ||
218 | if (mime != NULL) | ||
219 | { | ||
220 | i = 0; | ||
221 | while (m2d_map[i].mime_type != NULL) | ||
222 | { | ||
223 | if (!strcmp (m2d_map[i].mime_type, mime)) | ||
224 | { | ||
225 | is_image = 1; | ||
226 | image_codec_id = m2d_map[i].codec_id; | ||
227 | break; | ||
228 | } | ||
229 | i++; | ||
230 | } | ||
231 | } | ||
232 | |||
233 | #if DEBUG | ||
234 | printf ("is_image: %d codec:%d\n", is_image, image_codec_id); | ||
235 | #endif | ||
236 | if (!is_image) | ||
237 | { | ||
238 | pdat.filename = filename; | ||
239 | pdat.buf = (unsigned char *) data; | ||
240 | pdat.buf_size = (size > PROBE_MAX) ? PROBE_MAX : size; | ||
241 | |||
242 | fmt = av_probe_input_format (&pdat, 1); | ||
243 | if (fmt == NULL) | ||
244 | return prev; | ||
245 | #if DEBUG | ||
246 | printf ("format %p [%s] [%s]\n", fmt, fmt->name, fmt->long_name); | ||
247 | #endif | ||
248 | pdat.buf = (unsigned char *) data; | ||
249 | pdat.buf_size = size > PROBE_MAX ? PROBE_MAX : size; | ||
250 | score = fmt->read_probe (&pdat); | ||
251 | #if DEBUG | ||
252 | printf ("score: %d\n", score); | ||
253 | #endif | ||
254 | /*if (score < 50) return prev; */ | ||
255 | } | ||
256 | |||
257 | if (is_image) | ||
258 | { | ||
259 | codec_ctx = avcodec_alloc_context (); | ||
260 | codec = avcodec_find_decoder (image_codec_id); | ||
261 | if (codec != NULL) | ||
262 | { | ||
263 | if (avcodec_open (codec_ctx, codec) != 0) | ||
264 | { | ||
265 | #if DEBUG | ||
266 | printf ("open codec failed\n"); | ||
267 | #endif | ||
268 | codec = NULL; | ||
269 | } | ||
270 | } | ||
271 | } | ||
272 | else | ||
273 | { | ||
274 | bio_ctx = malloc (sizeof (ByteIOContext)); | ||
275 | bio_buffer = malloc (BIOBUF_SIZE); | ||
276 | |||
277 | reader_state.data = data; | ||
278 | reader_state.offset = 0; | ||
279 | reader_state.size = size; | ||
280 | |||
281 | init_put_byte (bio_ctx, bio_buffer, | ||
282 | BIOBUF_SIZE, 0, &reader_state, | ||
283 | stream_read, NULL, stream_seek); | ||
284 | |||
285 | fmt->flags |= AVFMT_NOFILE; | ||
286 | err = av_open_input_stream (&format_ctx, bio_ctx, "", fmt, NULL); | ||
287 | if (err < 0) | ||
288 | { | ||
289 | #if DEBUG | ||
290 | printf ("couldn't open input stream\n"); | ||
291 | #endif | ||
292 | goto out; | ||
293 | } | ||
294 | |||
295 | err = av_find_stream_info (format_ctx); | ||
296 | if (err < 0) | ||
297 | { | ||
298 | #if DEBUG | ||
299 | printf ("couldn't find codec params\n"); | ||
300 | #endif | ||
301 | goto out; | ||
302 | } | ||
303 | |||
304 | for (i = 0; i < format_ctx->nb_streams; i++) | ||
305 | { | ||
306 | codec_ctx = format_ctx->streams[i]->codec; | ||
307 | if (codec_ctx->codec_type == CODEC_TYPE_VIDEO) | ||
308 | { | ||
309 | video_stream_index = i; | ||
310 | codec = avcodec_find_decoder (codec_ctx->codec_id); | ||
311 | if (codec == NULL) | ||
312 | { | ||
313 | #if DEBUG | ||
314 | printf ("find_decoder failed\n"); | ||
315 | #endif | ||
316 | break; | ||
317 | } | ||
318 | err = avcodec_open (codec_ctx, codec); | ||
319 | if (err != 0) | ||
320 | { | ||
321 | #if DEBUG | ||
322 | printf ("failed to open codec\n"); | ||
323 | #endif | ||
324 | codec = NULL; | ||
325 | } | ||
326 | break; | ||
327 | } | ||
328 | } | ||
329 | } | ||
330 | |||
331 | if (codec_ctx == NULL || codec == NULL) | ||
332 | { | ||
333 | #if DEBUG | ||
334 | printf ("failed to open codec"); | ||
335 | #endif | ||
336 | goto out; | ||
337 | } | ||
338 | frame = avcodec_alloc_frame (); | ||
339 | if (frame == NULL) | ||
340 | { | ||
341 | #if DEBUG | ||
342 | printf ("failed to alloc frame"); | ||
343 | #endif | ||
344 | goto out; | ||
345 | } | ||
346 | |||
347 | if (!is_image) | ||
348 | { | ||
349 | #if DEBUG | ||
350 | printf ("duration: %lld\n", format_ctx->duration); | ||
351 | if (format_ctx->duration == AV_NOPTS_VALUE) | ||
352 | printf ("duration unknown\n"); | ||
353 | #endif | ||
354 | /* TODO: if duration is known seek to to some better place(?) */ | ||
355 | ts = 10; // s | ||
356 | ts = ts * AV_TIME_BASE; | ||
357 | err = av_seek_frame (format_ctx, -1, ts, 0); | ||
358 | if (err >= 0) | ||
359 | { | ||
360 | avcodec_flush_buffers (codec_ctx); | ||
361 | } | ||
362 | #if DEBUG | ||
363 | else | ||
364 | printf ("seeking failed %d\n", err); | ||
365 | #endif | ||
366 | } | ||
367 | |||
368 | frame_finished = 0; | ||
369 | if (is_image) | ||
370 | { | ||
371 | avcodec_decode_video (codec_ctx, frame, &frame_finished, data, size); | ||
372 | } | ||
373 | else | ||
374 | { | ||
375 | while (1) | ||
376 | { | ||
377 | err = av_read_frame (format_ctx, &packet); | ||
378 | if (err < 0) | ||
379 | break; | ||
380 | if (packet.stream_index == video_stream_index) | ||
381 | { | ||
382 | avcodec_decode_video (codec_ctx, | ||
383 | frame, | ||
384 | &frame_finished, | ||
385 | packet.data, packet.size); | ||
386 | if (frame_finished && frame->key_frame) | ||
387 | { | ||
388 | av_free_packet (&packet); | ||
389 | break; | ||
390 | } | ||
391 | } | ||
392 | av_free_packet (&packet); | ||
393 | } | ||
394 | } | ||
395 | |||
396 | if (!frame_finished || codec_ctx->width == 0 || codec_ctx->height == 0) | ||
397 | goto out; | ||
398 | |||
399 | sar_num = codec_ctx->sample_aspect_ratio.num; | ||
400 | sar_den = codec_ctx->sample_aspect_ratio.den; | ||
401 | if (sar_num <= 0 || sar_den <= 0) | ||
402 | { | ||
403 | sar_num = 1; | ||
404 | sar_den = 1; | ||
405 | } | ||
406 | if ((codec_ctx->width * sar_num) / sar_den > codec_ctx->height) | ||
407 | { | ||
408 | thumb_width = THUMBSIZE; | ||
409 | thumb_height = (thumb_width * codec_ctx->height) / | ||
410 | ((codec_ctx->width * sar_num) / sar_den); | ||
411 | } | ||
412 | else | ||
413 | { | ||
414 | thumb_height = THUMBSIZE; | ||
415 | thumb_width = (thumb_height * | ||
416 | ((codec_ctx->width * sar_num) / sar_den)) / | ||
417 | codec_ctx->height; | ||
418 | } | ||
419 | if (thumb_width < 8) | ||
420 | thumb_width = 8; | ||
421 | if (thumb_height < 1) | ||
422 | thumb_height = 1; | ||
423 | #if DEBUG | ||
424 | printf ("thumb dim: %d %d\n", thumb_width, thumb_height); | ||
425 | #endif | ||
426 | |||
427 | scaler_ctx = | ||
428 | sws_getContext (codec_ctx->width, codec_ctx->height, codec_ctx->pix_fmt, | ||
429 | thumb_width, thumb_height, PIX_FMT_RGB24, sws_flags, NULL, | ||
430 | NULL, NULL); | ||
431 | if (scaler_ctx == NULL) | ||
432 | { | ||
433 | #if DEBUG | ||
434 | printf ("failed to alloc scaler\n"); | ||
435 | #endif | ||
436 | goto out; | ||
437 | } | ||
438 | thumb_frame = avcodec_alloc_frame (); | ||
439 | thumb_buffer = | ||
440 | av_malloc (avpicture_get_size (PIX_FMT_RGB24, thumb_width, thumb_height)); | ||
441 | if (thumb_frame == NULL || thumb_buffer == NULL) | ||
442 | { | ||
443 | #if DEBUG | ||
444 | printf ("failed to alloc thumb frame\n"); | ||
445 | #endif | ||
446 | goto out; | ||
447 | } | ||
448 | avpicture_fill ((AVPicture *) thumb_frame, thumb_buffer, | ||
449 | PIX_FMT_RGB24, thumb_width, thumb_height); | ||
450 | |||
451 | sws_scale (scaler_ctx, | ||
452 | frame->data, frame->linesize, | ||
453 | 0, codec_ctx->height, thumb_frame->data, thumb_frame->linesize); | ||
454 | |||
455 | encoder_output_buffer_size = MAX_THUMB_SIZE; | ||
456 | encoder_output_buffer = av_malloc (encoder_output_buffer_size); | ||
457 | if (encoder_output_buffer == NULL) | ||
458 | { | ||
459 | #if DEBUG | ||
460 | printf ("couldn't alloc encoder output buf\n"); | ||
461 | #endif | ||
462 | goto out; | ||
463 | } | ||
464 | |||
465 | enc_codec = avcodec_find_encoder_by_name ("png"); | ||
466 | if (enc_codec == NULL) | ||
467 | { | ||
468 | #if DEBUG | ||
469 | printf ("couldn't find encoder\n"); | ||
470 | #endif | ||
471 | goto out; | ||
472 | } | ||
473 | enc_codec_ctx = avcodec_alloc_context (); | ||
474 | enc_codec_ctx->width = thumb_width; | ||
475 | enc_codec_ctx->height = thumb_height; | ||
476 | enc_codec_ctx->pix_fmt = PIX_FMT_RGB24; | ||
477 | |||
478 | if (avcodec_open (enc_codec_ctx, enc_codec) < 0) | ||
479 | { | ||
480 | #if DEBUG | ||
481 | printf ("couldn't open encoder\n"); | ||
482 | #endif | ||
483 | enc_codec = NULL; | ||
484 | goto out; | ||
485 | } | ||
486 | |||
487 | err = avcodec_encode_video (enc_codec_ctx, | ||
488 | encoder_output_buffer, | ||
489 | encoder_output_buffer_size, thumb_frame); | ||
490 | if (err <= 0) | ||
491 | goto out; | ||
492 | |||
493 | binary = | ||
494 | EXTRACTOR_binaryEncode ((const unsigned char *) encoder_output_buffer, | ||
495 | err); | ||
496 | if (binary != NULL) | ||
497 | prev = addKeyword (EXTRACTOR_THUMBNAIL_DATA, binary, prev); | ||
498 | |||
499 | out: | ||
500 | if (enc_codec != NULL) | ||
501 | avcodec_close (enc_codec_ctx); | ||
502 | if (enc_codec_ctx != NULL) | ||
503 | av_free (enc_codec_ctx); | ||
504 | if (encoder_output_buffer != NULL) | ||
505 | av_free (encoder_output_buffer); | ||
506 | if (scaler_ctx != NULL) | ||
507 | sws_freeContext(scaler_ctx); | ||
508 | if (codec != NULL) | ||
509 | avcodec_close (codec_ctx); | ||
510 | if (format_ctx != NULL) | ||
511 | av_close_input_file (format_ctx); | ||
512 | if (frame != NULL) | ||
513 | av_free (frame); | ||
514 | if (thumb_buffer != NULL) | ||
515 | av_free (thumb_buffer); | ||
516 | if (thumb_frame != NULL) | ||
517 | av_free (thumb_frame); | ||
518 | if (bio_ctx != NULL) | ||
519 | free (bio_ctx); | ||
520 | if (bio_buffer != NULL) | ||
521 | free (bio_buffer); | ||
522 | |||
523 | return prev; | ||
524 | } | ||
525 | |||
526 | int | ||
527 | EXTRACTOR_thumbnail_extract (const unsigned char *data, | ||
528 | size_t size, | ||
529 | EXTRACTOR_MetaDataProcessor proc, | ||
530 | void *proc_cls, | ||
531 | const char *options) | ||
532 | { | ||
533 | return EXTRACTOR_thumbnailffmpeg_extract (data, size, proc, proc_cls, options); | ||
534 | } | ||
535 | |||
536 | /* end of thumbnailffmpeg_extractor.c */ | ||