VideoMetadata.vala 20.7 KB
Newer Older
1
/* Copyright 2016 Software Freedom Conservancy Inc.
2 3
 *
 * This software is licensed under the GNU Lesser General Public License
4
 * (version 2.1 or later).  See the COPYING file in this distribution.
5 6 7 8
 */

public class VideoMetadata : MediaMetadata {
    
9 10
    private MetadataDateTime timestamp = null;
    private string title = null;
11
    private string comment = null;
12
   
13 14 15 16 17 18 19
    public VideoMetadata() {
    }
    
    ~VideoMetadata() {
    }
    
    public override void read_from_file(File file) throws Error {
20 21 22 23
        QuickTimeMetadataLoader quicktime = new QuickTimeMetadataLoader(file);
        if (quicktime.is_supported()) {
            timestamp = quicktime.get_creation_date_time();
            title = quicktime.get_title();
24
	        // TODO: is there an quicktime.get_comment ??
25
            comment = null;
26
            return;
27 28 29 30 31
        }    
        AVIMetadataLoader avi = new AVIMetadataLoader(file);
        if (avi.is_supported()) {
            timestamp = avi.get_creation_date_time();
            title = avi.get_title();
32
            comment = null;
33
            return;
34 35 36
        }
        
        throw new IOError.NOT_SUPPORTED("File %s is not a supported video format", file.get_path());
37 38 39
    }
    
    public override MetadataDateTime? get_creation_date_time() {
40 41 42 43 44 45 46
        return timestamp;
    }
    
    public override string? get_title() {
        return title;
    }
    
47 48 49 50
    public override string? get_comment() {
        return comment;
    }
    
51 52 53 54 55 56
}

private class QuickTimeMetadataLoader {

    // Quicktime calendar date/time format is number of seconds since January 1, 1904.
    // This converts to UNIX time (66 years + 17 leap days).
57
    public const time_t QUICKTIME_EPOCH_ADJUSTMENT = 2082844800;
58 59 60

    private File file = null;

61
    public QuickTimeMetadataLoader(File file) {
62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79
        this.file = file;
    }
    
    public MetadataDateTime? get_creation_date_time() {
        return new MetadataDateTime((time_t) get_creation_date_time_for_quicktime());
    }
    
    public string? get_title() {
        // Not supported.
        return null;
    }

    // Checks if the given file is a QuickTime file.
    public bool is_supported() {
        QuickTimeAtom test = new QuickTimeAtom(file);
        
        bool ret = false;
        try {
80
            test.open_file();
81 82 83 84
            test.read_atom();
            
            // Look for the header.
            if ("ftyp" == test.get_current_atom_name()) {
85
                ret = true;
86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102
            } else {
                // Some versions of QuickTime don't have
                // an ftyp section, so we'll just look
                // for the mandatory moov section.
                while(true) {
                    if ("moov" == test.get_current_atom_name()) {
                        ret = true;
                        break;
                    }
                    test.next_atom();
                    test.read_atom();
                    if (test.is_last_atom()) {
                        break;
                    }
                }
            }
        } catch (GLib.Error e) {
103
            debug("Error while testing for QuickTime file for %s: %s", file.get_path(), e.message);
104
        }
105
        
106 107 108 109 110
        try {
            test.close_file();
        } catch (GLib.Error e) {
            debug("Error while closing Quicktime file: %s", e.message);
        }
111 112 113 114 115
        return ret;
    }

    private ulong get_creation_date_time_for_quicktime() {
        QuickTimeAtom test = new QuickTimeAtom(file);
116
        time_t timestamp = 0;
117
        
118 119 120 121 122 123 124 125 126 127
        try {
            test.open_file();
            bool done = false;
            while(!done) {
                // Look for "moov" section.
                test.read_atom();
                if (test.is_last_atom()) break;
                if ("moov" == test.get_current_atom_name()) {
                    QuickTimeAtom child = test.get_first_child_atom();
                    while (!done) {
128
                        // Look for "mvhd" section, or break if none is found.
129
                        child.read_atom();
130 131 132 133 134
                        if (child.is_last_atom() || 0 == child.section_size_remaining()) {
                            done = true;
                            break;
                        }
                        
135 136 137 138 139 140 141 142
                        if ("mvhd" == child.get_current_atom_name()) {
                            // Skip 4 bytes (version + flags)
                            child.read_uint32();
                            // Grab the timestamp.
                            timestamp = child.read_uint32() - QUICKTIME_EPOCH_ADJUSTMENT;
                            done = true;
                            break;
                        }
143
                        child.next_atom();
144 145 146 147 148 149 150
                    }
                }
                test.next_atom();
            }
        } catch (GLib.Error e) {
            debug("Error while testing for QuickTime file: %s", e.message);
        }
151 152 153 154 155 156
        
        try {
            test.close_file();
        } catch (GLib.Error e) {
            debug("Error while closing Quicktime file: %s", e.message);
        }
157 158 159 160 161 162 163 164 165 166 167 168 169
        
        // Some Android phones package videos recorded with their internal cameras in a 3GP
        // container that looks suspiciously like a QuickTime container but really isn't -- for
        // the timestamps of these Android 3GP videos are relative to the UNIX epoch
        // (January 1, 1970) instead of the QuickTime epoch (January 1, 1904). So, if we detect a
        // QuickTime movie with a negative timestamp, we can be pretty sure it isn't a valid
        // QuickTime movie that was shot before 1904 but is instead a non-compliant 3GP video
        // file. If we detect such a video, we correct its time. See this Redmine ticket
        // (http://redmine.yorba.org/issues/3314) for more information.
        if (timestamp < 0)
            timestamp += QUICKTIME_EPOCH_ADJUSTMENT;
        
        return (ulong) timestamp;
170 171 172 173 174 175 176 177 178
    }
}

private class QuickTimeAtom {
    private GLib.File file = null;
    private string section_name = "";
    private uint64 section_size = 0;
    private uint64 section_offset = 0;
    private GLib.DataInputStream input = null;
179
    private QuickTimeAtom? parent = null;
180 181 182 183 184
    
    public QuickTimeAtom(GLib.File file) {
        this.file = file;
    }
    
185
    private QuickTimeAtom.with_input_stream(GLib.DataInputStream input, QuickTimeAtom parent) {
186
        this.input = input;
187 188
        this.parent = parent;
    }
189 190 191 192 193 194 195 196 197 198 199 200 201 202 203
    
    public void open_file() throws GLib.Error {
        close_file();
        input = new GLib.DataInputStream(file.read());
        input.set_byte_order(DataStreamByteOrder.BIG_ENDIAN);
        section_size = 0;
        section_offset = 0;
        section_name = "";
    }
    
    public void close_file() throws GLib.Error {
        if (null != input) {
            input.close();
            input = null;
        }
204 205 206 207 208 209 210 211
    }
    
    private void advance_section_offset(uint64 amount) {
        section_offset += amount;
        if (null != parent) {
            parent.advance_section_offset(amount);
        }
    }
212 213 214 215 216 217 218
    
    public QuickTimeAtom get_first_child_atom() {
        // Child will simply have the input stream
        // but not the size/offset.  This works because
        // child atoms follow immediately after a header,
        // so no skipping is required to access the child
        // from the current position.
219
        return new QuickTimeAtom.with_input_stream(input, this);
220 221 222
    }
    
    public uchar read_byte() throws GLib.Error {
223
        advance_section_offset(1);
224 225 226 227
        return input.read_byte();
    }
    
    public uint32 read_uint32() throws GLib.Error {
228
        advance_section_offset(4);
229 230 231 232
        return input.read_uint32();
    }
    
    public uint64 read_uint64() throws GLib.Error {
233
        advance_section_offset(8);
234 235 236 237 238 239
        return input.read_uint64();
    }

    public void read_atom() throws GLib.Error {
        // Read atom size.
        section_size = read_uint32();
240
        
241 242 243 244 245 246 247
        // Read atom name.
        GLib.StringBuilder sb = new GLib.StringBuilder();
        sb.append_c((char) read_byte());
        sb.append_c((char) read_byte());
        sb.append_c((char) read_byte());
        sb.append_c((char) read_byte());
        section_name = sb.str;
248
        
249 250 251 252 253 254 255 256 257 258 259
        // Check string.
        if (section_name.length != 4) {
            throw new IOError.NOT_SUPPORTED("QuickTime atom name length is invalid for %s", 
                file.get_path());
        }
        for (int i = 0; i < section_name.length; i++) {
            if (!section_name[i].isprint()) {
                throw new IOError.NOT_SUPPORTED("Bad QuickTime atom in file %s", file.get_path());
            }
        }
        
260 261 262 263 264
        if (1 == section_size) {
            // This indicates the section size is a 64-bit
            // value, specified below the atom name.
            section_size = read_uint64();
        }
265 266
    }
    
267 268 269 270 271
    private void skip(uint64 skip_amount) throws GLib.Error {
        skip_uint64(input, skip_amount);
    }
    
    public uint64 section_size_remaining() {
272
        assert(section_size >= section_offset);
273 274 275 276 277
        return section_size - section_offset;
    }
    
    public void next_atom() throws GLib.Error {
        skip(section_size_remaining());
278 279 280 281 282 283
        section_size = 0;
        section_offset = 0;
    }
    
    public string get_current_atom_name() {
        return section_name;
284
    }
285 286 287 288 289
   
    public bool is_last_atom() {
        return 0 == section_size;
    }
    
290
}
291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432

private class AVIMetadataLoader {

    private File file = null;
    
    // A numerical date string, i.e 2010:01:28 14:54:25
    private const int NUMERICAL_DATE_LENGTH = 19;
    
    // Marker for timestamp section in a Nikon nctg blob.
    private const uint16 NIKON_NCTG_TIMESTAMP_MARKER = 0x13;
    
    // Size limit to ensure we don't parse forever on a bad file.
    private const int MAX_STRD_LENGTH = 100;

    public AVIMetadataLoader(File file) {
        this.file = file;
    }
    
    public MetadataDateTime? get_creation_date_time() {
        return new MetadataDateTime((time_t) get_creation_date_time_for_avi());
    }
    
    public string? get_title() {
        // Not supported.
        return null;
    }

    // Checks if the given file is an AVI file.
    public bool is_supported() {
        AVIChunk chunk = new AVIChunk(file);
        bool ret = false;
        try {
            chunk.open_file();
            chunk.read_chunk();
            // Look for the header and identifier.
            if ("RIFF" == chunk.get_current_chunk_name() &&
                "AVI " == chunk.read_name()) {
                ret = true;
            }
        } catch (GLib.Error e) {
            debug("Error while testing for AVI file: %s", e.message);
        }
        
        try {
            chunk.close_file();
        } catch (GLib.Error e) {
            debug("Error while closing AVI file: %s", e.message);
        } 
        return ret;
    }
    
    // Parses a Nikon nctg tag.  Based losely on avi_read_nikon() in FFmpeg.
    private string read_nikon_nctg_tag(AVIChunk chunk) throws GLib.Error {
        bool found_date = false;
        while (chunk.section_size_remaining() > sizeof(uint16)*2) {
            uint16 tag = chunk.read_uint16();
            uint16 size = chunk.read_uint16();
            if (NIKON_NCTG_TIMESTAMP_MARKER == tag) {
                found_date = true;
                break;
            }
            chunk.skip(size);
        }
        
        if (found_date) {
            // Read numerical date string, example: 2010:01:28 14:54:25
            GLib.StringBuilder sb = new GLib.StringBuilder();
            for (int i = 0; i < NUMERICAL_DATE_LENGTH; i++) {
                sb.append_c((char) chunk.read_byte());
            }
            return sb.str;
        }
        return "";
    }
    
    // Parses a Fujifilm strd tag. Based on information from:
    // http://www.eden-foundation.org/products/code/film_date_stamp/index.html
    private string read_fuji_strd_tag(AVIChunk chunk) throws GLib.Error {
        chunk.skip(98); // Ignore 98-byte binary blob.
        chunk.skip(8); // Ignore the string "FUJIFILM"
        // Read until we find four colons, then two more chars.
        int colons = 0;
        int post_colons = 0;
        GLib.StringBuilder sb = new GLib.StringBuilder();
        // End of date is two chars past the fourth colon.
        while (colons <= 4 && post_colons < 2) {
            char c = (char) chunk.read_byte();
            if (4 == colons) {
                post_colons++;
            }
            if (':' == c) {
                colons++;
            }
            if (c.isprint()) {
                sb.append_c(c);
            }
            if (sb.len > MAX_STRD_LENGTH) {
                return ""; // Give up searching.
            }
        }
        
        if (sb.str.length < NUMERICAL_DATE_LENGTH) {
            return ""; 
        }
        // Date is now at the end of the string.
        return sb.str.substring(sb.str.length - NUMERICAL_DATE_LENGTH);
    }
    
    // Recursively read file until the section is found.
    private string? read_section(AVIChunk chunk) throws GLib.Error {
        while (true) {
            chunk.read_chunk();
            string name = chunk.get_current_chunk_name();
            if ("IDIT" == name) {
                return chunk.section_to_string();
            } else if ("nctg" == name) {
                return read_nikon_nctg_tag(chunk);
            } else if ("strd" == name) {
                return read_fuji_strd_tag(chunk);
            }
            
            if ("LIST" == name) {
                chunk.read_name(); // Read past list name.
                string result = read_section(chunk.get_first_child_chunk());
                if (null != result) {
                    return result;
                }
            }
            
            if (chunk.is_last_chunk()) {
                break;
            }
            chunk.next_chunk();
        }
        return null;
    }
    
    // Parses a date from a string.
    // Largely based on GStreamer's avi/gstavidemux.c 
    // and the information here: 
    // http://www.eden-foundation.org/products/code/film_date_stamp/index.html
    private ulong parse_date(string sdate) {
433
        if (sdate.length == 0) {
434 435
            return 0;
        }
436
        
437 438
        Date date = Date();
        uint seconds = 0;
439
        int year, month, day, hour, min, sec;
440 441 442
        char weekday[4];
        char monthstr[4];
        
443
        if (sdate[0].isdigit()) {
444 445
            // Format is: 2005:08:17 11:42:43
            // Format is: 2010/11/30/ 19:42
446 447 448 449 450 451 452
            // Format is: 2010/11/30 19:42
            string tmp = sdate.dup();
            tmp.canon("0123456789 ", ' '); // strip everything but numbers and spaces
            sec = 0;
            int result = tmp.scanf("%d %d %d %d %d %d", out year, out month, out day, out hour, out min, out sec);
            if(result < 5) {
                return 0;
453 454
            }
            date.set_dmy((DateDay) day, (DateMonth) month, (DateYear) year);
455
            seconds = sec + min * 60 + hour * 3600;
456 457
        } else {
            // Format is: Mon Mar  3 09:44:56 2008
458 459
            if(7 != sdate.scanf("%3s %3s %d %d:%d:%d %d", weekday, monthstr, out day, out hour,
                  out min, out sec, out year)) {
460 461
                return 0; // Error
            }
462
            date.set_dmy((DateDay) day, month_from_string((string) monthstr), (DateYear) year);
463 464 465 466 467
            seconds = sec + min * 60 + hour * 3600;
        }
        
        Time time = Time();
        date.to_time(out time);
468 469 470 471 472 473 474 475 476 477 478
        
        // watch for overflow (happens on quasi-bogus dates, like Year 200)
        time_t tm = time.mktime();
        ulong result = tm + seconds;
        if (result < tm) {
            debug("Overflow for timestamp in video file %s", file.get_path());
            
            return 0;
        }
        
        return result;
479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515
    }
    
    private DateMonth month_from_string(string s) {
        switch (s.down()) {
        case "jan":
            return DateMonth.JANUARY;
        case "feb":
            return DateMonth.FEBRUARY;
        case "mar":
            return DateMonth.MARCH;
        case "apr":
            return DateMonth.APRIL;
        case "may":
            return DateMonth.MAY;
        case "jun":
            return DateMonth.JUNE;
        case "jul":
            return DateMonth.JULY;
        case "aug":
            return DateMonth.AUGUST;
        case "sep":
            return DateMonth.SEPTEMBER;
        case "oct":
            return DateMonth.OCTOBER;
        case "nov":
            return DateMonth.NOVEMBER;
        case "dec":
            return DateMonth.DECEMBER;
        }
        return DateMonth.BAD_MONTH;
    }

    private ulong get_creation_date_time_for_avi() {
        AVIChunk chunk = new AVIChunk(file);
        ulong timestamp = 0;
        try {
            chunk.open_file();
516
            chunk.nonsection_skip(12); // Advance past 12 byte header.
517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567
            string sdate = read_section(chunk);
            if (null != sdate) {
                timestamp = parse_date(sdate.strip());
            }
        } catch (GLib.Error e) {
            debug("Error while reading AVI file: %s", e.message);
        }
        
        try {
            chunk.close_file();
        } catch (GLib.Error e) {
            debug("Error while closing AVI file: %s", e.message);
        } 
        return timestamp;
    }
}

private class AVIChunk {
    private GLib.File file = null;
    private string section_name = "";
    private uint64 section_size = 0;
    private uint64 section_offset = 0;
    private GLib.DataInputStream input = null;
    private AVIChunk? parent = null;
    private const int MAX_STRING_TO_SECTION_LENGTH = 1024;
    
    public AVIChunk(GLib.File file) {
        this.file = file;
    }
    
    private AVIChunk.with_input_stream(GLib.DataInputStream input, AVIChunk parent) {
        this.input = input;
        this.parent = parent;
    }   
    
    public void open_file() throws GLib.Error {
        close_file();
        input = new GLib.DataInputStream(file.read());
        input.set_byte_order(DataStreamByteOrder.LITTLE_ENDIAN);
        section_size = 0;
        section_offset = 0;
        section_name = "";
    }
    
    public void close_file() throws GLib.Error {
        if (null != input) {
            input.close();
            input = null;
        }
    }
    
568 569 570 571
    public void nonsection_skip(uint64 skip_amount) throws GLib.Error {
        skip_uint64(input, skip_amount);
    }
    
572 573
    public void skip(uint64 skip_amount) throws GLib.Error {
        advance_section_offset(skip_amount);
574
        skip_uint64(input, skip_amount);
575 576 577 578 579 580
    }
    
    public AVIChunk get_first_child_chunk() {
        return new AVIChunk.with_input_stream(input, this);
    }
    
581 582 583
    private void advance_section_offset(uint64 amount) {
        if ((section_offset + amount) > section_size)
            amount = section_size - section_offset;
584
        
585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601
        section_offset += amount;
        if (null != parent) {
            parent.advance_section_offset(amount);
        }
    }
    
    public uchar read_byte() throws GLib.Error {
        advance_section_offset(1);
        return input.read_byte();
    }
    
    public uint16 read_uint16() throws GLib.Error {
       advance_section_offset(2);
       return input.read_uint16();
    }
    
    public void read_chunk() throws GLib.Error {
602 603 604 605 606 607 608 609 610
        // don't use checked reads here because they advance the section offset, which we're trying
        // to determine here
        GLib.StringBuilder sb = new GLib.StringBuilder();
        sb.append_c((char) input.read_byte());
        sb.append_c((char) input.read_byte());
        sb.append_c((char) input.read_byte());
        sb.append_c((char) input.read_byte());
        section_name = sb.str;
        section_size = input.read_uint32();
611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655
        section_offset = 0;
    }
    
    public string read_name() throws GLib.Error {
        GLib.StringBuilder sb = new GLib.StringBuilder();
        sb.append_c((char) read_byte());
        sb.append_c((char) read_byte());
        sb.append_c((char) read_byte());
        sb.append_c((char) read_byte());
        return sb.str;
    }
    
    public void next_chunk() throws GLib.Error {
        skip(section_size_remaining());
        section_size = 0;
        section_offset = 0;
    }
    
    public string get_current_chunk_name() {
        return section_name;
    }
   
    public bool is_last_chunk() {
        return section_size == 0;
    }
    
    public uint64 section_size_remaining() {
        assert(section_size >= section_offset);
        return section_size - section_offset;
    }
    
    // Reads section contents into a string.
    public string section_to_string() throws GLib.Error {
        GLib.StringBuilder sb = new GLib.StringBuilder();
        while (section_offset < section_size) {
            sb.append_c((char) read_byte());
            if (sb.len > MAX_STRING_TO_SECTION_LENGTH) {
                return sb.str;
            }
        }
        return sb.str;
    }
    
}