1
0
mirror of https://github.com/moparisthebest/SickRage synced 2024-11-11 11:55:03 -05:00
SickRage/lib/hachoir_parser/image/jpeg.py
echel0n 0d9fbc1ad7 Welcome to our SickBeard-TVRage Edition ...
This version of SickBeard uses both TVDB and TVRage to search and gather it's series data from allowing you to now have access to and download shows that you couldn't before because of being locked into only what TheTVDB had to offer.

Also this edition is based off the code we used in our XEM editon so it does come with scene numbering support as well as all the other features our XEM edition has to offer.

Please before using this with your existing database (sickbeard.db) please make a backup copy of it and delete any other database files such as cache.db and failed.db if present, we HIGHLY recommend starting out with no database files at all to make this a fresh start but the choice is at your own risk!

Enjoy!
2014-03-09 22:39:12 -07:00

369 lines
14 KiB
Python

"""
JPEG picture parser.
Information:
- APP14 documents
http://partners.adobe.com/public/developer/en/ps/sdk/5116.DCT_Filter.pdf
http://java.sun.com/j2se/1.5.0/docs/api/javax/imageio/metadata/doc-files/jpeg_metadata.html#color
- APP12:
http://search.cpan.org/~exiftool/Image-ExifTool/lib/Image/ExifTool/TagNames.pod
Author: Victor Stinner
"""
from lib.hachoir_core.error import HachoirError
from lib.hachoir_parser import Parser
from lib.hachoir_core.field import (FieldSet, ParserError,
UInt8, UInt16, Enum,
Bit, Bits, NullBits, NullBytes,
String, RawBytes)
from lib.hachoir_parser.image.common import PaletteRGB
from lib.hachoir_core.endian import BIG_ENDIAN
from lib.hachoir_core.text_handler import textHandler, hexadecimal
from lib.hachoir_parser.image.exif import Exif
from lib.hachoir_parser.image.photoshop_metadata import PhotoshopMetadata
MAX_FILESIZE = 100 * 1024 * 1024
# The four tables (hash/sum for color/grayscale JPEG) comes
# from ImageMagick project
QUALITY_HASH_COLOR = (
1020, 1015, 932, 848, 780, 735, 702, 679, 660, 645,
632, 623, 613, 607, 600, 594, 589, 585, 581, 571,
555, 542, 529, 514, 494, 474, 457, 439, 424, 410,
397, 386, 373, 364, 351, 341, 334, 324, 317, 309,
299, 294, 287, 279, 274, 267, 262, 257, 251, 247,
243, 237, 232, 227, 222, 217, 213, 207, 202, 198,
192, 188, 183, 177, 173, 168, 163, 157, 153, 148,
143, 139, 132, 128, 125, 119, 115, 108, 104, 99,
94, 90, 84, 79, 74, 70, 64, 59, 55, 49,
45, 40, 34, 30, 25, 20, 15, 11, 6, 4,
0)
QUALITY_SUM_COLOR = (
32640,32635,32266,31495,30665,29804,29146,28599,28104,27670,
27225,26725,26210,25716,25240,24789,24373,23946,23572,22846,
21801,20842,19949,19121,18386,17651,16998,16349,15800,15247,
14783,14321,13859,13535,13081,12702,12423,12056,11779,11513,
11135,10955,10676,10392,10208, 9928, 9747, 9564, 9369, 9193,
9017, 8822, 8639, 8458, 8270, 8084, 7896, 7710, 7527, 7347,
7156, 6977, 6788, 6607, 6422, 6236, 6054, 5867, 5684, 5495,
5305, 5128, 4945, 4751, 4638, 4442, 4248, 4065, 3888, 3698,
3509, 3326, 3139, 2957, 2775, 2586, 2405, 2216, 2037, 1846,
1666, 1483, 1297, 1109, 927, 735, 554, 375, 201, 128,
0)
QUALITY_HASH_GRAY = (
510, 505, 422, 380, 355, 338, 326, 318, 311, 305,
300, 297, 293, 291, 288, 286, 284, 283, 281, 280,
279, 278, 277, 273, 262, 251, 243, 233, 225, 218,
211, 205, 198, 193, 186, 181, 177, 172, 168, 164,
158, 156, 152, 148, 145, 142, 139, 136, 133, 131,
129, 126, 123, 120, 118, 115, 113, 110, 107, 105,
102, 100, 97, 94, 92, 89, 87, 83, 81, 79,
76, 74, 70, 68, 66, 63, 61, 57, 55, 52,
50, 48, 44, 42, 39, 37, 34, 31, 29, 26,
24, 21, 18, 16, 13, 11, 8, 6, 3, 2,
0)
QUALITY_SUM_GRAY = (
16320,16315,15946,15277,14655,14073,13623,13230,12859,12560,
12240,11861,11456,11081,10714,10360,10027, 9679, 9368, 9056,
8680, 8331, 7995, 7668, 7376, 7084, 6823, 6562, 6345, 6125,
5939, 5756, 5571, 5421, 5240, 5086, 4976, 4829, 4719, 4616,
4463, 4393, 4280, 4166, 4092, 3980, 3909, 3835, 3755, 3688,
3621, 3541, 3467, 3396, 3323, 3247, 3170, 3096, 3021, 2952,
2874, 2804, 2727, 2657, 2583, 2509, 2437, 2362, 2290, 2211,
2136, 2068, 1996, 1915, 1858, 1773, 1692, 1620, 1552, 1477,
1398, 1326, 1251, 1179, 1109, 1031, 961, 884, 814, 736,
667, 592, 518, 441, 369, 292, 221, 151, 86, 64,
0)
JPEG_NATURAL_ORDER = (
0, 1, 8, 16, 9, 2, 3, 10,
17, 24, 32, 25, 18, 11, 4, 5,
12, 19, 26, 33, 40, 48, 41, 34,
27, 20, 13, 6, 7, 14, 21, 28,
35, 42, 49, 56, 57, 50, 43, 36,
29, 22, 15, 23, 30, 37, 44, 51,
58, 59, 52, 45, 38, 31, 39, 46,
53, 60, 61, 54, 47, 55, 62, 63)
class JpegChunkApp0(FieldSet):
UNIT_NAME = {
0: "pixels",
1: "dots per inch",
2: "dots per cm",
}
def createFields(self):
yield String(self, "jfif", 5, "JFIF string", charset="ASCII")
if self["jfif"].value != "JFIF\0":
raise ParserError(
"Stream doesn't look like JPEG chunk (wrong JFIF signature)")
yield UInt8(self, "ver_maj", "Major version")
yield UInt8(self, "ver_min", "Minor version")
yield Enum(UInt8(self, "units", "Units"), self.UNIT_NAME)
if self["units"].value == 0:
yield UInt16(self, "aspect_x", "Aspect ratio (X)")
yield UInt16(self, "aspect_y", "Aspect ratio (Y)")
else:
yield UInt16(self, "x_density", "X density")
yield UInt16(self, "y_density", "Y density")
yield UInt8(self, "thumb_w", "Thumbnail width")
yield UInt8(self, "thumb_h", "Thumbnail height")
thumb_size = self["thumb_w"].value * self["thumb_h"].value
if thumb_size != 0:
yield PaletteRGB(self, "thumb_palette", 256)
yield RawBytes(self, "thumb_data", thumb_size, "Thumbnail data")
class Ducky(FieldSet):
BLOCK_TYPE = {
0: "end",
1: "Quality",
2: "Comment",
3: "Copyright",
}
def createFields(self):
yield Enum(UInt16(self, "type"), self.BLOCK_TYPE)
if self["type"].value == 0:
return
yield UInt16(self, "size")
size = self["size"].value
if size:
yield RawBytes(self, "data", size)
class APP12(FieldSet):
"""
The JPEG APP12 "Picture Info" segment was used by some older cameras, and
contains ASCII-based meta information.
"""
def createFields(self):
yield String(self, "ducky", 5, '"Ducky" string', charset="ASCII")
while not self.eof:
yield Ducky(self, "item[]")
class StartOfFrame(FieldSet):
def createFields(self):
yield UInt8(self, "precision")
yield UInt16(self, "height")
yield UInt16(self, "width")
yield UInt8(self, "nr_components")
for index in range(self["nr_components"].value):
yield UInt8(self, "component_id[]")
yield UInt8(self, "high[]")
yield UInt8(self, "low[]")
class Comment(FieldSet):
def createFields(self):
yield String(self, "comment", self.size//8, strip="\0")
class AdobeChunk(FieldSet):
COLORSPACE_TRANSFORMATION = {
1: "YCbCr (converted from RGB)",
2: "YCCK (converted from CMYK)",
}
def createFields(self):
if self.stream.readBytes(self.absolute_address, 5) != "Adobe":
yield RawBytes(self, "raw", self.size//8, "Raw data")
return
yield String(self, "adobe", 5, "\"Adobe\" string", charset="ASCII")
yield UInt16(self, "version", "DCT encoder version")
yield Enum(Bit(self, "flag00"),
{False: "Chop down or subsampling", True: "Blend"})
yield NullBits(self, "flags0_reserved", 15)
yield NullBytes(self, "flags1", 2)
yield Enum(UInt8(self, "color_transform", "Colorspace transformation code"), self.COLORSPACE_TRANSFORMATION)
class StartOfScan(FieldSet):
def createFields(self):
yield UInt8(self, "nr_components")
for index in range(self["nr_components"].value):
comp_id = UInt8(self, "component_id[]")
yield comp_id
if not(1 <= comp_id.value <= self["nr_components"].value):
raise ParserError("JPEG error: Invalid component-id")
yield UInt8(self, "value[]")
yield RawBytes(self, "raw", 3) # TODO: What's this???
class RestartInterval(FieldSet):
def createFields(self):
yield UInt16(self, "interval", "Restart interval")
class QuantizationTable(FieldSet):
def createFields(self):
# Code based on function get_dqt() (jdmarker.c from libjpeg62)
yield Bits(self, "is_16bit", 4)
yield Bits(self, "index", 4)
if self["index"].value >= 4:
raise ParserError("Invalid quantification index (%s)" % self["index"].value)
if self["is_16bit"].value:
coeff_type = UInt16
else:
coeff_type = UInt8
for index in xrange(64):
natural = JPEG_NATURAL_ORDER[index]
yield coeff_type(self, "coeff[%u]" % natural)
def createDescription(self):
return "Quantification table #%u" % self["index"].value
class DefineQuantizationTable(FieldSet):
def createFields(self):
while self.current_size < self.size:
yield QuantizationTable(self, "qt[]")
class JpegChunk(FieldSet):
TAG_SOI = 0xD8
TAG_EOI = 0xD9
TAG_SOS = 0xDA
TAG_DQT = 0xDB
TAG_DRI = 0xDD
TAG_INFO = {
0xC4: ("huffman[]", "Define Huffman Table (DHT)", None),
0xD8: ("start_image", "Start of image (SOI)", None),
0xD9: ("end_image", "End of image (EOI)", None),
0xDA: ("start_scan", "Start Of Scan (SOS)", StartOfScan),
0xDB: ("quantization[]", "Define Quantization Table (DQT)", DefineQuantizationTable),
0xDC: ("nb_line", "Define number of Lines (DNL)", None),
0xDD: ("restart_interval", "Define Restart Interval (DRI)", RestartInterval),
0xE0: ("app0", "APP0", JpegChunkApp0),
0xE1: ("exif", "Exif metadata", Exif),
0xE2: ("icc", "ICC profile", None),
0xEC: ("app12", "APP12", APP12),
0xED: ("photoshop", "Photoshop", PhotoshopMetadata),
0xEE: ("adobe", "Image encoding information for DCT filters (Adobe)", AdobeChunk),
0xFE: ("comment[]", "Comment", Comment),
}
START_OF_FRAME = {
0xC0: u"Baseline",
0xC1: u"Extended sequential",
0xC2: u"Progressive",
0xC3: u"Lossless",
0xC5: u"Differential sequential",
0xC6: u"Differential progressive",
0xC7: u"Differential lossless",
0xC9: u"Extended sequential, arithmetic coding",
0xCA: u"Progressive, arithmetic coding",
0xCB: u"Lossless, arithmetic coding",
0xCD: u"Differential sequential, arithmetic coding",
0xCE: u"Differential progressive, arithmetic coding",
0xCF: u"Differential lossless, arithmetic coding",
}
for key, text in START_OF_FRAME.iteritems():
TAG_INFO[key] = ("start_frame", "Start of frame (%s)" % text.lower(), StartOfFrame)
def __init__(self, parent, name, description=None):
FieldSet.__init__(self, parent, name, description)
tag = self["type"].value
if tag == 0xE1:
# Hack for Adobe extension: XAP metadata (as XML)
bytes = self.stream.readBytes(self.absolute_address + 32, 6)
if bytes == "Exif\0\0":
self._name = "exif"
self._description = "EXIF"
self._parser = Exif
else:
self._parser = None
elif tag in self.TAG_INFO:
self._name, self._description, self._parser = self.TAG_INFO[tag]
else:
self._parser = None
def createFields(self):
yield textHandler(UInt8(self, "header", "Header"), hexadecimal)
if self["header"].value != 0xFF:
raise ParserError("JPEG: Invalid chunk header!")
yield textHandler(UInt8(self, "type", "Type"), hexadecimal)
tag = self["type"].value
if tag in (self.TAG_SOI, self.TAG_EOI):
return
yield UInt16(self, "size", "Size")
size = (self["size"].value - 2)
if 0 < size:
if self._parser:
yield self._parser(self, "content", "Chunk content", size=size*8)
else:
yield RawBytes(self, "data", size, "Data")
def createDescription(self):
return "Chunk: %s" % self["type"].display
class JpegFile(Parser):
endian = BIG_ENDIAN
PARSER_TAGS = {
"id": "jpeg",
"category": "image",
"file_ext": ("jpg", "jpeg"),
"mime": (u"image/jpeg",),
"magic": (
("\xFF\xD8\xFF\xE0", 0), # (Start Of Image, APP0)
("\xFF\xD8\xFF\xE1", 0), # (Start Of Image, EXIF)
("\xFF\xD8\xFF\xEE", 0), # (Start Of Image, Adobe)
),
"min_size": 22*8,
"description": "JPEG picture",
"subfile": "skip",
}
def validate(self):
if self.stream.readBytes(0, 2) != "\xFF\xD8":
return "Invalid file signature"
try:
for index, field in enumerate(self):
chunk_type = field["type"].value
if chunk_type not in JpegChunk.TAG_INFO:
return "Unknown chunk type: 0x%02X (chunk #%s)" % (chunk_type, index)
if index == 2:
# Only check 3 fields
break
except HachoirError:
return "Unable to parse at least three chunks"
return True
def createFields(self):
while not self.eof:
chunk = JpegChunk(self, "chunk[]")
yield chunk
if chunk["type"].value == JpegChunk.TAG_SOS:
# TODO: Read JPEG image data...
break
# TODO: is it possible to handle piped input?
if self._size is None:
raise NotImplementedError
has_end = False
size = (self._size - self.current_size) // 8
if size:
if 2 < size \
and self.stream.readBytes(self._size - 16, 2) == "\xff\xd9":
has_end = True
size -= 2
yield RawBytes(self, "data", size, "JPEG data")
if has_end:
yield JpegChunk(self, "chunk[]")
def createDescription(self):
desc = "JPEG picture"
if "sof/content" in self:
header = self["sof/content"]
desc += ": %ux%u pixels" % (header["width"].value, header["height"].value)
return desc
def createContentSize(self):
if "end" in self:
return self["end"].absolute_address + self["end"].size
if "data" not in self:
return None
start = self["data"].absolute_address
end = self.stream.searchBytes("\xff\xd9", start, MAX_FILESIZE*8)
if end is not None:
return end + 16
return None