File size: 15,148 Bytes
3a3c77b | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 | # Copyright (c) 2024, pypdf contributors
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright notice,
# this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright notice,
# this list of conditions and the following disclaimer in the documentation
# and/or other materials provided with the distribution.
# * The name of the author may not be used to endorse or promote products
# derived from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
# ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
# LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
# CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
# SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
# POSSIBILITY OF SUCH DAMAGE.
import logging
from io import BytesIO
from typing import IO
from .._utils import (
WHITESPACES,
WHITESPACES_AS_BYTES,
BinaryStreamType,
StreamType,
logger_warning,
read_non_whitespace,
)
from ..errors import PdfReadError
logger = logging.getLogger(__name__)
# An inline image should be used only for small images (4096 bytes or less),
# but allow twice this for cases where this has been exceeded.
BUFFER_SIZE = 8192
def _check_end_image_marker(stream: StreamType) -> bool:
ei_tok = read_non_whitespace(stream)
ei_tok += stream.read(2)
stream.seek(-3, 1)
return ei_tok[:2] == b"EI" and (ei_tok[2:3] == b"" or ei_tok[2:3] in WHITESPACES)
def extract_inline__ascii_hex_decode(stream: StreamType) -> bytes:
"""
Extract HexEncoded stream from inline image.
The stream will be moved onto the EI.
"""
data_out = bytearray()
# Read data until delimiter > and EI as backup.
while True:
data_buffered = read_non_whitespace(stream) + (read_bytes := stream.read(BUFFER_SIZE))
if not data_buffered or not read_bytes:
raise PdfReadError("Unexpected end of stream.")
pos_tok = data_buffered.find(b">")
if pos_tok >= 0: # found >
data_out += data_buffered[: pos_tok + 1]
stream.seek(-len(data_buffered) + pos_tok + 1, 1)
break
pos_ei = data_buffered.find(b"EI")
if pos_ei >= 0: # found EI
stream.seek(-len(data_buffered) + pos_ei - 1, 1)
c = stream.read(1)
while c in WHITESPACES:
stream.seek(-2, 1)
c = stream.read(1)
pos_ei -= 1
data_out += data_buffered[:pos_ei]
break
if len(data_buffered) == 2:
data_out += data_buffered
raise PdfReadError("Unexpected end of stream.")
# Neither > nor EI found
data_out += data_buffered[:-2]
stream.seek(-2, 1)
if not _check_end_image_marker(stream):
raise PdfReadError("EI stream not found.")
return bytes(data_out)
def extract_inline__ascii85_decode(stream: StreamType) -> bytes:
"""
Extract A85 stream from inline image.
The stream will be moved onto the EI.
"""
data_out = bytearray()
# Read data until delimiter ~>
while True:
data_buffered = read_non_whitespace(stream) + (read_bytes := stream.read(BUFFER_SIZE))
if not data_buffered or not read_bytes:
raise PdfReadError("Unexpected end of stream.")
pos_tok = data_buffered.find(b"~>")
if pos_tok >= 0: # found!
data_out += data_buffered[: pos_tok + 2]
stream.seek(-len(data_buffered) + pos_tok + 2, 1)
break
if len(data_buffered) == 2: # end of buffer
data_out += data_buffered
raise PdfReadError("Unexpected end of stream.")
data_out += data_buffered[
:-2
] # back by one char in case of in the middle of ~>
stream.seek(-2, 1)
if not _check_end_image_marker(stream):
raise PdfReadError("EI stream not found.")
return bytes(data_out)
def extract_inline__run_length_decode(stream: StreamType) -> bytes:
"""
Extract RL (RunLengthDecode) stream from inline image.
The stream will be moved onto the EI.
"""
data_out = bytearray()
# Read data until delimiter 128
while True:
data_buffered = stream.read(BUFFER_SIZE)
if not data_buffered:
raise PdfReadError("Unexpected end of stream.")
pos_tok = data_buffered.find(b"\x80")
if pos_tok >= 0: # found
# Ideally, we could just use plain run-length decoding here, where 80_16 = 128_10
# marks the EOD. But there apparently are cases like in issue #3517, where we have
# an inline image with up to 51 EOD markers. In these cases, be resilient here and
# use the default `EI` marker detection instead. Please note that this fallback
# still omits special `EI` handling within the stream, but for now assume that having
# both of these cases occur at the same time is very unlikely (and the image stream
# is broken anyway).
# For now, do not skip over more than one whitespace character.
after_token = data_buffered[pos_tok + 1 : pos_tok + 4]
if after_token.startswith(b"EI") or after_token.endswith(b"EI"):
data_out += data_buffered[: pos_tok + 1]
stream.seek(-len(data_buffered) + pos_tok + 1, 1)
else:
logger_warning("Early EOD in RunLengthDecode of inline image, using fallback.", source=__name__)
ei_marker = data_buffered.find(b"EI")
if ei_marker > 0:
data_out += data_buffered[: ei_marker]
stream.seek(-len(data_buffered) + ei_marker - 1, 1)
break
data_out += data_buffered
if not _check_end_image_marker(stream):
raise PdfReadError("EI stream not found.")
return bytes(data_out)
def extract_inline__dct_decode(stream: BinaryStreamType) -> bytes:
"""
Extract DCT (JPEG) stream from inline image.
The stream will be moved onto the EI.
"""
def read(length: int) -> bytes:
# If 0 bytes are returned, and *size* was not 0, this indicates end of file.
# If the object is in non-blocking mode and no bytes are available, `None` is returned.
_result = stream.read(length)
if _result is None or len(_result) != length:
raise PdfReadError("Unexpected end of stream.")
return _result
data_out = bytearray()
# Read Blocks of data (ID/Size/data) up to ID=FF/D9
# https://www.digicamsoft.com/itu/itu-t81-36.html
not_first = False
while True:
c = read(1)
if not_first or (c == b"\xff"):
data_out += c
if c != b"\xff":
continue
not_first = True
c = read(1)
data_out += c
if c == b"\xff":
stream.seek(-1, 1) # pragma: no cover
elif c == b"\x00": # stuffing
pass
elif c == b"\xd9": # end
break
elif c in (
b"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc9\xca\xcb\xcc\xcd\xce\xcf"
b"\xda\xdb\xdc\xdd\xde\xdf"
b"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef\xfe"
):
c = read(2)
data_out += c
sz = c[0] * 256 + c[1]
data_out += read(sz - 2)
if not _check_end_image_marker(stream):
raise PdfReadError("EI stream not found.")
return bytes(data_out)
def _is_preceded_by_whitespace(stream_out: BytesIO, position: int) -> bool:
"""
Check whether the byte in front of `position` is a whitespace.
Returns:
Whether the preceding byte is whitespace. If no preceding byte exists,
returns ``False`` because the ``EI`` marker is at the beginning of the
image data and is therefore not delimited by anything.
"""
if position <= 0:
return False
saved_position = stream_out.tell()
# Avoid exporting the BytesIO buffer: PyPy may keep the view alive,
# preventing later writes from resizing the output stream.
stream_out.seek(position - 1)
is_whitespace = stream_out.read(1) in WHITESPACES
stream_out.seek(saved_position)
return is_whitespace
def _is_followed_by_closing_operator(stream: StreamType, first_byte: bytes) -> bool:
"""
Check whether `first_byte` starts an exact `Q` or `EMC` operator.
`first_byte` has already been consumed from the stream, which is restored to its
current position afterwards. Looking at this byte alone is not sufficient, as
binary image data can easily be mistaken for one of these operators.
"""
position = stream.tell()
if first_byte == b"Q":
suffix = stream.read(1)
is_closing_operator = suffix == b"" or suffix in WHITESPACES
elif first_byte == b"E":
suffix = stream.read(3)
is_closing_operator = suffix[:2] == b"MC" and (suffix[2:3] == b"" or suffix[2:3] in WHITESPACES)
else:
is_closing_operator = False
stream.seek(position, 0)
return is_closing_operator
def extract_inline_default(stream: StreamType) -> bytes:
"""Legacy method, used by default"""
stream_out = BytesIO()
# Read the inline image, while checking for EI (End Image) operator.
while True:
data_buffered = stream.read(BUFFER_SIZE)
if not data_buffered:
raise PdfReadError("Unexpected end of stream.")
pos_ei = data_buffered.find(
b"E"
) # We can not look straight for "EI" because it may not have been loaded in the buffer
if pos_ei == -1:
stream_out.write(data_buffered)
else:
# Write out everything including E (the one from EI to be removed)
stream_out.write(data_buffered[0 : pos_ei + 1])
sav_pos_ei = stream_out.tell() - 1
# Seek back in the stream to read the E next
stream.seek(pos_ei + 1 - len(data_buffered), 1)
saved_pos = stream.tell()
# Check for End Image
tok2 = stream.read(1) # I of "EI"
if tok2 != b"I":
stream.seek(saved_pos, 0)
continue
tok3 = stream.read(1) # possible space after "EI"
if tok3 == b"":
# The `EI` marker is at the very end of the stream. There can be
# no trailing binary data, so this is unambiguously the end of the
# inline image (#3468).
stream.seek(saved_pos - 1, 0)
stream_out.truncate(sav_pos_ei)
break
if tok3 not in WHITESPACES:
stream.seek(saved_pos, 0)
continue
while tok3 in WHITESPACES:
tok3 = stream.read(1)
# The `EI` marker should be delimited by whitespace on both sides. Preserve the
# legacy tolerance for a missing leading delimiter only when it is followed by an
# exact closing operator, as binary image data would match otherwise.
if not _is_preceded_by_whitespace(stream_out, sav_pos_ei) and not _is_followed_by_closing_operator(
stream, tok3
):
stream.seek(saved_pos, 0)
continue
if is_followed_by_binary_data(stream):
# Inline image contains `EI ` sequence usually marking the end of it, but
# is followed by binary data which does not make sense for the actual end.
stream.seek(saved_pos, 0)
continue
# Data contains [\s]EI[\s](Q|EMC): 4 chars are sufficient
# remove E(I) wrongly inserted earlier
stream.seek(saved_pos - 1, 0)
stream_out.truncate(sav_pos_ei)
break
return stream_out.getvalue()
def is_followed_by_binary_data(stream: IO[bytes], length: int = 10) -> bool:
"""
Check if the next bytes of the stream look like binary image data or regular page content.
This is just some heuristics due to the PDF specification being too imprecise about
inline images containing the `EI` marker which would end an image. Starting with PDF 2.0,
we finally get a mandatory length field, but with (proper) PDF 2.0 support being very limited
everywhere, we should not expect to be able to remove such hacks in the near future - especially
considering legacy documents as well.
The actual implementation draws some inspiration from
https://github.com/itext/itext-java/blob/9.1.0/kernel/src/main/java/com/itextpdf/kernel/pdf/canvas/parser/util/InlineImageParsingUtils.java
"""
position = stream.tell()
data = stream.read(length)
stream.seek(position)
if not data:
return False
operator_start = None
operator_end = None
for index, byte in enumerate(data):
if byte < 32 and byte not in WHITESPACES_AS_BYTES:
# This covers all characters not being displayable directly, although omitting whitespace
# to allow for operator detection.
return True
is_whitespace = byte in WHITESPACES_AS_BYTES
if operator_start is None and not is_whitespace:
# Interpret all other non-whitespace characters as the start of an operation.
operator_start = index
if operator_start is not None and is_whitespace:
# A whitespace stops an operation.
# Assume that having an inline image with tons of whitespace is rather unlikely.
operator_end = index
break
if operator_start is None:
# Inline images should not have tons of whitespaces, which would lead to no operator start.
return False
if operator_end is None:
# We probably are inside an operation.
operator_end = length
operator_length = operator_end - operator_start
operator = data[operator_start:operator_end]
if operator.startswith(b"/") and operator_length > 1:
# Name object.
return False
if operator.replace(b".", b"").isdigit():
# Graphics operator, for example a move. A number (integer or float).
return False
if operator_length > 3: # noqa: SIM103
# Usually, the operators inside a content stream should not have more than three characters,
# especially after an inline image.
return True
return False
|