Merge pull request #8642 from radarhere/bigtiff

Author: hugovk

Tests/test_file_tiff.py

@@ -115,6 +115,13 @@ def test_bigtiff(self, tmp_path: Path) -> None:
             outfile = str(tmp_path / "temp.tif")
             im.save(outfile, save_all=True, append_images=[im], tiffinfo=im.tag_v2)
 
+    def test_bigtiff_save(self, tmp_path: Path) -> None:
+        outfile = str(tmp_path / "temp.tif")
+        hopper().save(outfile, big_tiff=True)
+
+        with Image.open(outfile) as im:
+            assert im.tag_v2._bigtiff is True
+
     def test_seek_too_large(self) -> None:
         with pytest.raises(ValueError, match="Unable to seek to frame"):
             Image.open("Tests/images/seek_too_large.tif")

docs/handbook/image-file-formats.rst

@@ -1208,6 +1208,11 @@ The :py:meth:`~PIL.Image.Image.save` method can take the following keyword argum
 
     .. versionadded:: 8.4.0
 
+**big_tiff**
+    If true, the image will be saved as a BigTIFF.
+
+    .. versionadded:: 11.1.0
+
 **compression**
     A string containing the desired compression method for the
     file. (valid only with libtiff installed) Valid compression

docs/releasenotes/11.1.0.rst

@@ -1,25 +1,6 @@
 11.1.0
 ------
 
-Security
-========
-
-TODO
-^^^^
-
-TODO
-
-:cve:`YYYY-XXXXX`: TODO
-^^^^^^^^^^^^^^^^^^^^^^^
-
-TODO
-
-Backwards Incompatible Changes
-==============================
-
-TODO
-^^^^
-
 Deprecations
 ============
 
@@ -66,6 +47,13 @@ zlib library, and what version of zlib-ng is being used::
     features.check_feature("zlib_ng")  # True or False
     features.version_feature("zlib_ng")  # "2.2.2" for example, or None
 
+Saving TIFF as BigTIFF
+^^^^^^^^^^^^^^^^^^^^^^
+
+TIFF images can now be saved as BigTIFF using a ``big_tiff`` argument::
+
+    im.save("out.tiff", big_tiff=True)
+
 Other Changes
 =============
 

src/PIL/TiffImagePlugin.py

@@ -582,7 +582,7 @@ class ImageFileDirectory_v2(_IFDv2Base):
 
     def __init__(
         self,
-        ifh: bytes = b"II\052\0\0\0\0\0",
+        ifh: bytes = b"II\x2A\x00\x00\x00\x00\x00",
         prefix: bytes | None = None,
         group: int | None = None,
     ) -> None:
@@ -949,28 +949,34 @@ def load(self, fp: IO[bytes]) -> None:
             warnings.warn(str(msg))
             return
 
+    def _get_ifh(self):
+        ifh = self._prefix + self._pack("H", 43 if self._bigtiff else 42)
+        if self._bigtiff:
+            ifh += self._pack("HH", 8, 0)
+        ifh += self._pack("Q", 16) if self._bigtiff else self._pack("L", 8)
+
+        return ifh
+
     def tobytes(self, offset: int = 0) -> bytes:
         # FIXME What about tagdata?
-        result = self._pack("H", len(self._tags_v2))
+        result = self._pack("Q" if self._bigtiff else "H", len(self._tags_v2))
 
         entries: list[tuple[int, int, int, bytes, bytes]] = []
-        offset = offset + len(result) + len(self._tags_v2) * 12 + 4
+        offset += len(result) + len(self._tags_v2) * (20 if self._bigtiff else 12) + 4
         stripoffsets = None
 
         # pass 1: convert tags to binary format
         # always write tags in ascending order
+        fmt = "Q" if self._bigtiff else "L"
+        fmt_size = 8 if self._bigtiff else 4
         for tag, value in sorted(self._tags_v2.items()):
             if tag == STRIPOFFSETS:
                 stripoffsets = len(entries)
             typ = self.tagtype[tag]
             logger.debug("Tag %s, Type: %s, Value: %s", tag, typ, repr(value))
             is_ifd = typ == TiffTags.LONG and isinstance(value, dict)
             if is_ifd:
-                if self._endian == "<":
-                    ifh = b"II\x2A\x00\x08\x00\x00\x00"
-                else:
-                    ifh = b"MM\x00\x2A\x00\x00\x00\x08"
-                ifd = ImageFileDirectory_v2(ifh, group=tag)
+                ifd = ImageFileDirectory_v2(self._get_ifh(), group=tag)
                 values = self._tags_v2[tag]
                 for ifd_tag, ifd_value in values.items():
                     ifd[ifd_tag] = ifd_value
@@ -993,10 +999,10 @@ def tobytes(self, offset: int = 0) -> bytes:
             else:
                 count = len(values)
             # figure out if data fits into the entry
-            if len(data) <= 4:
-                entries.append((tag, typ, count, data.ljust(4, b"\0"), b""))
+            if len(data) <= fmt_size:
+                entries.append((tag, typ, count, data.ljust(fmt_size, b"\0"), b""))
             else:
-                entries.append((tag, typ, count, self._pack("L", offset), data))
+                entries.append((tag, typ, count, self._pack(fmt, offset), data))
                 offset += (len(data) + 1) // 2 * 2  # pad to word
 
         # update strip offset data to point beyond auxiliary data
@@ -1007,13 +1013,15 @@ def tobytes(self, offset: int = 0) -> bytes:
                 values = [val + offset for val in handler(self, data, self.legacy_api)]
                 data = self._write_dispatch[typ](self, *values)
             else:
-                value = self._pack("L", self._unpack("L", value)[0] + offset)
+                value = self._pack(fmt, self._unpack(fmt, value)[0] + offset)
             entries[stripoffsets] = tag, typ, count, value, data
 
         # pass 2: write entries to file
         for tag, typ, count, value, data in entries:
             logger.debug("%s %s %s %s %s", tag, typ, count, repr(value), repr(data))
-            result += self._pack("HHL4s", tag, typ, count, value)
+            result += self._pack(
+                "HHQ8s" if self._bigtiff else "HHL4s", tag, typ, count, value
+            )
 
         # -- overwrite here for multi-page --
         result += b"\0\0\0\0"  # end of entries
@@ -1028,8 +1036,7 @@ def tobytes(self, offset: int = 0) -> bytes:
 
     def save(self, fp: IO[bytes]) -> int:
         if fp.tell() == 0:  # skip TIFF header on subsequent pages
-            # tiff header -- PIL always starts the first IFD at offset 8
-            fp.write(self._prefix + self._pack("HL", 42, 8))
+            fp.write(self._get_ifh())
 
         offset = fp.tell()
         result = self.tobytes(offset)
@@ -1680,10 +1687,13 @@ def _save(im: Image.Image, fp: IO[bytes], filename: str | bytes) -> None:
         msg = f"cannot write mode {im.mode} as TIFF"
         raise OSError(msg) from e
 
-    ifd = ImageFileDirectory_v2(prefix=prefix)
-
     encoderinfo = im.encoderinfo
     encoderconfig = im.encoderconfig
+
+    ifd = ImageFileDirectory_v2(prefix=prefix)
+    if encoderinfo.get("big_tiff"):
+        ifd._bigtiff = True
+
     try:
         compression = encoderinfo["compression"]
     except KeyError: