(https://jhove.openpreservation.org/modules/tiff/)

> python -c 'import tifftools;tifftools.tiff_concat(["good1.svs", "good2.svs"], "out.svs", overwrite=True)'
> jhove -m TIFF-hul out.svs

Jhove (Rel. 1.24.1, 2020-03-16)
 Date: 2021-07-15 20:14:30 MDT
 RepresentationInformation: out.svs
  ReportingModule: TIFF-hul, Rel. 1.9.2 (2019-12-10)
  LastModified: 2021-07-15 20:14:20 MDT
  Size: 875685820
  Format: TIFF
  Status: Not well-formed
  SignatureMatches:
   TIFF-hul
  ErrorMessage: Value offset not word-aligned: 8842289
   ID: TIFF-HUL-4
   Offset: 8858224
  MIMEtype: image/tiff

https://web.archive.org/web/20160324105748/https://partners.adobe.com/public/developer/en/tiff/TIFF6.pdf on page 15 (about IFD entries) says:

Bytes 8-11 The Value Offset, the file offset (in bytes) of the Value for the field. The Value is expected to begin on a word boundary; the correspond- ing Value Offset will thus be an even number. This file offset may point anywhere in the file, even after the image data.

Hi,

I am trying to add a new tag to a TIF file using Python3.

I have worked through your code and I am now able to understand the TIFF structure.

However, I need to add a a new Tag. My code is an attempt to adapt your example for a similar action from the command line:

info = tifftools.read_tiff(Fpath) info[info, 'ifds'][0]['tags'][tifftools.Tag.ImageDescription.value] = { 'data': 'A dog digging.', 'datatype': tifftools.Datatype.ASCII }

tifftools.write_tiff(Fpath, OutPath, info)

I am not the strongest Python programmer, but I am baffled why I cannot update the TIFF TAG dict structure. Can you guide me please?

regards

Phil

Running the sample code from the readme file gives:

$ python3 sample.py
Traceback (most recent call last):
  File "sample2.py", line 8, in <module>
    exififd['tags'][tifftools.constants.EXIFTag.FNumber.value] = {
TypeError: list indices must be integers or slices, not str

it seems that we are missing level for exifs (SubIDF?).

The following does works:

$ diff -u sample.py.orig sample.py
--- sample.py.orig      2022-02-28 18:19:35.000000000 +0100
+++ sample.py   2022-02-28 18:19:52.000000000 +0100
@@ -4,7 +4,7 @@
     'data': 'A dog digging.',
     'datatype': tifftools.Datatype.ASCII
 }
-exififd = info['ifds'][0]['tags'][tifftools.Tag.EXIFIFD.value]['ifds'][0]
+exififd = info['ifds'][0]['tags'][tifftools.Tag.EXIFIFD.value]['ifds'][0][0]
 exififd['tags'][tifftools.constants.EXIFTag.FNumber.value] = {
     'data': [54, 10],
     'datatype': tifftools.Datatype.RATIONAL
$ python3 sample.py
$ tifftools dump photograph_tagged.tif | grep FNumber
      FNumber 33437 (0x829D) RATIONAL: 54 10 (5.4)

Side note: the above code will fail if the TIFF file doesn't already contains EXIFs. Is the following snippet the correct way to add EXIF IFD?

try:
    exif = info["ifds"][0]["tags"][tifftools.Tag.EXIFIFD.value]
except KeyError:
    exif = {
        "datatype": tifftools.Datatype.IFD,
        "ifds": [[{"tags": {}, "path_or_fobj": info["ifds"][0]["path_or_fobj"]}]],
    }
    info["ifds"][0]["tags"][tifftools.Tag.EXIFIFD.value] = exif
exififd = exif["ifds"][0][0]
# add tags herunder

Please could you provide more examples for how to use this tool to write tiffs from scratch? Thanks SO much for providing this tool. I really hope I can make it work for me!

Problem:

• I would like to write out numpy arrays to bigtiff with custom metadata AND exif data
• I'm getting exif data from a jpeg, then generating outputs from that jpeg within a script, that I want to write those outputs to a tiff with exif and metadata

Alternatives tried:

• I cant use tifffile because it doesnt deal with exif data
• I cant use PIL because it doesnt write exif data
• Rasterio doesnt seem to have an exif option
• I cant follow the provided example because I dont have a tiff to read as a starting point

So I'm really hoping I can use tifftools, but I cant figure out (from the one provided example) how to write a tiff entirely from scratch using arrays

Question: The provided example is useful, but only deals with the case where all the info is obtained from an existing file

info = tifftools.read_tiff(file)

But how would you create that info object from an numpy array, metadata, and exif separately? where metadata and exif are two separate dicts or JSON objects (or whatever)?

I also found this that helpfully explains how to add info to an existing tiff

In summary, I'm looking for advice for how to construct a viable info object that can be passed to .write_tif like this

tifftools.write_tif(info, ...)

where info is constructed from 1) a numpy array containing the image, 2) an array or dictionary of exif data, and 3) an array or dictionary of other metadata (tags)

The path_or_fobj internal value is only required for ifds if they need to transfer data from an existing tiff file. If tags are entirely self-contained, this is no longer required to be set.

NDPI files aren't quite valid tiff files. Rather, they are marked as non-bigtiff, but use 64-bit values for ifd offsets and have some implied upper bits for some data offsets. If a file is larger than 4 Gb or of unknown length, and a NDPI-specific tag is encountered that could have otherwise invalid offsets, read additional IFD offsets as 64-bit values and adjust data offset values according to the NDPI methods as illustrated by the OpenSlide library.

Better handle saving to small tiff. Before, once written to a bigtiff, it was unlikely to convert back to a small tiff since some fields written as LONG8 didn't automatically convert to LONG.

Better handle saving to small tiff. Before, once written to a bigtiff, it was unlikely to convert back to a small tiff since some fields written as LONG8 didn't automatically convert to LONG.