(free/paid) Image Processing on Upload

[fortser]

i have an image-processing forum, so many heavy PNG images are being uploaded. I need a script that, at upload time, converts the images to JPEG at 80% quality, while also extracting technical data from the PNG file (similar to EXIF data, but in a slightly different, custom format), and publishes that data in the body of the post alongside the uploaded image.

I've already written a Python script on my local machine that converts the image to the format I need and saves the extracted text data to a .txt file next to the JPEG.

Please advise on the safest way, given the forum's architecture, to hook my script into the existing image-upload process.

[Kindred]

moved to mod requests

Please update the title to include whether this is a free or paid request

[fortser]

I don't see a button for editing messages (((
if it's really difficult to implement such functionality, then of course I'm ready for a small payment. We'll make an agreement

[GL700Wing]

@forster There is already a mod which I manage that converts/resizes images on upload (see  Automatic Attachment Rotation (and Resize)) but it doesn't extract any image data.

What image data do you want to extract and where/how do you want to use the data?

[dodos26]

You cannot view this attachment.
You cannot view this attachment.
You cannot view this attachment.

Want to get something like this inside the forum?

[fortser]

Here are examples of images (generated in the stable diffusion system) and the technical parameters of the generation. These parameters are located inside the technical data in the PNG file.

1
a detailed photo of  A close-up of a couple's faces inches apart, lips parted, hot breath mingling, eyes half-closed in pleasure, her hand gripping his jaw, intense passion, soft shadows, cinematic focus.

Steps: 25,  Euler, Schedule type: Simple, CFG scale: 1, Distilled CFG Scale: 3.5, Seed: 1735199045, Size: 896x1152, Model hash: 129032f322, Model: nsfwMASTERFLUXFp16LoraMerged_v12Q80Gguf, Version: f2.0.1v1.10.1-previous-659-gc055f2d4, Module 1: clip_l, Module 2: t5xxl_fp8_e4m3fn, Module 3: ae

2
Elderly bearded man with weathered face, deep wrinkles, kind eyes, wearing simple linen clothes, sitting by a wooden table in a rustic interior

(deformed face:1.3), (asymmetrical face:1.2), (disfigured:1.3), poorly drawn face, (bad anatomy:1.2), (extra limbs:1.3), missing fingers, (mutated hands:1.3), blurry, low resolution, pixelated, noisy, jpeg artifacts, (text:1.2), watermark, logo, signature, (cartoon style:1.3), (anime:1.3), 3d, illustration, sketch, painting, (low quality:1.3), (bad quality:1.3), distortion, wrong proportions, (multiple faces:1.3), cropped

Steps: 28,  Euler a, Schedule type: Automatic, CFG scale: 7, Seed: 111910324, Size: 1024x1024, Model hash: aeb7e9e689, Model: juggernautXL_v8Rundiffusion, Version: 1.10.1

[fortser]

GL700Wing, Yes, I was thinking about this extension. It would be great if it could be adapted to my needs.
dodos26, yes, it seems that your option does what it needs to do. can this script be used while uploading images to regular messages ?

[vbgamer45]

Interesting never knew PNG had data like EXIF inside them.

[dodos26]

It depends on what you want to achieve, do you want to create some database, do you have some gallery or only from attachments? Do you want to only read data after clicking a button? If you only want after clicking a button, you can do it easily, through javascript and fetch.

Before testing use check_config.php next try api_docs.php next can test this script via index.html

For ur photos return:

Code Select Expand

IHDR:
width: 896
height: 1152
depth: 8
color: 6
compression: 0
filter: 0
interlace: 0
iTXt:
parameters: [International Text]
other:
IDAT: 691
IEND: 0
And:

Code Select Expand

IHDR:
width: 1024
height: 1024
depth: 8
color: 2
compression: 0
filter: 0
interlace: 0
tEXt:
parameters: Elderly bearded man with weathered face, deep wrinkles, kind eyes, wearing simple linen clothes, sitting by a wooden table in a rustic interior Negative prompt: (deformed face:1.3), (asymmetrical face:1.2), (disfigured:1.3), poorly drawn face, (bad anatomy:1.2), (extra limbs:1.3), missing fingers, (mutated hands:1.3), blurry, low resolution, pixelated, noisy, jpeg artifacts, (text:1.2), watermark, logo, signature, (cartoon style:1.3), (anime:1.3), 3d, illustration, sketch, painting, (low quality:1.3), (bad quality:1.3), distortion, wrong proportions, (multiple faces:1.3), cropped Steps: 28, Sampler: Euler a, Schedule type: Automatic, CFG scale: 7, Seed: 111910324, Size: 1024x1024, Model hash: aeb7e9e689, Model: juggernautXL_v8Rundiffusion, Version: 1.10.1
other:
IDAT: 17853
IEND: 0