Resolution of PDF in example/doc-browser.py about pymupdf-utilities HOT 7 CLOSED

I found the solution and it has to do with scaling in Microsoft Windows and DPI awareness. The solution is to set DPI awareness by adding this in the code.

import ctypes

ctypes.windll.shcore.SetProcessDpiAwareness(2)

from pymupdf-utilities.

You already have found the doc-browser.py script, which contains the ability to zoom into a page - however at a fixed constant rate: the available window for page content in zooming mode, can be used for showing one of the 3 x 3 parts of the original page.

Look at the logic to achieve zooming: Use a matrix mat = fitz.Matrix(zoom, zoom) when making a pixmap: page.get_pixmap(matrix=mat,...).
The zoom factor is a float which magnifies (or shrinks) if >1 or <1. So if zoom=2 then magnification is 200% etc.
Assuming your display window has the dimensions WIDTH and HEIGHT, and the sub-rectangle of the page you want to show is rect = fitz.Rect(x0, y0, x1, y2), then do this to calculate the correct zoom factor:

if rect.width / rect.height < WIDTH / HEIGHT:
    zoom = HEIGHT / height
else:
    zoom = WIDTH / width
mat = fitz.Matrix(zoom, zoom)
pix = page.getPixmap(matrix=mat, clip=rect, alpha=False)

from pymupdf-utilities.

I'm aware on how to achieve zoom by changing the fitz.matrix. But, what I'm asking is how to get a better resolution at 1x Zoom as shown in the image, where Sumatra gives a better resolution than the doc-browser. If I do a 2x Zoom, I can get a better resolution but I can only see a part of the PDF. I'm a beginner at all of this and I'm confused on how to work with tkinter to fit a higher resolution image to the display window where every part is visible. I hope you understood.

from pymupdf-utilities.

I understand. I am not sure either.
You can try the following things:

• change the image format handed to tkinter from "ppm" to "png" and hope this will trigger TK using a different algorithm.
• simply use a larger zoom factor (e.g. twice at large than computed in previous post) and then see what TK does with it.
• if the preceeding doesn't work, use Pillow's thumbnail to shrink the image so it fits into TK's window. But after first try, the previous seemss to work ...

from pymupdf-utilities.

Just looked up Pillow again.
You can do e.g. this:

• compute the pixmap with a doubled zoom factor
• convert it to a Pillow Image img, and then use img.resize(size, resample=3, box=None, reducing_gap=1). Play with the resample and reducing_gap values to meet your requirements.

set size =(img.width//2, img.height//2) to restore the original dimensions.

from pymupdf-utilities.

Just looked up Pillow again.
You can do e.g. this:

• compute the pixmap with a doubled zoom factor
• convert it to a Pillow Image img, and then use img.resize(size, resample=3, box=None, reducing_gap=1). Play with the resample and reducing_gap values to meet your requirements.

set size =(img.width//2, img.height//2) to restore the original dimensions.

I've tried this and it doesn't work. The final quality is the same as when I just used a zoom factor of 1. I guess I have to read more about downscaling an image. Here's the sample code which I've used:

import fitz
from PIL import Image

pdf = fitz.open('test.pdf')
page_number = 5
page = pdf.loadPage(page_number)

scaling_matrix = fitz.Matrix(2, 2)
pix = page.getPixmap(matrix=scaling_matrix, alpha=False)

img = Image.frombytes('RGB', [pix.width, pix.height], pix.samples)
img.show()
img = img.resize((img.width // 2, img.height // 2),resample=3, box=None, reducing_gap=1)
img.show() 

from pymupdf-utilities.

Woah! How interesting - never heard of this before.
Thansk for digging this out!

from pymupdf-utilities.