There are multiple ways to reduce the size of a given PDF file. The easiest one is to remove content (e.g. images) or pages.
Some PDF documents contain the same object multiple times. For example, if an image appears three times in a PDF it could be embedded three times. Or it can be embedded once and referenced twice.
When adding data to a PdfWriter, the data is copied while respecting the original format. For example, if two pages include the same image which is duplicated in the source document, the object will be duplicated in the PdfWriter object.
Additionally, when you delete objects in a document, pypdf cannot easily identify whether the objects are used elsewhere or not or if the user wants to keep them in. When writing the PDF file, these objects will be hidden within (part of the file, but not displayed).
In order to reduce the file size, use a compression call: writer.compress_identical_objects(remove_identicals=True, remove_orphans=True)
remove_identicalsenables/disables compression merging identical objects.remove_orphansenables/disables suppression of unused objects.
It is recommended to apply this process just before writing to the file/stream.
It depends on the PDF how well this works, but we have seen an 86% file reduction (from 5.7 MB to 0.8 MB) within a real PDF.
from pypdf import PdfWriter
writer = PdfWriter(clone_from="example.pdf")
writer.remove_images()
with open("out.pdf", "wb") as f:
writer.write(f)If we reduce the quality of the images within the PDF, we can sometimes reduce the file size of the PDF overall. That depends on how well the reduced quality image can be compressed.
from pypdf import PdfWriter
writer = PdfWriter(clone_from="example.pdf")
for page in writer.pages:
for img in page.images:
img.replace(img.image, quality=80)
with open("out.pdf", "wb") as f:
writer.write(f)pypdf supports the FlateDecode filter which uses the zlib/deflate compression method. It is a lossless compression, meaning the resulting PDF looks exactly the same.
Deflate compression can be applied to a page via
{meth}page.compress_content_streams <pypdf._page.PageObject.compress_content_streams>:
from pypdf import PdfWriter
writer = PdfWriter(clone_from="example.pdf")
for page in writer.pages:
page.compress_content_streams() # This is CPU intensive!
with open("out.pdf", "wb") as f:
writer.write(f)page.compress_content_streams uses zlib.compress
and supports the level parameter: level=0 means no compression,
level=9 refers to the highest compression.
Using this method, we have seen a reduction by 70% (from 11.8 MB to 3.5 MB) with a real PDF.
When a page is removed from the page list, its content will still be present in the PDF file. This means that the data may still be used elsewhere.
Simply removing a page from the page list will reduce the page count but not the file size. In order to exclude the content completely, the pages should not be added to the PDF using the PdfWriter.append() function. Instead, only the desired pages should be selected for inclusion (note: PR #1843 will add a page deletion feature).
There can be issues with poor PDF formatting, such as when all pages are linked to the same resource. In such cases, dropping references to specific pages becomes useless because there is only one source for all pages.
Cropping is an ineffective method for reducing the file size because it only adjusts the viewboxes and not the external parts of the source image. Therefore, the content that is no longer visible will still be present in the PDF.
The presentation Putting a Squeeze on Your PDF has other suggestions. One takeaway is that most of the significant size optimizations usually come from image and font modification. However, font optimization, such as replacing, merging, and subsetting, is not within the functionality of pypdf at the moment.