Converting Files¶
Files to PDF¶
Document types supported by PyMuPDF can easily be converted to PDF by using the Document.convert_to_pdf() method. This method returns a buffer of data which can then be utilized by PyMuPDF to create a new PDF.
Example
import pymupdf
# Convert Markdown to PDF
md_doc = pymupdf.open("example.md")
pdfdata = md_doc.convert_to_pdf()
pdf_doc = pymupdf.open(stream=pdfdata)
pdf_doc.save("example.pdf")
# Convert XPS to PDF
xps = pymupdf.open("input.xps")
pdfdata = xps.convert_to_pdf()
pdf = pymupdf.open(stream=pdfdata)
pdf.save("output.pdf")
Markdown to PDF¶
As Markdown files are supported input files they can be easily converted to PDF using the Document.convert_to_pdf() method.
In the simplest case you can just open the Markdown file and call the method to get a PDF representation of the content.
Defining paper size¶
The default paper size is 400 x 600 Rect but you can specify a custom paper size if you wish, to do this just send through the Rect parameter as required, for example:
md_doc = pymupdf.open("example.md", rect=pymupdf.paper_rect("A4")) # A4 size
Defining CSS¶
By default, the Markdown content will be converted to PDF using a default CSS stylesheet. However, you can specify your own CSS stylesheet to customize the appearance of the resulting PDF. To do this, define your css and apply it.
For example, to make all h1 headers red (The single # symbol in Markdown), you could do the following:
md_doc = pymupdf.open( # open the Markdown document in A4 size
"example.md",
rect=pymupdf.paper_rect("A4")
)
css = "h1 {color:red;}"
md_doc.apply_css(css)
pdf_doc = pymupdf.open(stream=md_doc.convert_to_pdf())
pdf_doc.ez_save("red-colored-header.pdf")
Note
The support for CSS is currently limited.
Using CSS Custom Element Selectors¶
Markdown input supports custom element selectors in the CSS, so you can define your own tags in the Markdown and then refer to them in the CSS.
In this way you can specify custom styles for specific elements in the Markdown content. For example, you could define a custom tag called mytag in the Markdown and then refer to it in the CSS to make it red:
css = """
mytag {
color: red;
}
"""
And the corresponding Markdown:
# This is a header
This is some text.
<mytag>This text will be red.</mytag>
This is particularly useful for defining image sizes. For example, you could define a custom class called my_image_class in the CSS and then refer to it in the Markdown to style images:
css = """
my_image_class img {
width: 100px;
height: 100px;
}
"""
With the corresponding Markdown:
# This is a header
This is some text.
<my_image_class><img src="pie-chart.png" /></my_image_class>
Defining Fonts¶
Fonts can be defined by using the Archive parameter to provide a custom Archive containing the font files.
The fonts must exist in an archive which is provided to the Archive parameter when opening the Markdown file. The CSS can then refer to these fonts by their names as defined in the archive.
For example, assuming you have access to the source files for the “Comic Sans” font for all text, you could do the following:
# Global CSS instructions to use the "Comic Sans" font for all text. The font files must be provided in the archive.
css = """
@font-face {font-family: sans-serif; src: url(comic.ttf);}
@font-face {font-family: sans-serif; src: url(comicbd.ttf); font-weight: bold;}
@font-face {font-family: sans-serif; src: url(comicz.ttf); font-weight: bold; font-style: italic;}
@font-face {font-family: sans-serif; src: url(comici.ttf); font-style: italic;}
"""
archive = pymupdf.Archive("C:/Windows/Fonts") # the fonts are here
archive.add(".") # we've stored the archive image in this script's folder
md_file = "sample.md"
md_doc = pymupdf.open( # open the Markdown document
md_file,
archive=archive, # where to look for resources (fonts, images)
rect=pymupdf.paper_rect("A4"), # page dimension ISO A4
)
md_doc.apply_css(css)
PDF to Markdown¶
By utlilizing the PyMuPDF4LLM API we are able to convert PDF to a Markdown representation.
Example
import pymupdf4llm
import pathlib
md_text = pymupdf4llm.to_markdown("test.pdf")
print(md_text)
pathlib.Path("4llm-output.md").write_bytes(md_text.encode())
PDF to SVG¶
Technically, as SVG files cannot be multipage, we must export each page as an SVG.
To get an SVG representation of a page use the Page.get_svg_image() method.
Example
import pymupdf
doc = pymupdf.open("input.pdf")
page = doc[0]
# Convert page to SVG
svg_content = page.get_svg_image()
# Save to file
with open("output.svg", "w", encoding="utf-8") as f:
f.write(svg_content)
doc.close()
PDF to DOCX¶
Use the pdf2docx library which uses PyMuPDF to provide document conversion from PDF to DOCX format.
Example
from pdf2docx import Converter
pdf_file = 'input.pdf'
docx_file = 'output.docx'
# convert pdf to docx
cv = Converter(pdf_file)
cv.convert(docx_file) # all pages by default
cv.close()