JuPDF

JuPDF is a small Python package to convert .ipynb files to single .pdf files via the use of Pandoc.

Supported Features

• All standard markdown features supported in Pandoc conversion, such as lists, tables, etc.
• PNG images within markdown (i.e. ![...](...png)).
• LaTeX mathematics, including in-line LaTeX.
• stdout code-cell output.
• PNG code-cell output (e.g. the output of plot.show() from matplotlib).
• YAML metadata for use with the eisvogel.tex template.

Other image types are planned to be supported in future versions. For now, please ensure both markdown images and code output images are .png.

Requirements

• In order to use JuPDF, as well as any dependencies handled by pip, you must have Pandoc installed on your system, as the conversion process utilizes the pandoc command.

• You must also have a TeX engine installed on your system. For example, on Windows, I use MikTeX.

---

Basic Usage

Converting a single .ipynb file to .pdf

from jupdf.pypdfnb import PYPDFNB
from jupdf.pypdfnb_jobs import single_to_pdf

pypdfnb = PYPDFNB()
pypdfnb.read_ipynb('notebook.ipynb')
single_to_pdf(pypdfnb, 'notebook.pdf')

Converting multiple .ipynb files to .pdf

from jupdf.pypdfnb import PYPDFNB
from jupdf.pypdfnb_jobs import multiple_to_pdf

pypdfnb_a, pypdfnb_b = PYPDFNB(), PYPDFNB()
pypdfnb_a.read_ipynb('notebook_a.ipynb')
pypdfnb_b.read_ipynb('notebook_b.ipynb')
multiple_to_pdf([pypdfnb_a, pypdfnb_b], 'notebook.pdf')

Setting a PYPDFNB instance's contents property back to []

from jupdf.pypdfnb import PYPDFNB

pypdfnb_instance = PYPDFNB()
pypdfnb_instance.read_ipynb('notebook.ipynb')

# do something . . .

pypdfnb_instance.empty()

---

eisvogel.tex Metadata

Before converting a PYPDFNB instance to a .pdf file, you can set various YAML metadata attributes of the instance that will affect how Pandoc converts the notebook markdown to PDF using the eisvogel.tex template.

The code below shows the YAML metadata properties available for PYPDFNB instances, and their values on initialization.

def __init__(...):

    # other attributes ...

    self.title: Optional[str] = None
    self.author: Optional[str] = None
    self.date: Optional[str] = None
    self.subject: Optional[str] = None
    self.keywords: Optional[list[str]] = None
    self.lang: Optional[str] = None
    self.listings: bool = False
    self.titlepage: bool = False

• title - setting this will place the string in the top-left of the PDF pages.
• author - setting this will place the string in the bottom-left of the PDF pages.
• date - setting this will place the string in the top-right of the PDF pages.
• subject - non-visual - the subject of the PDF document.
• keywords - non-visual - keywords associated with the PDF document.
• lang - non-visual - the language code of the document (e.g. 'en').
• listings - whether or not to use Pandoc listings during conversion.
• titlepage - whether or not to insert a title page at the start of the PDF, which will include the title, author and date.

---

Parsers

JuPDF provides a few different parsing Callables that can be passed to PYPDFNB instances. These callables determine how a .ipynb is read, therefore determining how a PDF will look following conversion.

There are currently two distinct types of parsers: cell parsers and code parsers. Cell parsers will determine how cells are placed within the documuent, where as code parsers determine how code cells should be handled within the document.

Callable	Type	Description
cell_parser_regular	cell	Parses cells such that cells are placed in the next available space in a PDF.
cell_parser_one_cell_per_page	cell	Parses cells such that every cell ends with a page break.
cell_parser_one_md_cell_per_page	cell	Parses cells such that each markdown cell specifically starts with a page break.
code_parser_regular	code	Parses code cells such that both the code itself and the code's output are included within the PDF.
code_parser_source_only	code	Parses code cells such that only the code itself is included within the PDF.
code_parser_output_only	code	Parses code cells such that only the code's output is included within the PDF.

An example of using these parsers is shown below.

# Convert a Jupyter Notebook to PDF, whereby each cell starts on a seperate page, and only the output of
# code cells is included in the PDF.

from jupdf.pypdfnb import PYPDFNB
from jupdf.pypdfnb_parsing import cell_parser_one_cell_per_page, code_parser_output_only
from jupdf.pypdfnb_jobs import single_to_pdf

pypdfnb = PYPDFNB(cell_parser_one_cell_per_page, code_parser_output_only)
pypdfnb.read_ipynb('notebook.ipynb')
single_to_pdf(pypdfnb, 'notebook.pdf')

---

Saving time with Saved Parses

Suppose you have a massive .ipynb file that you think you'll need to convert several times. If the file is big enough, then reading and parsing the file may take some time. As such, you likely do not wish to repeat this process again and again. This is where .pypdfnb files come into play.

Using the write_pypdfnb instance method, you can write the current contents of a PYPDFNB instance to a .pypdfnb file. Now, whenever you need to convert that massive file, you can use the read_pypdfnb instance method instead of read_ipynb, which will require no parsing.

from jupdf.pypdfnb import PYPDFNB
from jupdf.pypdfnb_jobs import single_to_pdf

pypdfnb = PYPDFNB()
pypdfnb.open_ipynb('massive_notebook.ipynb')

# writes massive_notebook.pypdfnb to a pypdfnbs directory - this method handles the .pypdfnb extension for you!
pypdfnb.write_pypdfnb('massive_notebook', dir='pypdfnbs')

pypdfnb.empty()


# Later on . . .
pypdfnb.read_pypdfnb('pypdfnbs/massive_notebook.pypdfnb')  # No time spent parsing!
single_to_pdf(pypdfnb, 'massive_notebook.pdf')
pypdfnb.empty()