pypandoc

If you have installed pypandoc but not pandoc, you can download and install pandoc programmatically. This works on Windows, Mac OS X, and Linux (64-bit Intel). By default, it installs the latest version to a location that is automatically added to the search path.

from pypandoc.pandoc_download import download_pandoc

# Download the latest version to the default path
download_pandoc()

# Download a specific version
download_pandoc(version='1.19.1')

Alternatively, use the CLI:

# Install latest pandoc to default path
pypandoc download

# Download a specific version
pypandoc download --version 3.6

Converting to PDF requires a LaTeX engine. Pypandoc integrates with pytinytex to automate this setup. When converting to PDF, pypandoc will automatically add TinyTeX to the PATH and attempt to install missing LaTeX packages (up to 3 retries) if compilation fails.

# Install pypandoc with tinytex support
pip install pypandoc[tinytex]

# Download TinyTeX once
pytinytex download

import pypandoc
# PDF conversion now works seamlessly
pypandoc.convert_file('document.md', 'pdf', outputfile='document.pdf')

Pypandoc is available on conda-forge. Installing via conda will also install the pandoc package, making it available immediately.

# Install via conda-forge channel
conda install -c conda-forge pypandoc

You can install pypandoc using pip. Note that pypandoc is a thin wrapper and requires a pandoc installation on your system. If you want pandoc included automatically, install the pypandoc_binary package instead.

# Install only the wrapper (requires manual pandoc installation)
pip install pypandoc

# Install the wrapper with pandoc included out of the box
pip install pypandoc_binary

The download_pandoc function uses different unpacking logic based on the operating system:

• Linux: Uses ar and tar to extract .deb packages and copies pandoc and pandoc-citeproc to the target folder.
• macOS (Darwin): Uses pkgutil to expand .pkg files and extracts binaries from the Payload.
• Windows: Uses msiexec with the /a (administrative installation) flag to extract .msi files to a temporary directory.

If pandoc is not in your PATH, or if you need to point to a specific binary (e.g., for a web server with restricted user permissions), set the PYPANDOC_PANDOC environment variable to the full path of the binary. If this variable is set, pypandoc will only search there.

# Via shell
export PYPANDOC_PANDOC=/home/x/whatever/pandoc

# Via Python runtime
import os
os.environ.setdefault('PYPANDOC_PANDOC', '/home/x/whatever/pandoc')

You can run the pypandoc environment using a Docker Compose configuration. The setup includes a binding mount that synchronizes the current host directory with the container's /pypandoc/ directory, allowing code changes on the host to be immediately reflected in the running container. The container is configured to stay running (using tail -F) so that you can interact with it using docker compose exec.

services:
  app:
    image: pypandoc
    container_name: pypan_container
    build: .
    restart: always
    volumes:
    - .:/pypandoc/
    command: tail -F ./anything

Pypandoc uses the standard Python logging library. Messages from Pandoc are sent to the console by default. To mute logs, add a NullHandler to the pypandoc logger before calling any conversion functions.

import logging
# Mute all pypandoc logging
logging.getLogger('pypandoc').addHandler(logging.NullHandler())

Use these utility functions to inspect the environment and available capabilities of the installed Pandoc binary.

import pypandoc

print(pypandoc.get_pandoc_version())  # Get version string
print(pypandoc.get_pandoc_path())     # Get path to binary
print(pypandoc.get_pandoc_formats())  # Get supported formats

You can pass arbitrary Pandoc options using the extra_args parameter (a list of strings) and apply Pandoc filters using the filters parameter (a list of strings).

# Using extra_args for Pandoc options
output = pypandoc.convert_text(
    '<h1>Primary Heading</h1>',
    'md', 
    format='html', 
    extra_args=['--atx-headers']
)

# Using filters (must be a list)
filters = ['pandoc-citeproc']
pdoc_args = ['--mathjax', '--smart']
output = pypandoc.convert_file(
    'filename.md', 
    to='html5', 
    format='md', 
    extra_args=pdoc_args, 
    filters=filters
)