Convert .eml (email) files to PDF using Python, making them easier
to archive, share, and view without requiring an email client.
Depends on GNOME's Pango and various Python libraries but NOT on a full rendering engine like WebKit or Gecko. python-pdfkit and wkhtmltopdf are deprecated libraries
eml2pdf should run on Linux distributions with Pango and Python and macOS. The Pango dependency is a challenge on Windows at the moment.
- Converts email body from HTML or plain text message body.
- Tries to filter potential security or privacy issues.
- Preserves formatting, character encodings, embedded images.
- Generates a header section with email metadata From, To, Subject, Date.
- Generates a list of attachments with size and md5sum. (Attachments are not embedded in the PDF.)
- Python 3.11+
- weasyprint - a visual rendering engine for HTML and CSS that can export to PDF. Weasyprint depends on GNOME's Pango.
- python-markdown - for HTML'izing plain text.
- hurry.filesize - return human readable filesizes.
- beautifulsoup4 - HTML sanitization.
On a desktop system, chances are high that you have Pango installed. In this case you can install eml2pdf from PyPi using pip:
pip install eml2pdfIf weasyprint can't find Pango you can consult weasyprints install help to install weasyprint using your system's package manager.
Users of Arch linux or derived distro's like Manjora can use AUR package eml2pdf.
Check INSTALL.md on
Github or
{ref}install for more detailed installation instructions if you need more
help.
eml2pdf has two modes of operation controlled via subcommands convert_dir and
convert_file. Both modes share options to control page size, perform HTML
sanitization, produce HTML debug output, be more quiet or verbose. Refer to
usage help per subcommand below.
usage: eml2pdf [-h] {convert_dir,convert_file} ...
Convert EML files to PDF
options:
-h, --help show this help message and exit
supported subcommands::
{convert_dir,convert_file}
Use {subcommand} --help for options.
convert_dir Convert all EML files in an input dir to PDF files in an output dir.
convert_file Convert a single EML file to a single PDF
convert_dir will convert all .eml files in an input directory and
save converted PDF files in a specified output directory.
The output filenames are formatted as:
YYYY-MM-DD_subject[-counter].pdf, where:
- The date prefix is taken from the email's sent date.
- The email subject is taken from the email headers.
- Should there be any duplicate filenames, then a counter will be added.
- The extension is changed to
.pdf
For example, some_file.eml with subject "My Email" sent on March 15, 2024
will become 2024-03-15_My_Email.pdf.
convert_dir has -n/--number-of-procs, as a specific option to set the
number of parallel processes. eml2pdf guesses its default value from the
apparent number of CPU's available to the eml2pdf process. If verbose output is
requested then -n is set to 1 or the debugging output from different
subprocesses gets mixed.
$ eml2pdf convert_dir -h
usage: eml2pdf convert_dir [-h] [-p size] [--unsafe] [-d] [-v] [-q] [-n number] input_dir output_dir
positional arguments:
input_dir Directory containing EML files
output_dir Directory for PDF output
options:
-h, --help show this help message and exit
-p, --page size One of a3, a4, a5, b4, b5, letter, legal, or ledger, with or without "landscape", for example: "a4 landscape" or a3. Surround with quotes if there is a space in the argument value.
Defaults to "a4", implying portrait.
--unsafe Don't sanitize HTML from potentially unsafe elements such as remote images, scripts, etc. This may expose sensitive user information.
-d, --debug_html Write intermediate html file next to PDF's
-v, --verbose Show a lot of verbose debugging info. Forces number of procs to 1.
-q, --quiet Show only errors.
-n, --number-of-procs number
Number of parallel processes. Defaults to the number of
available logical CPU's to eml_to_pdf.
Example below renders all .eml files in ./emails to a4 landscape oriented pdf's
in ./pdf:
eml2pdf -p "a4 landscape" ./emails ./pdfsconvert_file works per file, taking the input filename of the EML to convert to PDF and output filename to convert to.
$ eml2pdf convert_file -h
usage: eml2pdf convert_file [-h] [-p size] [--unsafe] [-d] [-v] [-q] input_file output_file
positional arguments:
input_file Input EML file to convert
output_file Output PDF file to convert to
options:
-h, --help show this help message and exit
-p, --page size One of a3, a4, a5, b4, b5, letter, legal, or ledger, with or without "landscape", for example: "a4 landscape" or a3. Surround with quotes if there is a space in the argument value. Defaults
to "a4", implying portrait.
--unsafe Don't sanitize HTML from potentially unsafe elements such as remote images, scripts, etc. This may expose sensitive user information.
-d, --debug_html Write intermediate html file next to PDF's
-v, --verbose Show a lot of verbose debugging info. Forces number of procs to 1.
-q, --quiet Show only errors.
eml2pdf will first parse email header info such as date, subject, etc. Next the mail body will be parsed. If there is an HTML body, eml2pdf will clean this HTML body (ref. below under Security) and prepend this resulting HTML with a summary table.
In a next step this HTML is rendered by weasyprint to a PDF.
The --debug_html flag will save this intermediate HTML. You can use this to
check if there is an email parsing issue in eml2pdf or a PDF conversion issue
in weasyprint.
Not all emails are properly formatted. Part of your mail might not be visible in the pdf in case an email doesn't limit width of some elements such as images, tables or others. You can play with page sizes and orientations to try and accomodate wide emails.
Emails can contain HTML which can contain stuff you don't expect or want.
In the best case your emails contain clean HTML.
In common cases they will contain intentional tracking of end users using forged remote sources for images and other resources. This is a common practice in marketing or mass mailing solutions.
eml2pdf tries to keep the formatting in your mails ánd clean up potentially malicious content using custom filtering of tags, remote images, remote stylesheets, etc.
We try to cleanup. We can't give you a 100% guarantee. If you're very worried, please cleanup your mails yourself.
You can use the --unsafe flag if you don't want eml2pdf to try and
sanitize your mails. Check your mails' content before you use this flag!
eml2pdf lists attachments with their md5sums. You can use these md5sums for your convenience. They give a very strong indication that files are not altered. They will not be usable as proof in courts of law. They are not intended to be.
We've tested eml2pdf with a couple of cases with embedded images, tables, unicode or specific encodings. Refer to tests for example emails.
Please open an issue ticket if you have a mail where conversion results are not usable. Describe what you think your message contains and the output you expect. Attach verbose eml2pdf output of only this eml file and attach the eml file itself. We're not promising a solution, but we can have a look.
Please cleanup any attachments you add. Remove things you don't want to share with the world.
eml2pdf was originally forked from klokie/eml-to-pdf by Daniel Grossfeld.
- Inline non-image attachments - omusale
- convert_file mode and 8bit CTE with UTF-8 encoding - bastidest.
- Fix bug - Header class not renamed correctly benhxy
If you want to work on eml2pdf, read DEVELOPMENT.md on
GitHub or
{ref}development. PR's welcome ;-).
eml2pdf code is published under the MIT license.
Licenses for dependencies:
- weasyprint: BSD-3
- python-markdown: BSD-3
- hurry.filesize: ZPL 2.1
- beautifulsoup4: MIT
- Pango: GPLv2