I setup a forward to move my html only emails to an email address that this script processes. The colon in the subject line (Fwd:) breaks the file naming because files cannot contain non-standard characters, the file is saved simply as "Fwd".

It's not producing much error details only: email2pdf | standard_init_linux.go:228: exec user process caused: exec format error

My compose section:

  email2pdf:
    image: ghcr.io/rob-luke/emails-html-to-pdf:latest
    container_name: email2pdf
    environment:
      - IMAP_URL=xxxx
      - IMAP_USERNAME=xxxx
      - IMAP_PASSWORD=xxxx
      - IMAP_FOLDER=paperless
      - SMTP_URL=xxxx
      - MAIL_SENDER=xxxx
      - MAIL_DESTINATION=xxxx
      - INTER_RUN_INTERVAL=300
      - PRINT_FAILED_MSG='false'

This PR fixes #11

This PR introduces 2 new Environment Variables:

HOSTS

This var is a semicolon separated list of hosts that should be added to /etc/hosts to prevent dns lookup failures. e.x.: HOSTS=127.0.0.1 tracking.paypal.com;127.0.0.1 my.custom.host.tld

WKHTMLTOPDF_OPTIONS

This var is a python dict (json) representation of wkhtmltopdf_options that can be passed to the used pdfkit library.

e.x.: WKHTMLTOPDF_OPTIONS='{"load-media-error-handling":"ignore"}'

More options for wkhtmltopdf can be found here. More about the usage of those options with pdfkit can be found here

Examples

I had the problem that the tracking pixel of PayPal caused a HostNotFoundError. This was because the container was not able to resolve the tracking.paypal.com domain.

With this PR, I would add following to the docker-compose.yml:

...
    environment:
        ...
        HOSTS=127.0.0.1 tracking.paypal.com
        WKHTMLTOPDF_OPTIONS={"load-media-error-handling":"ignore"}

I tried to send a Paypal receipt email to Paperless via the Email to PDF and got the below error. I don't understand why it's trying to load this page instead of the actual email... thoughts?

email2pdf    |
email2pdf    | No attachments in: You have authorized a payment to XXXXXXXX Inc.
email2pdf    |
email2pdf    | PDF: You-have-authorized-a-payment-to-XXXXXXXX_.pdf
email2pdf    |
email2pdf    | PDF: You-have-authorized-a-payment-to-XXXXXXXX_.pdf
email2pdf    | Traceback (most recent call last):
email2pdf    |   File "/app/main.py", line 98, in <module>
email2pdf    |     process_mail(imap_url=server_imap,
email2pdf    |   File "/app/main.py", line 72, in process_mail
email2pdf    |     pdfkit.from_string(html, filename)
email2pdf    |   File "/usr/local/lib/python3.9/site-packages/pdfkit/api.py", line 72, in from_string
email2pdf    |     return r.to_pdf(output_path)
email2pdf    |   File "/usr/local/lib/python3.9/site-packages/pdfkit/pdfkit.py", line 156, in to_pdf
email2pdf    |     raise IOError('wkhtmltopdf reported an error:\n' + stderr)
email2pdf    | OSError: wkhtmltopdf reported an error:
email2pdf    | QStandardPaths: XDG_RUNTIME_DIR not set, defaulting to '/tmp/runtime-root'
email2pdf    | Loading page (1/2)
Error: Failed to load https://t.paypal.com/ts?xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx, with network status code 1 and http status code 0 - Connection refused
Printing pages (2/2)                                                        ] 25%
Done                                                                        ]
email2pdf    | Exit with code 1 due to network error: ConnectionRefusedError
email2pdf    |

• Adds an option to specify a custom search criteria
• Fixes mail being processed multiple times if flag is not "SEEN"
  • Search was always looking for "unseen" mail. If the mail flag was not set to seen, the mail would remain unread and be processed next run. A suitable filter should now be determined, or an error raised if the user must specify one.
• Remove "DRAFT" flag option
  • Causes strange behaviour where inbound mail can be converted to an outbound draft. This can then be accidentally discarded deleting the original message or converting it back can result in the metadata of the email being lost.
• Add "UNFLAGGED" mail flag option
  • Removes the "FLAGGED" option from the message once processed

This looks perfect in combination with paperless, thanks for spending your time creating this!

Sometimes, I do not only get HTML emails but simply text emails where I would need the actual headers (from, date, subject, etc.) - is it possible to include these in the generated PDF?WKHTMLTOPDF options appear to control the PDF's layout, not the content.

I added smtp_username and smtp_password options in case your outgoing email goes through an smtp server with different login credentials than your imap server.

It would be cool to have an option to delete the mails instead of marking them as read. Mails I want to consume will be moved into a "Processing" folder where this container picks them up. Afterwards, they are sent to the paperless instance and can therefore be deleted to keep the mail inbox empty.

Hi,

since this project got some good improvements over the last days, the image should be shipped with a versioning system. Since only the latest Tag is currently used, clients will not pull a later image because the tag name did not change.

My Recommendation:

• create a Github Release (Tag) for every new Feature / Group of Features
• Modify the Buildprocess to get triggered by the "onTag" action of GitHub and use the tagname as Docker tag
• Build the latest tag with the current master (independently of current releases) branch.

Currently, the derived base image (python 3.9) is based on a default debian install and therefore is about 1,5GB tall. We should create an image based on alpine-linux to reduce image size.

I will create an PR tomorrow probably.

Hey there, thanks for your effort! Is it possible to setup the docker-compose.yml variables to define if SSL/TLS os in or off and another option to define the port that should be used?

Hi

Just started to use your docker instance and it seemed to be going great, but then suddenly the container stopped, looking in the logs I saw the following error a few times..

Hi

I’m looking to use something just like this, however it seems the likes of gmail and outlook (hotmail, live) etc. are now no longer supporting basic authentication (username,password) is that your understanding too ? If so what providers can be used with your set up ?

Many thanks

My account has 2FA enabled on it, and I don't have the option to disable 2FA. I'm assuming that is why I'm getting this error, but I don't even get a prompt on my Microsoft Authenticator app to approve the login attempt, so it doesn't seem to be getting that far.

600
Skipping virtualenv creation, as specified in config file.
Running emails-html-to-pdf
Starting mail processing run
    raise self.error(dat[-1])
imaplib.IMAP4.error: b'LOGIN failed.'
Traceback (most recent call last):
  File "/app/main.py", line 257, in <module>
    process_mail(
  File "/app/main.py", line 102, in process_mail
    with MailBox(imap_url).login(imap_username, imap_password, imap_folder) as mailbox:
  File "/usr/local/lib/python3.9/site-packages/imap_tools/mailbox.py", line 44, in login
    login_result = self.box.login(username, password)
  File "/usr/local/lib/python3.9/imaplib.py", line 612, in login
    raise self.error(dat[-1])
imaplib.IMAP4.error: b'LOGIN failed.'

This is a pull request to add the feature requested here: #27

The following ENV variables have been added:

EMAIL_HEADER EMAIL_HEADER_EXT

These are both False by default.

By turning on the Header, you will now have some basic information added to the PDF such as the subject, date/time, sender and recipient. Example:

Right now it displays:

• Subject
• Date
• From
• Reply-to
• To

Additional fields such as CC / BCC should probably be added.

By turning on the Header (extended) it will display the full headers as received. However, it currently does not display them very cleanly. Maybe it would be worthwhile to look into how to make them display better. It could be as simple as adding a <pre> tag.

Tested and works.

I thought I had a failing installation because my first test email was not being converted.

The email had an inline image that was attached to the email. I guess I didn't know that in some cases (Gmail in this instance) an HTML image in the body of the email is also attached to the email itself.

Could there be a way to handle attached images in this manner? Perhaps it allows attachments, but only if the attachment is also used in the body of the message? That way it skips legitimately attached files but does act when the attachment is part of the email body?

My Paperless-ng(x) filters out all non-PDF attachments already, so it wouldn't import the email with the inline images anyways.