20 releases (9 breaking)

0.10.3 Sep 4, 2024
0.9.0 Aug 28, 2024
0.3.0 Jul 31, 2024

#106 in Email

Download history 76/week @ 2024-07-20 375/week @ 2024-07-27 714/week @ 2024-08-03 34/week @ 2024-08-10 3/week @ 2024-08-17 198/week @ 2024-08-24 273/week @ 2024-08-31 19/week @ 2024-09-07 7/week @ 2024-09-14 36/week @ 2024-09-21 8/week @ 2024-09-28

85 downloads per month

Custom license

60KB
1.5K SLoC

ma

Mail Archivist. Downloads all mail from multiple accounts into a SQLite database, for backup and analysis.

Usage

Environment

  1. install rust: https://www.rust-lang.org/tools/install
  2. ensure $HOME/.cargo/bin is in your $PATH
  3. cargo install ma

Initial

  1. cd $mail_archive_directory (any directory in which you want ma to maintain the mail archive)
  2. ma fetch (will generate initial config file and then exit with a failure to connect)
  3. Create an App Password with your provider, for example:
  4. $EDITOR ma.toml (fill in the correct values in the generated config file)
  5. ma fetch (should now work)

Routine

Like with git, you can either cd into $mail_archive_directory or provide it to ma as an argument, like so:

  • ma --dir $mail_archive_directory fetch to update
  • sqlite3 $mail_archive_directory/ma.db to enjoy exploring your mail archive with SQL!

TODO

  • fetch all messages from all mailboxes from all accounts
  • store raw messages in content-addressed file tree (dump/[hash..2]/[hash].eml.gz)
  • insert headers and text body into SQLite
  • store state, the highest seen msg per account per mailbox, and avoid re-downloads
  • fetch directly to database and rebrand file-tree storing as export
  • snapshot (log?) mailboxes and message locations
  • poll/idle for new messages (maybe not necessary, since once can just periodically re-fetch)
  • post-update hooks (Can be used for custom notifications, aggregate query reruns, etc.)
  • timeouts
  • parallelize fetch
  • parallelize import/export
  • example analytics
    • Received based route trace graph
    • From -> To directed graph with edges weighted by
      • frequency
      • msg size
  • export into InfluxDB or something similar that Grafana can read from

Questions

Automatic Address Book

How to measure confidence in sender/receiver identity, given that:

  • any name can have any address
  • any address can have any name

Dependencies

~73MB
~1.5M SLoC