Athens exporter

Standalone tool that converts an Athens Research database into a logseq directory.

Usage

The easiest way is to download a released JAR, and run it with

$ wget https://github.com/bshepherdson/athens-export/releases/latest/download/athens-export.jar
$ java -jar athens-export.jar path/to/athens/index.transit logseq/parent/dir

To run from source: install the Clojure CLI tool, clj, clone this repo and cd into it. Then:

$ clj -X athens.export/export \
  :athens '"path/to/athens/index.transit"' \
  :logseq '"path/to/empty/logseq/dir"'

(Yes, with the weird double quoting - we want to pass the literal "s to Clojure, so we wrap them with ' so the shell doesn't eat them.)

Features

This transforms an Athens database and exports it as Markdown files in a directory tree, suitable for importing into logseq.

• Pages named eg. July 16, 2021 in Athens are recognized as journal pages and written accordingly for logseq (eg. journals/2021_07_16.md).
• Block references are detected, and rewritten with logseq-style UUIDs.
• File names and title:: metadata are properly written
  • / becomes .
  • : becomes _
  • . is preserved but the title:: is written explicitly

Caveats

There are several known weaknesses of this tool.

• TODO/DONE in Athens is rudimentarily converted to logseq's format using TODO and DONE markers. Keep in mind that logseq will only recognize them if they are the begining of a block (Athens allows you to randomly add them anywhere on the block, if you rely on that, you will have to manually fix your files).
• iframe and YouTube embeds are not converted
  • they are visible (flagged as errors) in logseq, and can be edited to work.
• code blocks are clumsy in Athens, and show up misformatted in logseq
  • but they render legibly, and can be fixed by hand in logseq.
• there may be more file name special characters I don't know about

PRs to fix any of these issues would be welcome!