Further to my limabean announcement a couple of weeks ago, I present to you 🥁
https://github.com/tesujimath/limabean-harvestlimabean-harvest is a new Beancount importer, built on similar principles to
limabean, that is, a mix of Rust and Clojure.
Its compelling (to me) features are:
- configuration as data not code
- inference of secondary accounts from payee and narration fields, which in particular enables:
-
pairing of transactions across accounts where both accounts are imported in the same group
I am aware that the first point is contrary to
the experience of beancount_reds_importers in particular, so this approach may be insufficiently flexible in the general case. But for now, I am pursuing this approach of adding institution support by means of minimal field mapping in configuration, as in
this config example. Consider this to be a further exploration of the design space.
It is usable out of the box for OFX imports. CSV will always need some configuration, but here limited to field mapping.
Feedback welcome.
Happy Beancounting, as always!