There's a bit of uncertainty right now whether conversion to synapser is right yet given that it removes some functionality and hasn't seen updates/maintenance in 5 months. One motivation for synapser was the thought that we would benefit from new developments, but if we try to do add dependencies/complexity while losing functionality instead, I think we should just stick with reticulate to retain flexibility and control for now. Hence I've moved the WIP synapser work to https://github.com/nf-osi/develop-synapser, to be able to focus on needed functional updates and new features rather than migration. This reverts commit 4320d35b99b48466de41c5337e0ff6dd8da5b2f4, which was the first commit for the transition, and since it's only one commit back this makes a relatively easy "downclimb".
There's a bit of uncertainty right now whether conversion to synapser is right yet given that it removes some functionality and hasn't seen updates/maintenance in 5 months. One motivation for synapser was the thought that we would benefit from new developments, but if we try to do add dependencies/complexity while losing functionality instead, I think we should just stick with reticulate to retain flexibility and control for now. Hence I've moved the WIP synapser work to https://github.com/nf-osi/develop-synapser, to be able to focus on needed functional updates and new features rather than migration. This reverts commit 4320d35b99b48466de41c5337e0ff6dd8da5b2f4, which was the first commit for the transition, and since it's only one commit back this makes a relatively easy "downclimb".