Alexander Lenail

Good idea. A good way to do this would be to expose the choice of whether to have the [pcst_fast](https://github.com/fraenkel-lab/OmicsIntegrator2/blob/82245952a555c20e34b346009e3d59c956be99d9/src/graph.py#L295) routine be rooted or unrooted in the [pcsf](https://github.com/fraenkel-lab/OmicsIntegrator2/blob/82245952a555c20e34b346009e3d59c956be99d9/src/graph.py#L251) method.

Do you want to take this on, @iamjli ?

@sgosline thanks for the suggestion. [pcst_fast](https://github.com/fraenkel-lab/pcst_fast) has a num_clusters option, which specifies a hard constraint on the number of trees in the output forest. The dummy node / omega-weighted edges...

@AmandaKedaigle I'd love some more details on this. First, could you report changes you'd like to see in Garnet on the garnet repo? (https://github.com/fraenkel-lab/GarNet) and then also spell out the...

@AmandaKedaigle okay, changes have been made to GarNet as part of https://github.com/fraenkel-lab/GarNet/pull/30. On this repo, I started this work as part of #37. I think that you'll be the best...

np. Have a good trip!

I'm open to it. Would gladly review a PR.

approve

Note: `LSODA` has an `events` argument for this purpose underneath, but it isn't exposed by the scipy wrapper: https://github.com/scipy/scipy/blob/4ec4ab8d6ccc1cdb34b84fdcb66fde2cc0210dbf/scipy/integrate/lsoda.pyf#L27 However, R's deSolve and others _do_ expose this option (see `events`...

@drhagen you're right about the adjoint method, I had been mis-reading the math. I've moved on to using diffeqpy for my particular project, but would still love to see this...