aepsych/server.py (5 lines):
	- line 177: TODO actually make an abstract interface to subclass from here
	- line 215: TODO make things a little more robust to bad messages from client; this
	- line 283: # TODO this should probably be the other way around when we refactor
	- line 655: if x is None:  # TODO: ensure if x is between lb and ub
	- line 761: TODO better types


aepsych/benchmark/problem.py (2 lines):
	- line 166: # TODO handle the special cases for no contextual dims at some point
	- line 193: # assume mono_dim is last dim (TODO make this better)


aepsych/strategy.py (2 lines):
	- line 112: # TODO: allow user to pass in generator options
	- line 203: strat_list (list[Strategy]): TODO make this nicely typed / doc'd


aepsych/benchmark/benchmark.py (1 line):
	- line 34: TODO make a benchmarking tutorial and link/refer to it here.


aepsych/models/gp_classification.py (1 line):
	- line 116: from a configuration. TODO: document how this works in some tutorial.


clients/unity/Assets/Scripts/AEPsychClient.cs (1 line):
	- line 24: // TODO make this a little more strongly typed


aepsych/acquisition/mutual_information.py (1 line):
	- line 124: # TODO this is identical to nono-monotonic BALV acquisition with a different


aepsych/models/base.py (1 line):
	- line 16: torch.set_default_dtype(torch.double)  # TODO: find a better way to prevent type errors


aepsych/acquisition/rejection_sampler.py (1 line):
	- line 65: ), "Batches not supported"  # TODO T68656582 handle batches later


aepsych/acquisition/monotonic_rejection.py (1 line):
	- line 69: # TODO T68656582 support batches properly.


aepsych/factory/factory.py (1 line):
	- line 27: TODO write a modular AEPsych tutorial.


aepsych/plotting.py (1 line):
	- line 104: # TODO implement 3d plots


aepsych/config.py (1 line):
	- line 35: TODO write a tutorial on writing configs.