Add Benchmarks for RRT

[eholum]

Not really what I was supposed to be doing but at least it would catch if things are broken. Plus I think they would be useful...

-------------------------------------------------------------------------------------------- benchmark: 6 tests --------------------------------------------------------------------------------------------
Name (time in ms)                           Min                 Max                Mean             StdDev              Median                IQR            Outliers      OPS            Rounds  Iterations
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
test_benchmark_rrt_connect[kinova]      22.7227 (1.0)       64.8647 (1.10)      38.9695 (1.0)      10.8353 (3.33)      37.7636 (1.0)      14.6174 (3.28)          6;0  25.6611 (1.0)          23           1
test_benchmark_rrt[kinova]              46.5273 (2.05)      58.7023 (1.0)       52.2198 (1.34)      3.2497 (1.0)       52.3314 (1.39)      4.4541 (1.0)           3;0  19.1498 (0.75)         12           1
test_benchmark_rrt_connect[ur5]         93.4375 (4.11)     187.5150 (3.19)     129.8622 (3.33)     33.1143 (10.19)    117.8829 (3.12)     50.7555 (11.40)         3;0   7.7005 (0.30)         10           1
test_benchmark_rrt[ur5]                119.8362 (5.27)     134.7720 (2.30)     127.2674 (3.27)      5.1344 (1.58)     127.2968 (3.37)      7.5349 (1.69)          3;0   7.8575 (0.31)          8           1
test_benchmark_rrt_connect[franka]     192.6760 (8.48)     359.8098 (6.13)     254.2296 (6.52)     66.5204 (20.47)    239.0451 (6.33)     94.2356 (21.16)         1;0   3.9335 (0.15)          6           1
test_benchmark_rrt[franka]             227.0057 (9.99)     248.0522 (4.23)     240.8397 (6.18)      8.1034 (2.49)     243.9057 (6.46)      7.2561 (1.63)          1;0   4.1521 (0.16)          5           1
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

[sea-bass]

Should be range(iterations) yeah?

[sea-bass]

I feel like I want this to land before the continuous joint expanding PR to know how much worse it's going to make the results (if even noticeable)

[eholum]

I've played around with, but haven't actually connected, two options here for continuous benchmark monitoring (assuming we want this)...

1. https://github.com/benchmark-action/github-action-benchmark, which can store benchmarking results in gh-pages. It'll be a little annoying to configure so if we go this route it's probably easier to just wait until the repo is public. On the plus side we wouldn't need a separate app and it supports more benchmarking suites out of the box. On the down side it's more to configure.

2. https://github.com/CodSpeedHQ/action, third part app to which benchmarking results are uploaded and are viewable on their dashboard. It's pretty easy to setup and use but requires granting access to our repo.

Thoughts? Or do we just want the benchmarks in place and then can come back to this later and rely on checking manually?

[eholum]

Sample output here.

https://github.com/open-planning/roboplan/actions/runs/18103021233/job/51510609862?pr=67#step:6:27

If this is enough for now then we can call it enough for now...

[sea-bass]

I think waiting for the repo to be public is good for me. We can just make this a non-required check for now.

As for 1 or 2, I think whichever is easier tbh