| 4 Nov 2018 |
| 01:06:55 | |  ebretl set a profile picture. |
| 01:06:55 | ebretl | Note on general software training: this coming week will be our last week of training |
| 6 Nov 2018 |
| 20:12:24 |  jneiger3 | Hey guys! I just wrote up the answer for the pass or shoot exercise that we did last week. There is more than one right answer so don't worry if it doesn't match what you have. Let me know if you have any questions or anything. https://gist.github.com/JNeiger/73802a57607539b768618fc2d5ec752e |
| 20:55:32 |  alan.penkar | Are using a linter or do you just write very neat Python? |
| 20:56:51 | jneiger3 | I usually don' |
| 20:57:35 | jneiger3 | *I usually don't but we do have one that runs if you use the command "make pretty". |
| 21:02:33 | alan.penkar | 1 - You mention not being able to call self.all_subbehaviors_completed (or whatever the exact function is) because there are also defensive behaviors, but on exiting setup you call self.remove_all_subbehaviors() - why is that allowed? |
| 21:05:17 | alan.penkar | 2- is there a stylistic preference for more readable functions with the state transitions vs. just saying for example self.target_pass_pos is None ? |
| 21:06:58 |  jgkamat | 1 - the defense subbehavior is added back every frame if it was removed : https://github.com/RoboJackets/robocup-software/blob/master/soccer/gameplay/standard_play.py#L36 |
| 21:07:23 | jgkamat | (maybe all_subbehaviors_completed shouldn't take into account the defense play by default or something to make that easier) |
| 21:07:31 | jneiger3 | 1. I'm kinda cheating here. The proper way to remove subbehaviors is using their name. If you do remove_all_subbehaviors here, it removes the defense also, but the next frame this code runs and adds it back. Let me change it so it makes it clearer |
| 21:07:56 | alan.penkar | ahh, gotcha |
| 21:10:41 | jneiger3 | 2. It's really a personal preference. I don't like when they start hitting 5 or 6 lines, so I try to cut things out of the transition using a function. As long as it looks clean, it's really up to you |
| 21:16:07 | alan.penkar | so for setup->shoot transition, if I'd used lambda: self.target_pass_pos is None. Is that clean to you or more problematic since you may have to lookup what it is? |
| 21:21:47 | jneiger3 | That would be fine with me. If you look at aim.py in the skills folder. If we didn't abstract away the is_aimed, is_steady, fumbled and is_desperate, that lambda for aiming->aimed would be like 20 lines long and would be pretty ugly. If it's less than 3 or 4 lines, it's fine generally. |
| 7 Nov 2018 |
| 00:01:51 |  tjones320 | Do this |
| 00:01:55 | tjones320 | https://goo.gl/forms/euBtwZBXyIBzLh6t2 |
| 00:02:06 | tjones320 | It is required for all new members!!! |
| 00:03:29 | tjones320 | @room |
| 15:51:46 |  clindbeck3 | If I recall face commands are not subbehaviors right? Is there any way to tell when a face command is considered finished aside from just checking for an arbitrary rotational error threshold? |
| 15:53:04 | clindbeck3 | Nvm, there is a face subbehavior you can use |
| 15:57:02 | clindbeck3 | Do we actually use the "face.py" move and face behavior? It uses a 3 degree threshold. I'm thinking about the correct way to benchmark rotation and rotation motion combinations. |
| 15:59:57 | jneiger3 | I don't think so. I believe we only use robot.face(pt) and that's it. |
| 15:59:58 | jgkamat | If you are benchmarking the behavior you should probably use the underlying cpp methods instead to be more accurate |
| 8 Nov 2018 |
| 14:50:30 |  guyfleeman | 2YQPBPM2 29BB7JE8 is this something that we think we'd want to do? |
| 14:52:00 | guyfleeman | Maybe CP1F9S0Z could participate with his sim? |
| 14:52:12 | guyfleeman | Might be a fair bit of work this |
| 15:27:12 | jneiger3 | The only thing I can think of that would work well in simulink is the controls stuff. But I'm not sure if that is worth my time to work on instead of the vision filter. If someone else want to learn simulink though, I can walk you through it |
| 30 Nov 2018 |
| 16:12:57 | jneiger3 | 0JDCA057 What's the new checks tab in PR's on github? From what I can tell it's the same as the checks at the bottom, but you get more information in github about what fails? |
| 2 Dec 2018 |
| 03:15:53 | jgkamat | Circle ci is now adding those for us so that's probably added by them directly |
