df7832aa43
This PR replaces our current NPC AI with a [behavior tree system](https://en.wikipedia.org/wiki/Behavior_tree_(artificial_intelligence,_robotics_and_control)). Behavior trees are a common way of creating AI in which you place nodes in a tree structure to define what actions an AI should take. AI controllers defined a list of /datum/ai_planning_subtree types in behavior_nodes. Each subtree was a self-contained unit that could call queue_behavior() to fire off /datum/ai_behavior actions. The controller iterated subtrees in order, each one deciding independently whether to queue something and deciding whether the next subtree would run. This has a few issues: 1. There's no real structure; you are just defining a list of things to try in order. 2. There was a loooot of subtrees that were basically the same as another but with some slight modification 3. It was hard to understand. Controllers now define a single json file describing a tree of nodes. The tree is composed of structural composites: Sequence - do A, then B, then C (and so on) Selector - try A, if it fails try B, then C (and so on) Parallel - run A and B simultaneously, with configurable failure/success policies and or looping behavior Subplan - loop a child continiously Along that we also have "Decorators". These are nodes that basically check a condition (E.g.; do we have a combat target). These decorators can be used to gate behavior and are re-useable across behavior trees. They also have a concept known as "Observers". Which lets them cancel lower priority behavior in case their condition changes (Which we check whenever a signal fires that fits that specific decorator). This makes the AI much more responsive to change in environment. For behaviors, we still use the ai_behavior datums. These are the actual behaviors such as "Move to X", "Attack X". The only major change is that these can no longer sleep() since they now run in the ai_controller. Lastly, we now also have subtrees, except now they are essentially pieces of behavior tree that can be re-used, or even overriden at runtime or as a variable. Allowing for making modular AI made out of several smaller trees. You can set variables on these nodes directly via the extension (see below), which should reduce the need to make subtypes of behaviors by a lot. All of these vars are saved on the JSON and will be applied at runtime. If you are using subtrees, you can also assign "bindings" to these variables, which will allow instances of the subtree to override those variables. Since a tree structure with variables becomes hard to parse in a JSON, I've made a VSCode extension to edit these JSONs: https://marketplace.visualstudio.com/items?itemName=BehaviorTreeG.behaviortreeg https://github.com/CabinetOnFire/BehaviorTreeG <img width="1795" height="1268" alt="image" src="https://github.com/user-attachments/assets/56aa2f0b-3cf9-449f-bca4-8281fca82db6" /> This extension allows you to edit the behavior tree JSONs, and browse through all the behaviors/decorators/subtrees we have If you'd like more info on how to build these AI check out the learn_ai.md. I will also make a tutorial to go over more depth on what the system offers because I kind of suck at doing technical write-ups. Targetting has been changed to. I've made a new acquire_targets behavior that takes a target_source (what am I targetting) and targetting_strategy (what does the candidate need to fulfill to be considered a target). This allows us to make composites targetting combinations to reduce the amount of specific find_and_set esque behaviors we had before. Not everything is ported to this system but that would be a longer term goal. I've added a new build_bt script that converts all the behavior tree JSONs into compiled versions. Why is this needed? Because I wanted to keep using defines in behavior trees, so we need a way to convert this into literal values before we send it to DM. This script runs on compile and should also run in CI (If I didn't fuck that up!). This saves to a new build/ folder. I've ported every single AI in the game to this system (except raptors, Kobsa is working on those so should be in soon!), so I do expect some bugs to come out of this. But I also fixed some issues that have probably been in the game for a long time such as: - Fixed penguins being unable to fish - Fixed bileworms not being able to devour people - Fixes goldgrubs not grubbing gold (they could not mine!) - Lizards actually eat food they find Either way, I'd reccomend a long TM on this. 1. (Hopefully) a better development experience for making AI 2. Less copy-paste for behaviors, we should be able to re-use more pieces to make behavior 3. Behavior trees is a more common pattern in making AI, so it should be easier to find resources to find out how to do things. 🆑 CabinetOnFire, Iamgoofball, SmartKar, Ben10omintrix refactor: Replaces our AI system with behavior trees, porting all datum/ai to it /🆑 I will add this PR with more details down the line. I think I got the big picture but its a big PR, so sorry if I missed something important. --------- Co-authored-by: Iamgoofball <iamgoofball@gmail.com> Co-authored-by: SmArtKar <44720187+SmArtKar@users.noreply.github.com> Co-authored-by: Ghom <42542238+Ghommie@users.noreply.github.com> Co-authored-by: Ben10Omintrix <138636438+Ben10Omintrix@users.noreply.github.com> Co-authored-by: SyncIt21 <110812394+SyncIt21@users.noreply.github.com>
Unit Tests
What is unit testing?
Unit tests are automated code to verify that parts of the game work exactly as they should. For example, a test to make sure that the amputation surgery actually amputates the limb. These are ran every time a PR is made, and thus are very helpful for preventing bugs from cropping up in your code that would've otherwise gone unnoticed. For example, would you have thought to check that beach boys would still work the same after editing pizza? If you value your time, probably not.
On their most basic level, when UNIT_TESTS is defined, all subtypes of /datum/unit_test will have their Run proc executed. From here, if Fail is called at any point, then the tests will report as failed.
How do I write one?
- Find a relevant file.
All unit test related code is in code/modules/unit_tests. If you are adding a new test for a surgery, for example, then you'd open surgeries.dm. If a relevant file does not exist, simply create one in this folder, then #include it in _unit_tests.dm.
- Create the unit test.
To make a new unit test, you simply need to define a /datum/unit_test.
For example, let's suppose that we are creating a test to make sure a proc square correctly raises inputs to the power of two. We'd start with first:
/datum/unit_test/square/Run()
This defines our new unit test, /datum/unit_test/square. Inside this function, we're then going to run through whatever we want to check. Tests provide a few assertion functions to make this easy. For now, we're going to use TEST_ASSERT_EQUAL.
/datum/unit_test/square/Run()
TEST_ASSERT_EQUAL(square(3), 9, "square(3) did not return 9")
TEST_ASSERT_EQUAL(square(4), 16, "square(4) did not return 16")
As you can hopefully tell, we're simply checking if the output of square matches the output we are expecting. If the test fails, it'll report the error message given as well as whatever the actual output was.
- Run the unit test
Open code/_compile_options.dm and uncomment the following line.
//#define UNIT_TESTS //If this is uncommented, we do a single run though of the game setup and tear down process with unit tests in between
There are 3 ways to run unit tests
-
Run tgstation.dmb in Dream Daemon. Don't bother trying to connect, you won't need to. You'll be able to see the outputs of all the tests. You'll get to see which tests failed and for what reason. If they all pass, you're set!
-
Launch game from VS Code. Launch the game as normal & you will see the output of your unit tests in your fancy chat window. This is preferred as you can use the debugger to step through each line of your unit test & can use the games inbuilt debugging tools to further aid in testing
-
Use VS Code Tgstation Test Explorer Extension. This allows you to run tests without launching the game & can also run focused tests(either a single or a selected group)
How to think about tests
Unit tests exist to prevent bugs that would happen in a real game. Thus, they should attempt to emulate the game world wherever possible. For example, the quick swap sanity test emulates a real scenario of the bug it fixed occurring by creating a character and giving it real items. The unrecommended alternative would be to create special test-only items. This isn't a hard rule, the reagent method exposure tests create a test-only reagent for example, but do keep it in mind.
Unit tests should also be just that--testing units of code. For example, instead of having one massive test for reagents, there are instead several smaller tests for testing exposure, metabolization, etc.
The unit testing API
You can find more information about all of these from their respective doc comments, but for a brief overview:
/datum/unit_test - The base for all tests to be ran. Subtypes must override Run(). New() and Destroy() can be used for setup and teardown. To fail, use TEST_FAIL(reason).
/datum/unit_test/proc/allocate(type, ...) - Allocates an instance of the provided type with the given arguments. Is automatically destroyed when the test is over. Commonly seen in the form of var/mob/living/carbon/human/human = allocate(/mob/living/carbon/human/consistent).
TEST_FAIL(reason) - Marks a failure at this location, but does not stop the test.
TEST_ASSERT(assertion, reason) - Stops the unit test and fails if the assertion is not met. For example: TEST_ASSERT(powered(), "Machine is not powered").
TEST_ASSERT_NOTNULL(a, message) - Same as TEST_ASSERT, but checks if !isnull(a). For example: TEST_ASSERT_NOTNULL(myatom, "My atom was never set!").
TEST_ASSERT_NULL(a, message) - Same as TEST_ASSERT, but checks if isnull(a). If not, gives a helpful message showing what a was. For example: TEST_ASSERT_NULL(delme, "Delme was never cleaned up!").
TEST_ASSERT_EQUAL(a, b, message) - Same as TEST_ASSERT, but checks if a == b. If not, gives a helpful message showing what both a and b were. For example: TEST_ASSERT_EQUAL(2 + 2, 4, "The universe is falling apart before our eyes!").
TEST_ASSERT_NOTEQUAL(a, b, message) - Same as TEST_ASSERT_EQUAL, but reversed.
TEST_FOCUS(test_path) - Only run the test provided within the parameters. Useful for reducing noise. For example, if we only want to run our example square test, we can add TEST_FOCUS(/datum/unit_test/square). Should never be pushed in a pull request--you will be laughed at.
TEST_REPEAT(test_path, times_to_run) - Run the provided test times_to_run times in a row. Useful for debugging flaky tests that don't always fail. Should also never be pushed in a pull request.
Final Notes
- Writing tests before you attempt to fix the bug can actually speed up development a lot! It means you don't have to go in game and folllow the same exact steps manually every time. This process is known as "TDD" (test driven development). Write the test first, make sure it fails, then start work on the fix/feature, and you'll know you're done when your tests pass. If you do try this, do make sure to confirm in a non-testing environment just to double check.
- Make sure that your tests don't accidentally call RNG functions like
prob. Since RNG is seeded during tests, you may not realize you have until someone else makes a PR and the tests fail! - Do your best not to change the behavior of non-testing code during tests. While it may sometimes be necessary in the case of situations such as the above, it is still a slippery slope that can lead to the code you're testing being too different from the production environment to be useful.