Using Zenject to wire up Mapbox SDK - unity3d

I'm using the Mapbox SDK and RoadArchitect and trying to build some game initialization code that uses Mapbox to generate a list of vectors, then take that list of vectors and call functions in Road Architect to generate road GameObjects. In order to do this, I need to wait until Mapbox has finished it's work before I can run the code to generate roads. I extended the Mapbox MergedModifierStack as ObservedModifierStack and added a StackComplete event.
My current scene setup has 2 relevant GameObjects:
Map (Mapbox AbstractMap)
which contains (pseudo propery path, might not be 100% accurate): Map.VectorTileFactory.Factories[m].Visualizer.ObservedModifierStack .Modifiers[n].MyModifier
RoadArchitect GSDRoadSystem
Most of the objects in the Map hierarchy are ScriptableObject and all are added/configured through inspectors in the editor UI.
public class MyModifier {
public MyModifier (MyDataObject myDataObject) { ... }
}
public class MyObjectCreator {
public MyObjectCreator(MyDataObject myDataObject, ModifierStack modifierStack) {
...
modifierStack.StackComplete += OnStackComplete();
}
}
In My OnStackComplete event handler I am attempting to execute functions against GSDRoadSystem. I say "attempt" because I haven't been able to figure out how to wire everything up to be injected, short of rewriting large portions of the Mapbox code simply to decorate everything with [Inject], something I'm reluctant to do given the SDK is still in heavy development.
The main issue seems to be that MyModifier isn't having MyDataObject injected, I'm guessing because Mapbox/Unity is controlling the lifecycle (due to configuration through inspectors/editor), but I don't know how to get around or work with that through Zenject.
My other option is to mess around with the Unity messaging system, and passing the vector list in the message to MyObjectCreator which would locate the appropriate GSDRoadSystem through the Unity provided service locators. Not ideal, but given this is a one-time setup and the RoadArchitect processing is going to dwarf any extra time or memory that might take I'm okay with it, even if it is a bit inelegant, although I'd rather learn how to get Zenject to wire things up, if that is at all possible (maybe it isn't because of how Mapbox built things).
I do have a fairly extensive background working with DI (Autofac, Ninject, poor-man's, etc.) but this is my first foray into Unity3d. It's possible I'm allowing Unity to cloud things and the solution is obvious, but I'm not seeing it right now.

For objects that are not created by Zenject and are not in the initial scene, you need to manually inject them yourself. You can do this by first injecting DiContainer into one of your classes and then calling DiContainer.Inject or DiContainer.InjectGameObject on the object that was created by third party code.

Related

Duplicating ScritableObjects with Values per Script Unity

I am making a Multiplayer Game with MLAPI. The problem is, i have a ability system so every ability get executed over an ScritableObject. The problem is, if i have for example five Players in the scene all Players are referenced to the same ScritableObject. And so when somebody execute a ability. The other players are also doing it.
So my Question is can I duplicate my default ability ScritableObject so that every player has its own abilities ScritableObject but with the same Values as the default ability ScritableObject. I must do that over Script.
I know i could create a new ScritableObject and then make a method who paste my values. But maybe somebody has a better idea??????????
Afaik you can use Instantiate also for cloning ScriptableObject on runtime! Only you can of course not use any component taking more then the original parameter (which makes sense).
And the same way as ScriptableObject.CreateInstance this will of course not create an asset but only a runtime instance.

scala for mapbox vector tiles - getting an 'id' field into the Features written to vector tiles

I'm writing MapBox vector tiles using geotrellis vectorpipe.
see here for the basic flow: https://geotrellis.github.io/vectorpipe/usage.html
Typically GeoJson Features can have an id field, so that Features can be rolled up into FeatureCollections. I need to make use of this field, but vectorpipe doesn't (natively) have this capability.
This is the Feature type used, and you can see it only has space for 1) a Geometry and 2) a data object D (which ends up populating properties in the output). There is no spot for an id.
https://geotrellis.github.io/scaladocs/latest/index.html#geotrellis.vector.Feature
Upstream there's a method called writeFeatureJsonWithID() that does let you inject an id field into a Feature when writing GeoJson.
https://github.com/locationtech/geotrellis/blob/master/vector/src/main/scala/geotrellis/vector/io/json/FeatureFormats.scala#L41-L49
My question is this:
I have worked through the vectorpipe code (https://github.com/geotrellis/vectorpipe), and I can't figure out if/where the data ever exists as GeoJson in a way where I can override and inject the id, maybe using the writeFeatureJsonWithID() or something I write explicitly. A lot of conversions are implicit, but it also may never explicitly sit as json.
Any ideas for how to get an id field in the final GeoJson written to vector tiles?
EDIT
Right now I think the trick is going to be finding a way to override .unfeature() method here:
https://github.com/locationtech/geotrellis/blob/master/vectortile/src/main/scala/geotrellis/vectortile/Layer.scala
The problem is that the internal.vector_tile.Tile is private, so I can construct it without forking the project.
Ended up having to fork geotrellis, hard-code a metadata => id function in Layer.unfeature() and compile locally to include in my project. Not ideal, but it works fine.
Also opened an issue here: https://github.com/locationtech/geotrellis/issues/2884

qt.webChannelTransport from sub-frame

Is it possible to use the qt.webChannelTransport from a sub-frame? It seems like QWebEnginePage.setWebChannel only creates the qt.webChannelTransport object in the top frame. I've considered trying to inject scripts to handle cross document messages and then use the Qt WebChannel between the main frame and the qt application, but that seems like a total pain.

how to run the example of uima-text-segmenter?

I want to call the API of uima-text-segmenter https://code.google.com/p/uima-text-segmenter/source/browse/trunk/INSTALL?r=22 to run an example.
But I don`t know how to call the API...
the readme said,
With the DocumentAnalyzer, run the following descriptor
`desc/textSegmenter/wst-snowball-C99-JTextTilingAAE.xml` by taking the
uima-examples data as input.
Could anyone give me some code which could be run directly in main func for example?
Thanks a lot!
Long answer:
The link describes how you would set up the application from within the Eclipse UIMA environment. This sort of set-up is typically targeted at subject matter specialists with little or no coding experience. It allows them to work (relatively fast) with UIMA in a declarative way: all data structures and analysis engines (computing blocks within UIMA) are declared in xml (with a GUI on top of it), after which the framework takes care of the rest. In this scenario, you would typically run a UIMA pipeline using a run configuration from within Eclipse (or the included UIMA pipeline runner application). Luckily, UIMA allows you to do exactly the same from code, but I would recommend using UIMAFit (http://uima.apache.org/d/uimafit-current/tools.uimafit.book.html#d5e137) for this purpose instead of of UIMA, as it bundles lots of useful things and coding shortcuts.
Short answer:
Using UIMAFit, you can call Factory methods that create CollectionReader (read input), AnalysisEngine (process input) and Consumer objects (write/do other stuff) from (third-party provided) XML files. Use these methods to construct your pipeline and the SimplePipeline class to run it. To extract the data you need, you would manipulate the CAS object (containing your data) in a Consumer object, possibly with a callback. You could also do this in a Analysis Engine object. I recommend using DKPro's FeaturePathFactory (https://code.google.com/p/dkpro-core-asl/source/browse/de.tudarmstadt.ukp.dkpro.core-asl/trunk/de.tudarmstadt.ukp.dkpro.core.api.featurepath-asl/src/main/java/de/tudarmstadt/ukp/dkpro/core/api/featurepath/FeaturePathFactory.java?spec=svn1811&r=1811) to quickly access the feature you are after.
Code examples:
http://uima.apache.org/d/uimafit-current/tools.uimafit.book.html#d5e137 contains examples, but they all go in the opposite direction (class objects are used in the factory methods, instead of XML files - XML is generated from these classes). Take a look at the UIMAFit API to find the method you need, AnalysisEngineDescription from XML for example: http://uima.apache.org/d/uimafit-current/api/org/apache/uima/fit/factory/AnalysisEngineFactory.html#createEngineDescriptionFromPath-java.lang.String-java.lang.Object...-

Game: General Structure and Game Item Structure

I am making a graphically simple 3D game in C++ using DirectX. The main problem I am having is with how to structure some things efficiently. Right now I know my goal for certain areas but not how to ideally perform them.
For instance, right now I am storing all meshes and textures in an Asset class with enumerated definitions pointing to each asset. All meshes and textures are loaded when the game starts by creating an Asset object and initializing it. From there I load meshes and textures to objects by assigning the pointer given by the Asset object. Is this the best way to go about this?
A harder topic is that of items. For the sake of argument I am going to say we are dealing with scenery. Right now my scenery needs the following:
A name
A mesh
Textures (optional)
Flags (such as Destructible, Flammable, etc.)
Status (such as Burning, Locked, Cursed, etc.)
Abilities (things the object actually does, such as becoming damaged when burning)
The first 5 things on this list are simply variables. But abilities are bothering me. The first 5 can all exist in one item class. But what I do not know how to do is somehow attach abilities to the object.
For example; I want the object to have the "Fire Nova" ability. At the start of each turn, the object damages anything near it.
For me, this means each object would need a "Trigger_Phase_TurnStart()" or some similarly named method and would then need a new "Fire Nova" class with its own unique action for that trigger and its own extra variables (Damage, Range, etc). So now I have the base object class and a new Fire Nova class that inherits from it.
But now what if I needed an object that has Fire Nova and Frost Nova and also Slowing Aura and other such abilities. Basically I need a way to add triggered effects to an object without needing a new object class for each of them.
So by the end I would be able to have say:
(pseudo code of the object's components)
name = Elemental Orb
mesh = "Mesh_Sphere"
textures[] = "Tex_Ice", "Tex_Fire"
flags = OF_Destructible
status = SF_Cursed | SF_Burning
abilities[] = Fire_Nova, Frost_Nova, Slowing_Aura
And this orb would simply be the object class with all these attributes. The object would then activate each stored ability's trigger at the appropriate turn phase for any actions to perform.
Right now I am thinking I might need to make a class for each ability possessing every turn-phase or action trigger (inherited from a base ability class) and have them perform the appropriate action when the object calls them from it's array of abilities. Would this be the best way to do this?
As a separate issue. Some flags would require additional variables that would otherwise be unnecessary. For example, Destructible would mean the object would have health whereas without the flag it wouldn't need it, or an Openable item would need an array of contents. What would be a good way to ration these variables to when they are needed? I do not need every wall to have health and an empty contents array for example.
Finally. Assuming the bullet-listed attributes above are all an item needs, how might you suggest I best store them? To clarify, I would like to be able to write:
CreateItem(ITEM_CHAIR);
and have this return the created object with name, mesh, textures, flags, status and abilities. What structure might be suitable for achieving such an end effect?
To summarise:
Does my current Asset storage seem feasible?
What is the best way to create abilities to attach to the object class without making numerous separate object classes?
Is there a way to limit the variables (or at least memory usage) when the associated flag is not present?
What structure or format would be best for storing fixed item definitions?
Sorry if this is a little long winded. If you cannot answer all the questions then an answer to one would still be appreciated. Question two is probably the largest priority for me right now.
Thanks for your time :)
I want only answer question 2 and it looks like you need a decorator pattern. In OOP a decorator pattern is useful when you have many ingredients and wants to decorate an object, for example a pizza. Unfortunately you need to create for each ingredient a separate class hence I think you have the right approach. The good thing is that with the decorator pattern you can wrap the object over and over with abilites classes and call only one method at the end to do all the stuff.