class based api

Content/Scripts/TFPluginAPI.py

@@ -0,0 +1,61 @@
+import sys
+
+import tensorflow as tf
+import unreal_engine as ue
+
+class TFPluginAPI():
+	@classmethod
+	def getInstance(cls):
+		#This should return an instance of your class even if you subclassed it
+		return cls()
+
+	def __init__(self):
+		#class scoped variable for stopping
+		self.shouldstop = False
+		self.stored = {}
+
+	#expected api: setup your model for training
+	def setup(self):
+		#setup or load your model and pass it into stored
+
+		#Usually store session, graph, and model if using keras
+		#self.sess = tf.InteractiveSession()
+		#self.graph = tf.get_default_graph()
+		pass
+
+	#expected api: storedModel and session, json inputs
+	def runJsonInput(self, jsonInput):
+		#e.g. our json input could be a pixel array
+		#pixelarray = jsonInput['pixels']
+
+		#run input on your graph
+		#e.g. sess.run(model['y'], feed_dict)
+		# where y is your result graph and feed_dict is {x:[input]}
+
+		#...
+
+		#return a json you will parse e.g. a prediction
+		result = {}
+		result['prediction'] = 0
+
+		return result
+
+	#expected api: early stopping
+	def stop(self):
+		self.shouldstop = True
+
+	#expected api: no params forwarded for training? TBC
+	def train(self):
+		#train here
+
+		#...
+
+		#inside your training loop check if we should stop early
+		#if(this.shouldstop):
+		#	break
+		pass
+
+#required function to get our api
+def getApi():
+	#return CLASSNAME.getInstance()
+	return TFPluginAPI.getInstance()

Content/Scripts/TensorFlowComponent.py

@@ -18,8 +18,15 @@ def begin_play(self):
 		if(self.uobject.VerbosePythonLog):
 			ue.log('BeginPlay, importing TF module: ' + self.uobject.TensorFlowModule)
 
-		self.tf = importlib.import_module(self.uobject.TensorFlowModule)
-		imp.reload(self.tf)
+		#import the module
+		self.tfModule = importlib.import_module(self.uobject.TensorFlowModule)
+		imp.reload(self.tfModule)
+
+		#tfc or the class instance holding the pluginAPI
+		self.tfapi = self.tfModule.getApi()
+
+		#init
+		self.tfapi.setup()
 
 		#train
 		if(self.uobject.ShouldTrainOnBeginPlay):
@@ -30,15 +37,15 @@ def begin_play(self):
 
 	def end_play(self):
 		self.ValidGameWorld = False
-		self.tf.stop()
+		self.tfapi.stop()
 
 	#tensor input
 	def tensorinput(self, args):
 		if(self.uobject.VerbosePythonLog):
 			ue.log(self.uobject.TensorFlowModule + ' input passed: ' + args)
 
 		#pass the raw json to the script to handle
-		resultJson = self.tf.runJsonInput(self.trained, json.loads(args))
+		resultJson = self.tfapi.runJsonInput(json.loads(args))
 
 		#pass prediction json back
 		self.uobject.OnResultsFunction(json.dumps(resultJson))
@@ -57,7 +64,7 @@ def trainBlocking(self):
 
 		#calculate the time it takes to train your network
 		start = time.time()
-		self.trained = self.tf.train()
+		self.trained = self.tfapi.train()
 		stop = time.time()
 
 		if self.trained is None: