Add multiAgentGroup capabilities flag (#5096)

* Add multiAgentGroup capabilities flag

* Add proto

* Fix compiler error

* Add warning for multiagent group

* Add comment

* Fix spelling mistake

Author: Ervin T

com.unity.ml-agents/Runtime/Academy.cs

@@ -97,7 +97,7 @@ public class Academy : IDisposable
         ///     </item>
         ///     <item>
         ///         <term>1.5.0</term>
-        ///         <description>Support variable length observation training.</description>
+        ///         <description>Support variable length observation training and multi-agent groups.</description>
         ///     </item>
         /// </list>
         /// </remarks>

com.unity.ml-agents/Runtime/Communicator/GrpcExtensions.cs

@@ -20,6 +20,11 @@ namespace Unity.MLAgents
     internal static class GrpcExtensions
     {
         #region AgentInfo
+        /// <summary>
+        /// Static flag to make sure that we only fire the warning once.
+        /// </summary>
+        private static bool s_HaveWarnedTrainerCapabilitiesAgentGroup = false;
+
         /// <summary>
         /// Converts a AgentInfo to a protobuf generated AgentInfoActionPairProto
         /// </summary>
@@ -55,6 +60,22 @@ public static AgentInfoActionPairProto ToInfoActionPairProto(this AgentInfo ai)
         /// <returns>The protobuf version of the AgentInfo.</returns>
         public static AgentInfoProto ToAgentInfoProto(this AgentInfo ai)
         {
+            if(ai.groupId > 0)
+            {
+                var trainerCanHandle = Academy.Instance.TrainerCapabilities == null || Academy.Instance.TrainerCapabilities.MultiAgentGroups;
+                if (!trainerCanHandle)
+                {
+                    if (!s_HaveWarnedTrainerCapabilitiesAgentGroup)
+                    {
+                        Debug.LogWarning(
+                            $"Attached trainer doesn't support Multi Agent Groups; group rewards will be ignored." +
+                            "Please find the versions that work best together from our release page: " +
+                            "https://github.com/Unity-Technologies/ml-agents/releases"
+                        );
+                        s_HaveWarnedTrainerCapabilitiesAgentGroup = true;
+                    }
+                }
+            }
             var agentInfoProto = new AgentInfoProto
             {
                 Reward = ai.reward,
@@ -457,6 +478,7 @@ public static UnityRLCapabilities ToRLCapabilities(this UnityRLCapabilitiesProto
                 HybridActions = proto.HybridActions,
                 TrainingAnalytics = proto.TrainingAnalytics,
                 VariableLengthObservation = proto.VariableLengthObservation,
+                MultiAgentGroups = proto.MultiAgentGroups,
             };
         }
 
@@ -470,6 +492,7 @@ public static UnityRLCapabilitiesProto ToProto(this UnityRLCapabilities rlCaps)
                 HybridActions = rlCaps.HybridActions,
                 TrainingAnalytics = rlCaps.TrainingAnalytics,
                 VariableLengthObservation = rlCaps.VariableLengthObservation,
+                MultiAgentGroups = rlCaps.MultiAgentGroups,
             };
         }
 

com.unity.ml-agents/Runtime/Communicator/UnityRLCapabilities.cs

@@ -10,6 +10,7 @@ internal class UnityRLCapabilities
         public bool HybridActions;
         public bool TrainingAnalytics;
         public bool VariableLengthObservation;
+        public bool MultiAgentGroups;
 
         /// <summary>
         /// A class holding the capabilities flags for Reinforcement Learning across C# and the Trainer codebase.  This
@@ -21,14 +22,16 @@ public UnityRLCapabilities(
             bool compressedChannelMapping = true,
             bool hybridActions = true,
             bool trainingAnalytics = true,
-            bool variableLengthObservation = true)
+            bool variableLengthObservation = true,
+            bool multiAgentGroups = true)
         {
             BaseRLCapabilities = baseRlCapabilities;
             ConcatenatedPngObservations = concatenatedPngObservations;
             CompressedChannelMapping = compressedChannelMapping;
             HybridActions = hybridActions;
             TrainingAnalytics = trainingAnalytics;
             VariableLengthObservation = variableLengthObservation;
+            MultiAgentGroups = multiAgentGroups;
         }
 
         /// <summary>

com.unity.ml-agents/Runtime/Grpc/CommunicatorObjects/Capabilities.cs

@@ -25,17 +25,18 @@ static CapabilitiesReflection() {
       byte[] descriptorData = global::System.Convert.FromBase64String(
           string.Concat(
             "CjVtbGFnZW50c19lbnZzL2NvbW11bmljYXRvcl9vYmplY3RzL2NhcGFiaWxp",
-            "dGllcy5wcm90bxIUY29tbXVuaWNhdG9yX29iamVjdHMi0gEKGFVuaXR5UkxD",
+            "dGllcy5wcm90bxIUY29tbXVuaWNhdG9yX29iamVjdHMi7AEKGFVuaXR5UkxD",
             "YXBhYmlsaXRpZXNQcm90bxIaChJiYXNlUkxDYXBhYmlsaXRpZXMYASABKAgS",
             "IwobY29uY2F0ZW5hdGVkUG5nT2JzZXJ2YXRpb25zGAIgASgIEiAKGGNvbXBy",
             "ZXNzZWRDaGFubmVsTWFwcGluZxgDIAEoCBIVCg1oeWJyaWRBY3Rpb25zGAQg",
             "ASgIEhkKEXRyYWluaW5nQW5hbHl0aWNzGAUgASgIEiEKGXZhcmlhYmxlTGVu",
-            "Z3RoT2JzZXJ2YXRpb24YBiABKAhCJaoCIlVuaXR5Lk1MQWdlbnRzLkNvbW11",
-            "bmljYXRvck9iamVjdHNiBnByb3RvMw=="));
+            "Z3RoT2JzZXJ2YXRpb24YBiABKAgSGAoQbXVsdGlBZ2VudEdyb3VwcxgHIAEo",
+            "CEIlqgIiVW5pdHkuTUxBZ2VudHMuQ29tbXVuaWNhdG9yT2JqZWN0c2IGcHJv",
+            "dG8z"));
       descriptor = pbr::FileDescriptor.FromGeneratedCode(descriptorData,
           new pbr::FileDescriptor[] { },
           new pbr::GeneratedClrTypeInfo(null, new pbr::GeneratedClrTypeInfo[] {
-            new pbr::GeneratedClrTypeInfo(typeof(global::Unity.MLAgents.CommunicatorObjects.UnityRLCapabilitiesProto), global::Unity.MLAgents.CommunicatorObjects.UnityRLCapabilitiesProto.Parser, new[]{ "BaseRLCapabilities", "ConcatenatedPngObservations", "CompressedChannelMapping", "HybridActions", "TrainingAnalytics", "VariableLengthObservation" }, null, null, null)
+            new pbr::GeneratedClrTypeInfo(typeof(global::Unity.MLAgents.CommunicatorObjects.UnityRLCapabilitiesProto), global::Unity.MLAgents.CommunicatorObjects.UnityRLCapabilitiesProto.Parser, new[]{ "BaseRLCapabilities", "ConcatenatedPngObservations", "CompressedChannelMapping", "HybridActions", "TrainingAnalytics", "VariableLengthObservation", "MultiAgentGroups" }, null, null, null)
           }));
     }
     #endregion
@@ -78,6 +79,7 @@ public UnityRLCapabilitiesProto(UnityRLCapabilitiesProto other) : this() {
       hybridActions_ = other.hybridActions_;
       trainingAnalytics_ = other.trainingAnalytics_;
       variableLengthObservation_ = other.variableLengthObservation_;
+      multiAgentGroups_ = other.multiAgentGroups_;
       _unknownFields = pb::UnknownFieldSet.Clone(other._unknownFields);
     }
 
@@ -170,6 +172,20 @@ public bool VariableLengthObservation {
       }
     }
 
+    /// <summary>Field number for the "multiAgentGroups" field.</summary>
+    public const int MultiAgentGroupsFieldNumber = 7;
+    private bool multiAgentGroups_;
+    /// <summary>
+    /// Support for multi agent groups and group rewards
+    /// </summary>
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public bool MultiAgentGroups {
+      get { return multiAgentGroups_; }
+      set {
+        multiAgentGroups_ = value;
+      }
+    }
+
     [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
     public override bool Equals(object other) {
       return Equals(other as UnityRLCapabilitiesProto);
@@ -189,6 +205,7 @@ public bool Equals(UnityRLCapabilitiesProto other) {
       if (HybridActions != other.HybridActions) return false;
       if (TrainingAnalytics != other.TrainingAnalytics) return false;
       if (VariableLengthObservation != other.VariableLengthObservation) return false;
+      if (MultiAgentGroups != other.MultiAgentGroups) return false;
       return Equals(_unknownFields, other._unknownFields);
     }
 
@@ -201,6 +218,7 @@ public override int GetHashCode() {
       if (HybridActions != false) hash ^= HybridActions.GetHashCode();
       if (TrainingAnalytics != false) hash ^= TrainingAnalytics.GetHashCode();
       if (VariableLengthObservation != false) hash ^= VariableLengthObservation.GetHashCode();
+      if (MultiAgentGroups != false) hash ^= MultiAgentGroups.GetHashCode();
       if (_unknownFields != null) {
         hash ^= _unknownFields.GetHashCode();
       }
@@ -238,6 +256,10 @@ public void WriteTo(pb::CodedOutputStream output) {
         output.WriteRawTag(48);
         output.WriteBool(VariableLengthObservation);
       }
+      if (MultiAgentGroups != false) {
+        output.WriteRawTag(56);
+        output.WriteBool(MultiAgentGroups);
+      }
       if (_unknownFields != null) {
         _unknownFields.WriteTo(output);
       }
@@ -264,6 +286,9 @@ public int CalculateSize() {
       if (VariableLengthObservation != false) {
         size += 1 + 1;
       }
+      if (MultiAgentGroups != false) {
+        size += 1 + 1;
+      }
       if (_unknownFields != null) {
         size += _unknownFields.CalculateSize();
       }
@@ -293,6 +318,9 @@ public void MergeFrom(UnityRLCapabilitiesProto other) {
       if (other.VariableLengthObservation != false) {
         VariableLengthObservation = other.VariableLengthObservation;
       }
+      if (other.MultiAgentGroups != false) {
+        MultiAgentGroups = other.MultiAgentGroups;
+      }
       _unknownFields = pb::UnknownFieldSet.MergeFrom(_unknownFields, other._unknownFields);
     }
 
@@ -328,6 +356,10 @@ public void MergeFrom(pb::CodedInputStream input) {
             VariableLengthObservation = input.ReadBool();
             break;
           }
+          case 56: {
+            MultiAgentGroups = input.ReadBool();
+            break;
+          }
         }
       }
     }

ml-agents-envs/mlagents_envs/communicator_objects/capabilities_pb2.py

@@ -31,6 +31,7 @@ class UnityRLCapabilitiesProto(google___protobuf___message___Message):
     hybridActions = ... # type: builtin___bool
     trainingAnalytics = ... # type: builtin___bool
     variableLengthObservation = ... # type: builtin___bool
+    multiAgentGroups = ... # type: builtin___bool
 
     def __init__(self,
         *,
@@ -40,12 +41,13 @@ class UnityRLCapabilitiesProto(google___protobuf___message___Message):
         hybridActions : typing___Optional[builtin___bool] = None,
         trainingAnalytics : typing___Optional[builtin___bool] = None,
         variableLengthObservation : typing___Optional[builtin___bool] = None,
+        multiAgentGroups : typing___Optional[builtin___bool] = None,
         ) -> None: ...
     @classmethod
     def FromString(cls, s: builtin___bytes) -> UnityRLCapabilitiesProto: ...
     def MergeFrom(self, other_msg: google___protobuf___message___Message) -> None: ...
     def CopyFrom(self, other_msg: google___protobuf___message___Message) -> None: ...
     if sys.version_info >= (3,):
-        def ClearField(self, field_name: typing_extensions___Literal[u"baseRLCapabilities",u"compressedChannelMapping",u"concatenatedPngObservations",u"hybridActions",u"trainingAnalytics",u"variableLengthObservation"]) -> None: ...
+        def ClearField(self, field_name: typing_extensions___Literal[u"baseRLCapabilities",u"compressedChannelMapping",u"concatenatedPngObservations",u"hybridActions",u"multiAgentGroups",u"trainingAnalytics",u"variableLengthObservation"]) -> None: ...
     else:
-        def ClearField(self, field_name: typing_extensions___Literal[u"baseRLCapabilities",b"baseRLCapabilities",u"compressedChannelMapping",b"compressedChannelMapping",u"concatenatedPngObservations",b"concatenatedPngObservations",u"hybridActions",b"hybridActions",u"trainingAnalytics",b"trainingAnalytics",u"variableLengthObservation",b"variableLengthObservation"]) -> None: ...
+        def ClearField(self, field_name: typing_extensions___Literal[u"baseRLCapabilities",b"baseRLCapabilities",u"compressedChannelMapping",b"compressedChannelMapping",u"concatenatedPngObservations",b"concatenatedPngObservations",u"hybridActions",b"hybridActions",u"multiAgentGroups",b"multiAgentGroups",u"trainingAnalytics",b"trainingAnalytics",u"variableLengthObservation",b"variableLengthObservation"]) -> None: ...

ml-agents-envs/mlagents_envs/communicator_objects/capabilities_pb2.pyi

@@ -63,7 +63,7 @@ class UnityEnvironment(BaseEnv):
     #  * 1.2.0 - support compression mapping for stacked compressed observations.
     #  * 1.3.0 - support action spaces with both continuous and discrete actions.
     #  * 1.4.0 - support training analytics sent from python trainer to the editor.
-    #  * 1.5.0 - support variable length observation training.
+    #  * 1.5.0 - support variable length observation training and multi-agent groups.
     API_VERSION = "1.5.0"
 
     # Default port that the editor listens on. If an environment executable
@@ -124,6 +124,7 @@ def _get_capabilities_proto() -> UnityRLCapabilitiesProto:
         capabilities.hybridActions = True
         capabilities.trainingAnalytics = True
         capabilities.variableLengthObservation = True
+        capabilities.multiAgentGroups = True
         return capabilities
 
     @staticmethod

ml-agents-envs/mlagents_envs/environment.py

@@ -25,4 +25,7 @@ message UnityRLCapabilitiesProto {
 
     // Support for variable length observations of rank 2
     bool variableLengthObservation = 6;
+
+    // Support for multi agent groups and group rewards
+    bool multiAgentGroups = 7;
 }