using System;
using System.Collections.Generic;
using Unity.MLAgents.Actuators;
using Unity.MLAgents.Sensors;
namespace Unity.MLAgents
{
internal struct CommunicatorInitParameters
{
///
/// Port to listen for connections on.
///
public int port;
///
/// The name of the environment.
///
public string name;
///
/// The version of the Unity SDK.
///
public string unityPackageVersion;
///
/// The version of the communication API.
///
public string unityCommunicationVersion;
///
/// The RL capabilities of the C# codebase.
///
public UnityRLCapabilities CSharpCapabilities;
}
internal struct UnityRLInitParameters
{
///
/// A random number generator (RNG) seed sent from the python process to Unity.
///
public int seed;
///
/// The library version of the python process.
///
public string pythonLibraryVersion;
///
/// The version of the communication API that python is using.
///
public string pythonCommunicationVersion;
///
/// The RL capabilities of the Trainer codebase.
///
public UnityRLCapabilities TrainerCapabilities;
}
internal struct UnityRLInputParameters
{
///
/// Boolean sent back from python to indicate whether or not training is happening.
///
public bool isTraining;
}
///
/// Delegate for handling quit events sent back from the communicator.
///
internal delegate void QuitCommandHandler();
///
/// Delegate for handling reset parameter updates sent from the communicator.
///
internal delegate void ResetCommandHandler();
///
/// Delegate to handle UnityRLInputParameters updates from the communicator.
///
///
internal delegate void RLInputReceivedHandler(UnityRLInputParameters inputParams);
/**
This is the interface of the Communicators.
This does not need to be modified nor implemented to create a Unity environment.
When the Unity Communicator is initialized, it will wait for the External Communicator
to be initialized as well. The two communicators will then exchange their first messages
that will usually contain information for initialization (information that does not need
to be resent at each new exchange).
By convention a Unity input is from External to Unity and a Unity output is from Unity to
External. Inputs and outputs are relative to Unity.
By convention, when the Unity Communicator and External Communicator call exchange, the
exchange is NOT simultaneous but sequential. This means that when a side of the
communication calls exchange, the other will receive the result of its previous
exchange call.
This is what happens when A calls exchange a single time:
A sends data_1 to B -> B receives data_1 -> B generates and sends data_2 -> A receives data_2
When A calls exchange, it sends data_1 and receives data_2
Since the messages are sent back and forth with exchange and simultaneously when calling
initialize, External sends two messages at initialization.
The structure of the messages is as follows:
UnityMessage
...Header
...UnityOutput
......UnityRLOutput
......UnityRLInitializationOutput
...UnityInput
......UnityRLInput
......UnityRLInitializationInput
UnityOutput and UnityInput can be extended to provide functionalities beyond RL
UnityRLOutput and UnityRLInput can be extended to provide new RL functionalities
*/
internal interface ICommunicator : IDisposable
{
///
/// Quit was received by the communicator.
///
event QuitCommandHandler QuitCommandReceived;
///
/// Reset command sent back from the communicator.
///
event ResetCommandHandler ResetCommandReceived;
///
/// Sends the academy parameters through the Communicator.
/// Is used by the academy to send the AcademyParameters to the communicator.
///
/// The External Initialization Parameters received.
/// The Unity Initialization Parameters to be sent.
UnityRLInitParameters Initialize(CommunicatorInitParameters initParameters);
///
/// Registers a new Brain to the Communicator.
///
/// The name or key uniquely identifying the Brain.
/// Description of the actions for the Agent.
void SubscribeBrain(string name, ActionSpec actionSpec);
///
/// Sends the observations of one Agent.
///
/// Batch Key.
/// Agent info.
/// The list of ISensors of the Agent.
void PutObservations(string brainKey, AgentInfo info, List sensors);
///
/// Signals the ICommunicator that the Agents are now ready to receive their action
/// and that if the communicator has not yet received an action for one of the Agents
/// it needs to get one at this point.
///
void DecideBatch();
///
/// Gets the AgentActions based on the batching key.
///
/// A key to identify which behavior actions to get.
/// A key to identify which Agent actions to get.
///
ActionBuffers GetActions(string key, int agentId);
}
}