Skip to content

HeadlessServer

Baptiste Lesquoy edited this page Jan 25, 2024 · 29 revisions

Headless Server

Running a Gama Headless server

Before doing anything, make sure that you possess the rights to create files and directories at the location you are running gama-server because it will need it to create workspaces on the fly.

From the release

Go to the headless directory in your Gama installation folder and run the script gama-headless.sh (or gama-headless.bat) with the argument -socket followed by the port number you want your Gama server to run on.

For example on Mac OS you could do:

cd Gama.app/Contents/headless

to move to the right directory, then run the script to listen on port 6868 with:

gama-headless.sh -socket 6868

From the command-line tool

The users who installed gama through a .deb file or aur have access to the command gama-headless and thus only need to open a terminal and run

gama-headless -socket 6868

to run a Gama server on the port 6868.

From the source code

In Eclipse, instantiate a headless server by running msi.gama.headless.id4_full with the following argument -os ${target.os} -ws ${target.ws} -arch ${target.arch} -nl ${target.nl} -socket 6868 (you can specify any other port)

From docker

First ensure to pull the official docker image

docker pull gamaplatform/gama:<version>

Then, run the container with the image you just pulled.

Do not forget to (1) expose the port you're starting your server on, and (2) mount your workspace inside the started container as below :

docker run -v <path/to/your/workspace>:/working_dir -p 6868:6868 gamaplatform/gama:<version> -socket 6868

For more informations, please refer to Docker's official documentation or GAMA image's repository.

Interacting with Gama Server

Connection

To connect to gama-server as a client you just need to access the following address: ws://<ip>:<port>. For example if you try to connect to a gama-server running on your current computer and started with the command gama-headless -socket 6868, you will have to connect to ws://localhost:6868.

Once a client is connected, gama-server will send a json object of type ConnectionSuccessful.

General description of interactions

Once connected, you can ask gama-server to execute different commands to control the execution of different simulations.

If you close your client application (or just close the socket on client-side) gama-server will destroy all running simulations of that client, so you have to keep your client alive.

For every command treated by gama-server, it will send back a json object describing if the command has been executed correctly or if there was a problem. If an unexpected exception is raised in gama-server, it will try to send the connected clients a json-object describing it. The same goes if a simulation throws an exception/error while running, the client that asked for it to run will receive it as a json-object.

In addition, the client can ask gama-server to receive (or not) the different outputs of a simulation: write statements, dialogs, status-bar changes etc. they will be sent as they come, in a specific json wrapper.

Available commands

All the commands sent to gama-server must be formatted as a json object.

The available commands are:

The exit command

This command is used to kill gama-server. It is triggered by sending a json object formatted as follows to the server

{
  "type": "exit"
}

Answer from gama-server

It is the only command that won't send back a json object.

The load command

This command is used to ask the server to initialize a specific experiment in a gaml file on the server's file-system. It is triggered by sending a json object formatted as follows to the server:

{
  "type": "load",
  "model": "<gaml_file_path>",
  "experiment": "<experiment_name>",
  "console": "<console>", //optional
  "status": "<status>", //optional
  "dialog": "<dialog>", //optional
  "runtime": "<runtime>",//optional
  "parameters": "<params>", //optional
  "until": "<end_condition>", //optional
}

The model parameter indicates the path of the experiment file on the server's file-system, and experiment is the actual name of the experiment to run. The four parameters console, status, dialog and runtime are booleans used to determine if the messages from respectively the console, the status-bar, the dialogs and the runtime errors should be redirected to the client. They are optional as per default console and runtime are set to true and the two others to false. You can add an array of parameters that will be used to initialize the experiment's variables with the values you picked. The value of parameters should be formatted as follows:

[
  {
    "type": "<type of the first parameter>",
    "value": "<value of the first parameter>",
    "name": "<name of the first parameter in the gaml file>"
  },
  {
    "type": "<type of the second parameter>",
    "value": "<value of the second parameter>",
    "name": "<name of the second parameter in the gaml file>"
  },
 ...
]

You can also add an ending condition to your simulation with the parameter until, the condition must be expressed in gaml.

Answer from gama-server

The content field of the response json sent by gama-server after processing this command will directly contain the experiment_id value stored as a string. The experiment id should be used in all the other commands to refer to that specific experiment in order to control it.

The play command

This command is used to actually run an experimented already initialized. It is triggered by sending a json object formatted as follows to the server

{
  "type": "play",
  "exp_id": "<experiment_id>",
  "sync": "<synchronized>", //optional
}

The experiment_id is used to identify the experiment to play, and the optional sync is a boolean used in the case where there was an end condition defined in the load command, if it is true, gama-server will not send a response to the command, but only a end of simulation message once the condition is reached, if it's false gama-server will send both the response to the command and the SimulationEnded message.

Answer from gama-server

This command has an empty content field in the response json sent by gama-server after processing it. In case where the end condition is reached, a message of type SimulationEnded is sent to the client with an empty content.

The pause command

This command is used to pause a running experiment. It is triggered by sending a json object formatted as follows to the server

{
  "type": "pause",
  "exp_id": "<experiment_id>"
}

Answer from gama-server

This command has an empty content field in the response json sent by gama-server after processing it.

The step command

This command is used to process one (or a defined number of) step(s) of a simulation that has already been loaded. It is triggered by sending a json object formatted as follows to the server

{
  "type": "step",
  "exp_id": "<experiment_id>",
  "nb_step": "<number_of_steps>", //optional
  "sync": "<synchronized>", // optional
}

As usual exp_id refers to the experiment you want to apply the command to. The nb_step parameter indicates how many steps you want to execute, if you do not give that parameter gama-server will execute one step. The sync parameter indicates whether gama-server must wait for the end of the step(s) to send back a success message (when its value is true), or just plan the step(s) and send one directly after (when its value is false), this parameter can be ignored and will be interpreted as if it were false.

Answer from gama-server

This command has an empty content field in the response json sent by gama-server after processing it.

The stepBack command

This command is used to rollback the simulation one (or a defined number of) step(s) back. This command only works on experiments of type memorize. It is triggered by sending a json object formatted as follows to the server

{
  "type": "stepBack",
  "exp_id": "<experiment_id>",
  "nb_step": "<number_of_steps>", //optional
  "sync": "<synchronized>", // optional
}

The parameters are exactly the same as in the step command.

Answer from gama-server

This command has an empty content field in the response json sent by gama-server after processing it.

The stop command

This command is used to stop (kill) a running experiment. It is triggered by sending a json object formatted as follows to the server

{
  "type": "stop",
  "exp_id": "<experiment_id>",
}

Answer from gama-server

This command has an empty content field in the response json sent by gama-server after processing it.

The reload command

This command is used to reload an experiment. The experiment will be stop and the initialization process run again. You can use this command to change the simulation parameters or the ending condition. It is triggered by sending a json object formatted as follows to the server

{
  "type": "reload",
  "exp_id": "<experiment_id>",
  "parameters": "<params>", //optional
  "until": "<end_condition>", //optional
}

Just like for the load command, the parameters and the until parameters are optional and must follow the same formatting.

Answer from gama-server

This command has an empty content field in the response json sent by gama-server after processing it.

The expression command

This command is used to ask the server to evaluate a gaml expression having an experiment as context.
It is triggered by sending a json object formatted as follows to the server

{
  "type": "expression",
  "exp_id": "<experiment_id>",
  "expr": "<expression to evaluate>"
}

For example if you want to know the number of agents of species people currently present in the simulation represented by the id 123, you could send this command to gama-server:

{
  "type": "expression",
  "exp_id": "123",
  "expr": "length(people)"
}

Answer from gama-server

If the command is executed successfully by gama-server the content field of the response json will directly contain the result of the evaluated expression as a string.

The ask command

This command is used to execute an action defined in an agent inside of an experiment.
It is triggered by sending a json object formatted as follows to the server:

{
  "type": "ask",
  "exp_id": "<experiment_id>", // only required in headless
  "socket_id": "<socket_id>", // only required in headless
  "action": "<expression to evaluate>",
  "args": "<the arguments of the action>",
  "agent": "<the agent to apply the action to>",
  "escaped": "<the expression is escaped or not>", //optional
}

Answer from gama-server

If the action has been executed without any problem, gama-server will return a CommandExecutedSuccessfully message, else if errors occur it will be a UnableToExecuteRequest type of message.

The validate command

This command is used to ask the server to validate (compile and check for errors) a set of gaml expressions.
It is triggered by sending a json object formatted as follows to the server

{
  "type": "validate",
  "expr": "<expressions to validate>",
  "syntax": "<check the syntax only>", // optional
  "escaped": "<the expressions are escaped or not>", //optional
}

The syntax parameter is a boolean and can be used to ask for a syntactical check only (true) or syntactical + semantic check (false) if not provided it will be considered false.

Answer from gama-server

If no error has been found, the server will return a CommandExecutedSuccessfully message, else it will be a UnableToExecuteRequest where the content will be the list of errors.

The download command

This command is used to download a file from the gama-server file system. It is triggered by sending a json object formatted as follows to the server

{
  "type": "download",
  "file": "<path_to_file>",
}

The file parameter is mandatory, it represents the path of the file to download expressed in gama-server's file system.

Answer from gama-server

On success, this command will trigger a json response message which content field will be filled with the file's content.

The upload command

This command is used to upload a file to the gama-server file system. It is triggered by sending a json object formatted as follows to the server

{
  "type": "download",
  "file": "<path_to_file>",
  "content": "<file_content>",
}

The file and content parameters are mandatory, they represent respectively the path where the file will be uploaded on gama-server's file system and the content of the file as a string of characters.

Answer from gama-server

This command has an empty content field in the response json sent by gama-server after processing it.

Gama-server messages

All messages send by gama-server follow a json architecture that is formatted as follows:

{
  "type": "some string describing the type of message",
  "content": "a field containing everything additional information for the message", //It can be a string, an int or a json object
  "exp_id": "contains the experiment id (as a string) to which this message is linked to", //Optional, its presence depends on the message's type
  "command": "a json containing the original command to which gama is responding",//Optional, is only present in messages responding directly to a command sent by the client
}

Messages types

All messages have in common a type field that informs the client of the type of message sent. The different types possibles are:

  • ConnectionSuccessful: Used when a client connected without any problem to gama-server
  • SimulationStatus: Signals a message representing a simulation status
  • SimulationStatusInform: Signals a message representing a simulation inform status
  • SimulationStatusError: Signals a message representing a simulation error status
  • SimulationStatusNeutral: Signals a message representing a simulation neutral status
  • SimulationOutput: Signals a message as would be written in the console by a write statement in gama with an interface
  • SimulationDebug: Signals a message as would be written in the console by a debug statement in gama with an interface
  • SimulationDialog: Signals a message representing what would be a dialog in gama with an interface
  • SimulationErrorDialog: Signals a message representing what would be an dialog in gama with an interface
  • SimulationError: Signals a message representing an error raised in a running simulation
  • RuntimeError: Signals a message representing an exception raised in gama-server while trying to process a command
  • GamaServerError: Signals a message representing an unknown exception raised in gama-server (can be unrelated to any command)
  • MalformedRequest: Signals that a command sent by the client doesn't follow the expected format (lack of parameter, wrong type etc.)
  • CommandExecutedSuccessfully: Signals that a command sent by the client was executed without any problem on gama-server
  • SimulationEnded: Signals that a running simulation reached its end condition and stopped. Beware if the simulation stops for another reason, this message won't be send.
  • UnableToExecuteRequest: Signals that a command cannot be executed, though it may be formatted correctly. It mainly occurs when trying to execute a command on a simulation that is not currently running.

Connection related answers

When your client is connected correctly to gama-server, a message is sent. Its type is ConnectionSuccessful with an empty content. In case of problem, the client may receive a message of type GamaServerError or just get a timeout/broken connection message at the socket level.

Command answers

For every command described in the commands section, the client will received a json answer formatted as follows:

{
  "type": "some string describing the type of message",
  "content": "a field containing every additional information for the message", //It can be a string, an int or a json object, depending on the type of message, it could also be empty
  "command": "a json containing the original command to which gama is responding"
}

So for example if you send an expression command to gama, with an experiment_id of value 2 and you want to evaluate the expression length(people) to know the number of agent people in that simulation. You may receive an answer looking like this:

{
  "type": "CommandExecutedSuccessfully", //The type indicates that everything went normally
  "content": 102, //There are 102 agents of the species people in your simulation at the time of evaluation
  "command": //The description of the command you sent, as interpreted by gama-server and turned into a json
  {
    "type": "expression",
    "exp_id": "2",
    "expr": "length(people)",
  }
}

The command field is very useful for clients that run multiple simulations and commands at the same time, as it can be used to retrace which command the message responds to. Note: The command field contains all the parameters of the command sent by the client, including those that are not useful for GAMA to execute the command, you can thus use it to store more data, like an internal id used by the client, some kind of counter etc..

In case there is an error resulting from the processing of your command, you may receive an error message of type:

  • MalformedRequest if you forgot a mandatory parameter to execute the command or gave objects that couldn't be de-serialized. The list of required parameters will be sent as a string in the content field.
  • UnableToExecuteRequest if you are trying to execute a command on a simulation that is not currently running or some other problem of "logic". You will find more informations in the content field.
  • RuntimeError and GamaServerError if while executing your command, an exception happens, either in gaml code for RuntimeError or in gama's code for GamaServerError. The exception's description will be given in the content field, as a json object containing the error message and the stack trace.

There is no exp_id field in those messages, because it is already included in all the command fields that are related to an experiment.

Errors and exceptions

In addition to the error messages you can receive when directly requesting to execute a command (MalformedRequest, UnableToExecuteRequest) described in the command answers section, or the network errors that can be raised for external problems, it is possible that gama-server encounters an exception while running. In that case gama-server will send a json message formatted as described in the Gama-server messages section, the two different types would either be GamaServerError, RuntimeError or SimulationError and the content field would be filled as follows:

{
  "exception": "The java class of the exception",
  "message": "The message describing the problem",
  "stack": [],//The stack trace of the exception given as a list of strings
}

the exp_id and command fields would be present if possible, depending on where the exception happens.

Simulations outputs

As mentioned in the introduction and the description of the load command. You can ask gama-server to redirect the simulation's outputs. There are 3 different types of output produced by a simulation that you can chose to redirect or not:

  • the messages in the dialogs
  • the messages in the status-bar
  • the messages in the gama console Each has an associated boolean that you have to set to true in the load command in order to have it redirected to the client.

The output messages are sent directly to the client as soon as they are asked by the simulation. The format of the output messages follows the usual message format. The exp_id will always be filled with the current experiment id, the 'command' field won't be present. The different types of messages possible are:

  • for dialog messages: SimulationDialog and SimulationErrorDialog respectively for normal dialogs and error dialogs
  • for status messages: SimulationStatusNeutral, SimulationStatusError, SimulationStatusInform, SimulationStatus
  • for console messages: SimulationOutput for the messages written with the write statement and SimulationDebug for the ones written with the debug statement.

The content field will be formatted as follows:

  • for dialog messages, it's directly a string containing the message
  • for status messages:
{
  "message": "the status message",
  "icon": "the name/path of the associated icon", //only present for some SimulationStatus and SimulationStatusInform messages
  "color": 
  {
    "r": "red value", 
    "g": "green value", 
    "b": "blue value",
  }, // The background color in the status-bar, only present in some SimulationStatus messages
}
  • for console messages:
{
  "message": "the message as it would be written in the console",
  "color": 
  {
    "r": red_value, 
    "g": green_value, 
    "b": blue value,
  }, // The text color
  "cycle": simulation_cycle, // the cycle of the simulation at the moment the debug statement is executed, only for SimulationDebug messages
}

Python wrapper

A python package is available to interact with Gama server as a client, you can find it here. It will take care of formatting the queries to the server and receiving the answers. You simply have to install the package into your python environment with the command pip install gama-client and then import gama_client into your python files. For more information follow the README.md available on the package's github.

Javascript Client

There is also a javascript client being developed in this repository gama.client

Hello World Visualization in MapBox

  • Clone the repository gama.client
  • In js/gama_client.js edit the following variable ABSOLUTE_PATH_TO_GAMA to your local path (e.g var ABSOLUTE_PATH_TO_GAMA = '/Users/arno/';)
  • open index.htmlin a browser

Hello World Message example

  • In js/simple_syntax.js edit the following variable modelPath to your model's path
  • open syntax.htmlin a browser

Troubleshooting

crash on load command

It is possible that gama-server starts and accepts connections, but crashes when receiving a load command with a message of the type:

java.lang.Exception: java.lang.NoClassDefFoundError: org/eclipse/core/resources/ResourcesPlugin
        at org.java_websocket.server.WebSocketServer$WebSocketWorker.run(WebSocketServer.java:1093)
Caused by: java.lang.NoClassDefFoundError: org/eclipse/core/resources/ResourcesPlugin
        at msi.gama.lang.gaml.indexer.GamlResourceIndexer.<clinit>(GamlResourceIndexer.java:54)
        at msi.gama.lang.gaml.resource.GamlResource.doLinking(GamlResource.java:362)
        at org.eclipse.xtext.resource.XtextResource.updateInternalState(XtextResource.java:304)
        at org.eclipse.xtext.resource.XtextResource.updateInternalState(XtextResource.java:292)
        at msi.gama.lang.gaml.resource.GamlResource.updateInternalState(GamlResource.java:308)
        at org.eclipse.xtext.resource.XtextResource.doLoad(XtextResource.java:182)
        at org.eclipse.xtext.linking.lazy.LazyLinkingResource.doLoad(LazyLinkingResource.java:115)
        at org.eclipse.emf.ecore.resource.impl.ResourceImpl.load(ResourceImpl.java:1563)
        at org.eclipse.emf.ecore.resource.impl.ResourceImpl.load(ResourceImpl.java:1342)
        at org.eclipse.emf.ecore.resource.impl.ResourceSetImpl.demandLoad(ResourceSetImpl.java:259)
        at org.eclipse.emf.ecore.resource.impl.ResourceSetImpl.demandLoadHelper(ResourceSetImpl.java:274)
        at org.eclipse.xtext.resource.XtextResourceSet.getResource(XtextResourceSet.java:266)
        at org.eclipse.xtext.resource.SynchronizedXtextResourceSet.getResource(SynchronizedXtextResourceSet.java:33)
        at msi.gama.lang.gaml.validation.GamlModelBuilder.buildModelDescription(GamlModelBuilder.java:111)
        at msi.gama.lang.gaml.validation.GamlModelBuilder.compile(GamlModelBuilder.java:97)
        at msi.gama.headless.core.HeadlessSimulationLoader.loadModel(HeadlessSimulationLoader.java:155)
        at msi.gama.headless.core.HeadlessSimulationLoader.loadModel(HeadlessSimulationLoader.java:126)
        at msi.gama.headless.job.JobListFactory.constructAllJobs(JobListFactory.java:46)
        at msi.gama.headless.script.ExperimentationPlanFactory.buildExperiment(ExperimentationPlanFactory.java:183)
        at msi.gama.headless.listener.LoadCommand.launchGamlSimulation(LoadCommand.java:74)
        at msi.gama.headless.listener.LoadCommand.execute(LoadCommand.java:37)
        at msi.gama.headless.listener.LoadCommand.execute(LoadCommand.java:1)
        at msi.gama.headless.listener.CommandExecutor.process(CommandExecutor.java:43)
        at msi.gama.headless.listener.GamaWebSocketServer.onMessage(GamaWebSocketServer.java:209)
        at org.java_websocket.server.WebSocketServer.onWebsocketMessage(WebSocketServer.java:712)
        at org.java_websocket.drafts.Draft_6455.processFrameText(Draft_6455.java:986)
        at org.java_websocket.drafts.Draft_6455.processFrame(Draft_6455.java:910)
...

This issue arises because gama-server tries to create a workspace for your experiment but does not have the appropriate rights to do it. It can be the case in windows if you run gama-server directly from the headless directory from the installation folder (protected by default) and that you are not an Admin

  1. What's new (Changelog)
  1. Installation and Launching
    1. Installation
    2. Launching GAMA
    3. Updating GAMA
    4. Installing Plugins
  2. Workspace, Projects and Models
    1. Navigating in the Workspace
    2. Changing Workspace
    3. Importing Models
  3. Editing Models
    1. GAML Editor (Generalities)
    2. GAML Editor Tools
    3. Validation of Models
  4. Running Experiments
    1. Launching Experiments
    2. Experiments User interface
    3. Controls of experiments
    4. Parameters view
    5. Inspectors and monitors
    6. Displays
    7. Batch Specific UI
    8. Errors View
  5. Running Headless
    1. Headless Batch
    2. Headless Server
    3. Headless Legacy
  6. Preferences
  7. Troubleshooting
  1. Introduction
    1. Start with GAML
    2. Organization of a Model
    3. Basic programming concepts in GAML
  2. Manipulate basic Species
  3. Global Species
    1. Regular Species
    2. Defining Actions and Behaviors
    3. Interaction between Agents
    4. Attaching Skills
    5. Inheritance
  4. Defining Advanced Species
    1. Grid Species
    2. Graph Species
    3. Mirror Species
    4. Multi-Level Architecture
  5. Defining GUI Experiment
    1. Defining Parameters
    2. Defining Displays Generalities
    3. Defining 3D Displays
    4. Defining Charts
    5. Defining Monitors and Inspectors
    6. Defining Export files
    7. Defining User Interaction
  6. Exploring Models
    1. Run Several Simulations
    2. Batch Experiments
    3. Exploration Methods
  7. Optimizing Model Section
    1. Runtime Concepts
    2. Optimizing Models
  8. Multi-Paradigm Modeling
    1. Control Architecture
    2. Defining Differential Equations
  1. Manipulate OSM Data
  2. Diffusion
  3. Using Database
  4. Using FIPA ACL
  5. Using BDI with BEN
  6. Using Driving Skill
  7. Manipulate dates
  8. Manipulate lights
  9. Using comodel
  10. Save and restore Simulations
  11. Using network
  12. Headless mode
  13. Using Headless
  14. Writing Unit Tests
  15. Ensure model's reproducibility
  16. Going further with extensions
    1. Calling R
    2. Using Graphical Editor
    3. Using Git from GAMA
  1. Built-in Species
  2. Built-in Skills
  3. Built-in Architecture
  4. Statements
  5. Data Type
  6. File Type
  7. Expressions
    1. Literals
    2. Units and Constants
    3. Pseudo Variables
    4. Variables And Attributes
    5. Operators [A-A]
    6. Operators [B-C]
    7. Operators [D-H]
    8. Operators [I-M]
    9. Operators [N-R]
    10. Operators [S-Z]
  8. Exhaustive list of GAMA Keywords
  1. Installing the GIT version
  2. Developing Extensions
    1. Developing Plugins
    2. Developing Skills
    3. Developing Statements
    4. Developing Operators
    5. Developing Types
    6. Developing Species
    7. Developing Control Architectures
    8. Index of annotations
  3. Introduction to GAMA Java API
    1. Architecture of GAMA
    2. IScope
  4. Using GAMA flags
  5. Creating a release of GAMA
  6. Documentation generation

  1. Predator Prey
  2. Road Traffic
  3. 3D Tutorial
  4. Incremental Model
  5. Luneray's flu
  6. BDI Agents

  1. Team
  2. Projects using GAMA
  3. Scientific References
  4. Training Sessions

Resources

  1. Videos
  2. Conferences
  3. Code Examples
  4. Pedagogical materials
Clone this wiki locally