Model

class Melodie.model.ModelRunRoutine(max_step: int, model: Model)

Bases: object

A simple iterator for model run.

When calling Model.iterator() method, a ModelRunRoutine object will be created, yielding an int value reprensenting the current number of step, ranging [0, max_step - 1].

class Melodie.model.Model(config: Config, scenario: Scenario, run_id_in_scenario: int = 0, visualizer: Visualizer | None = None)

Bases: object

The base class for Model.

There are three major methods, create(), setup() and run(). create() and then setup() are called when the model creates, and run() is called for model running.

To build up your own model, inherit this class and override create(), setup() and run()

Parameters:

  • config – Type Melodie.Config

  • scenario – Type Melodie.Scenario containing model parameters.

  • run_id_in_scenario – Current run_id in the current scenario, an int from [0, number_of_run ), and 0 by default.

  • visualizer – Visualizer instance if needs visualization, None by default, indicating no need for visualization.

create()

An initialization method, which is called immediately right after the Model object is created.

Returns:

None

setup()

General method for model setup, which is called after Model.create()

Returns:

None

create_db_conn() → DBConn

Create a database connection with the project configuration.

Returns:

DBConn object

create_agent_list(agent_class: Type[Agent])

Create an agent list object. A model could contain multiple ``AgentList``s.

Parameters:

agent_class – The class of desired agent type.

Returns:

Agentlist object created

create_environment(env_class: Type[Environment])

Create the environment of model. Notice that a model has only one environment.

Parameters:

env_class –

Returns:

Environment object created

create_grid(grid_cls: Type[Grid] | None = None, spot_cls: Type[Spot] | None = None)

Create a grid.

Parameters:

  • grid_cls – The class of grid, Melodie.Grid by default.

  • spot_cls – The class of spot, Melodie.Spot by default.

Returns:

Grid object.

create_network(network_cls: Type[Network] | None = None, edge_cls: Type[Edge] | None = None)

Create the network of model.

Parameters:

  • network_cls – The type of network object, Melodie.Network by default.

  • edge_cls – The type of edge object, Melodie.Edge by default.

Returns:

Network object created

create_data_collector(data_collector_cls: Type[DataCollector])

Create the data collector of model.

Parameters:

data_collector_cls – The datacollector class, must be a custom class inheriting Melodie.DataCollector.

Returns:

Datacollector object created.

create_agent_container(agent_class: Type[Agent], initial_num: int, params_df: pd.DataFrame = None, container_type: str = 'list') → AgentList | AgentDict

Create a container for agents.

Parameters:

  • agent_class –

  • initial_num – Initial number of agents

  • params_df – Pandas DataFrame

  • container_type – a str, “list” or “dict”

Returns:

Agent container created

run()

Model run. Be sure to inherit this method on your model.

Returns:

None

iterator(period_num: int)

Return an iterator which iterates from 0 to period_num-1. In each iteration, the iterator updates the visualizer if it exists.

Parameters:

period_num – How many periods will this model run.

Returns:

None

init_visualize()

Be sure to implement it if you would like to use visualizer.

Returns: