CopyMeGraph — A good place to start with your graph-based model

CopyMeGraph — A good place to start with your graph-based model#

This template model supplies a minimal graph-based Utopia model. On top of what the CopyMeBare model template provides, this model contains:

  • Infrastructure for graph setup and data writing

  • A customisable GraphType (based on the Boost Graph Library), including vertex and edge properties

  • Configuration files showcasing some of the graph setup and plotting capabilities

  • A cfgs directory with further example configuration files

Use this model if you want to start “from scratch” but want to avoid setting up the infrastructure itself.

Hint

Refer to Step-by-step Guide for information on how to implement your own model based on this template.

Default configuration parameters#

Below are the default configuration parameters for the CopyMeGraph model.

# --- Space parameters
# The physical space this model is embedded in
space:
  periodic: true

# -- Network and agent initialization -----------------------------------------
create_graph:
  model: regular
  num_vertices: 30
  mean_degree: 1

  # --- Model-specific parameters
  regular:
    # Whether the underlying lattice is oriented (directed only)
    oriented: true

  ErdosRenyi:
    # Allow parallel edges
    parallel: false

    # Allow self edges
    self_edges: false

  WattsStrogatz:
    # Rewiring probability
    p_rewire: 0.

    # Orientation of the underlying lattice (directed only)
    oriented: false

  BarabasiAlbert:
    # Allow parallel edges
    parallel: false

  BollobasRiordan:
    # Graph generating parameters
    alpha: 0.2
    beta: 0.8
    gamma: 0.
    del_in: 0.
    del_out: 0.5

# Vertex initialization parameters
init_some_state: 3.14
init_some_trait: 42

# Edge initialization parameters
init_weight: 1.

# Use a random weight between zero and `init_weight`
init_random_weight: true

# --- Other initialization parameters
# Add parameters and their documentation here ...

# -- Dynamics
# Some parameter that controls some mechanism during some part of the dynamics
some_parameter: 0.1

# Add further parameters and their documentation here ...

Available plots#

The following plot configurations are available for the CopyMeGraph model:

Default Plot Configuration#

# -- Graph plots --------------------------------------------------------------
graph:
  based_on: graph


# -- Time series plots --------------------------------------------------------
time_series/some_state_mean_and_std:
  based_on: some_state_mean_and_std

time_series/some_trait_mean_and_std:
  based_on: some_trait_mean_and_std

time_series/single_vertex_states:
  based_on: single_vertex_states

time_series/single_vertex_states_dynamic:
  based_on: single_vertex_states_dynamic

Base Plot Configuration#

.variables:
  base_path: &base_path data/CopyMeGraph



# =============================================================================
#  ╔╦╗╔═╗╔╦╗╔═╗╦  ╔═╗╔╦╗╔═╗╔═╗
#   ║ ║╣ ║║║╠═╝║  ╠═╣ ║ ║╣ ╚═╗
#   ╩ ╚═╝╩ ╩╩  ╩═╝╩ ╩ ╩ ╚═╝╚═╝
# =============================================================================
# -- Overloads ----------------------------------------------------------------
# Overload some configs to insert model-specific settings

# Model-specific defaults
.defaults:
  based_on: .defaults

  # Can define something here ...


# .. Creators .................................................................
.creator.universe:
  based_on:
    - .creator.universe
    - .defaults

  dag_options:
    select_path_prefix: *base_path

.creator.multiverse:
  based_on:
    - .creator.multiverse
    - .defaults

  select_and_combine:
    base_path: *base_path



# -- Plot templates -----------------------------------------------------------
# Time series facet grid plots
.plot.time_series.base:
  based_on:
    - .hlpr.kind.time_series
    - .plot.facet_grid.with_auto_encoding
    - .plot.facet_grid.line

  x: time

.plot.time_series.uni:
  based_on:
    - .creator.universe
    - .plot.time_series.base

.plot.time_series.mv:
  based_on:
    - .creator.multiverse
    - .plot.time_series.base

# A graph plot for universe data
.plot.graph:
  based_on:
    - .creator.universe
    - .plot.graph




# =============================================================================
#  ╔═╗╦  ╔═╗╔╦╗╔═╗
#  ╠═╝║  ║ ║ ║ ╚═╗
#  ╩  ╩═╝╚═╝ ╩ ╚═╝
# =============================================================================

# .. Mean value time series ...................................................
some_state_mean_and_std:
  # Base it on existing configurations
  based_on:
    - .plot.time_series.uni
    - .plot.facet_grid.errorbands

  # Select the data by defining a variable that is also used to set the y label
  dag_options:
    define:
      data_to_select: g_static/some_state

  select:
    _data:
      path: "."
      transform:
        - getitem: [!dag_prev , !dag_tag data_to_select]

  # Compute mean and std and assemble into Dataset for errorbar plot
  transform:
    - .mean: [!dag_tag _data, [vertex_idx]]
      tag: mean
    - .std: [!dag_tag _data, [vertex_idx]]
      tag: std

    - xr.Dataset:
      - mean: !dag_tag mean
        std: !dag_tag std
      tag: data

  y: mean
  yerr: std

  helpers:
    set_labels:
      y: !dag_result data_to_select

# Same for the trait value
some_trait_mean_and_std:
  based_on: some_state_mean_and_std

  # Need only adapt the path to the dataset, rest is inherited
  dag_options:
    define:
      data_to_select: g_static/some_trait


# .. Individual cell states ...................................................
# Plot individual vertex states over time using a generic plotting function
single_vertex_states:
  based_on: .plot.time_series.uni

  # Select some single vertices to plot the states of
  select:
    data:
      path: g_static/some_state
      with_previous_result: true  # makes the select data available as first
                                  # positional argument to the transform
                                  # operation; and the result of the first
                                  # transformation to the first argument of the
                                  # second transformation, etc.
      transform:
        - .sel: {vertex_idx: [0, 1, 2, 4, 8]}

  helpers:
    set_title:
      title: Individual Vertex States

# Perform the same plot with the data from the graph group 'g_dynamic'. In this
# case, the data is loaded as a `TimeSeriesGroup` (and not as Xarray) which
# also supports important xarray-methods such as coordinate access using .sel.
single_vertex_states_dynamic:
  based_on: single_vertex_states

  select:
    data:
      path: g_dynamic/some_state

  helpers:
    set_labels:
      y: cell state


# .. Graph plot ...............................................................
# Plot the final graph. Map the 'some_state' data to the node color, the node
# degree to the size of the nodes, and the weights to the edge width.
graph:
  based_on: .plot.graph

  select:
    graph_group: g_dynamic

  # Configure the created networkx graph object
  graph_creation:
    at_time_idx: -1                           # Select the last timestep
    node_props: ['some_trait', 'some_state']  # Select node and edge
    edge_props: ['weights']                   # properties from the GraphGroup

  # Configure the graph layout
  graph_drawing:
    positions:
      model: spring                   # The spring model positions the nodes
                                      # such that edge lengths are reduced.
      k: 1                            # Tweaking the spring model by adjusting
                                      # the optimal edge length (k) and the
      iterations: 100                 # maximum number of iterations done.
    nodes:
      node_color:
        from_property: some_state     # Map the 'some_state' data to the color
      node_size:
        from_property: degree         # Map the node degree to the node size
        scale_to_interval: [10, 200]  # Scale the (degree-)data to the given
                                      # interval (the default size is 300).
    edges:
      width:
        from_property: weights        # Map the 'weights' data to the edge
        scale_to_interval: [0.1, 1.2] # width and scale the data accordingly.

  helpers:
    set_title:
      title: Final Graph

For available base plots, see Base Plot Configuration Pool.

Contents