agentlab.experiments.view_dep_graph

Dirty script to visualize the dependency graph of a benchmark, e.g. webarena, vsisualwebarena, etc. You may have to detust it to make it work for you.

Functions

clean_dict(dependency_dict)

compress_chains(G)

Compress chains in a directed graph by merging nodes that have single parent and single child.

compress_sequential_chains(dep_dict)

Compress chains of sequential numbers in a dependency dictionary.

dict_to_networkx(dependency_dict)

plot_components_grid(components[, max_cols, ...])

Plot components in a grid layout.

plot_graph(G, ax[, title, node_color, ...])

Plot a single graph component on the given matplotlib axis.
agentlab.experiments.view_dep_graph.clean_dict(dependency_dict: dict[str, list[str]]) → dict[str, list[str]]
agentlab.experiments.view_dep_graph.compress_chains(G)

Compress chains in a directed graph by merging nodes that have single parent and single child.

Parameters:

G – NetworkX directed graph

Returns:

NetworkX directed graph with compressed chains

agentlab.experiments.view_dep_graph.compress_sequential_chains(dep_dict: dict[str, list[str]]) → dict[str, list[str]]

Compress chains of sequential numbers in a dependency dictionary. Returns a new dictionary with compressed chains using range notation.

Parameters:

dep_dict – Dictionary mapping string numbers to list of string number dependencies

Returns:

Dictionary with compressed chains using range notation

agentlab.experiments.view_dep_graph.dict_to_networkx(dependency_dict: dict[str, list[str]]) → DiGraph
agentlab.experiments.view_dep_graph.plot_components_grid(components, max_cols=4, node_color='lightblue', node_size=2000, font_size=10)

Plot components in a grid layout.

Parameters:

  • components – List of NetworkX graphs, one per component

  • max_cols – Maximum number of columns in the grid

  • node_color – Color for the nodes

  • node_size – Size of the nodes

  • font_size – Size of the node labels

Returns:

matplotlib figure

agentlab.experiments.view_dep_graph.plot_graph(G, ax, title=None, node_color='lightblue', node_size=40, font_size=8)

Plot a single graph component on the given matplotlib axis.

Parameters:

  • G – NetworkX graph (should be a single connected component)

  • ax – Matplotlib axis to plot on

  • title – Optional title for the subplot

  • node_color – Color for the nodes

  • node_size – Size of the nodes

  • font_size – Size of the node labels