timelink.networks package

Generation and analysis of Networks in Timelink

Submodules

timelink.networks.network_draw module

timelink.networks.network_draw.draw_network(Graph: Graph, title='Timelink network', node_tooltips=None, edge_tooltips=None, iterations=50, scale=1, node_color_attr='type', node_colors: dict | None = None, color_palette: dict | None = None, circle_size=10, line_color='#CCCCCC', line_alpha=0.8, line_width=2, filename=None, width=500, height=500)[source]

Draws a network using Bokeh and networkx layout

Parameters:

  • Graph – a networkx graph object

  • title – the title of the plot

  • node_tooltips – a list of tuples with the information to show when hovering over nodes defaults to [(“desc”, “@desc”), (“type”, “@type”)]

  • edge_tooltips – a list of tuples with the information to show when hovering over edges defaults to None, which means no tooltips for edges

  • iterations – number of iterations for the spring layout

  • node_color_attr – the attribute to color nodes by (default is “type”)

  • node_colors – a dictionary mapping node types to colors. If None, a default color palette is used.

  • color_palette – a dictionary mapping node types to colors. If none a default color palette is used.

  • filename – if provided, saves the plot to this file

  • width – width of the plot

  • height – height of the plot

Returns:

None, but saves the plot to a file if filename is provided

Note on coloring of nodes:

Coloring is done by the attribute specified in node_color_attr. Each node is inspected to get its attribute node_color_attr. If node_colors is provided, it uses that dictionary to set the color. node_colors should be a dictionary mapping attributes to colors. Colors are set to “gray” if the attribute is not found in node_colors. If node_colors is None, this functions will inspect the nodes to get the number of different values for node_color_attr and build a color palette using bokeh.palettes.Category10 or a custom color_palette if provided in the arguments.

timelink.networks.network_generation module

Generation of networks

timelink.networks.network_generation.network_from_attribute(attribute: str, ignore_values: list[str] | None = None, mode='cliques', by_year=False, user=None, db: TimelinkDatabase | None = None, session=None) → Graph[source]

Generate a network from common attribute values.

This function will generate a network connecting the entities that have the same value for the attribute given in the parameter.

Parameters:

  • attribute (str) – The name (type) of the attribute used for generating the graph.

  • ignore_values (list[str], optional) – A list of values to ignore when generating the network. Defaults to None.

  • mode (str, optional) – The topology of the generated network (see bellow). Valid values are “cliques” and “value-node”. Defaults to “cliques”.

  • user (str, optional) – Use real persons identified by this user. Defaults to “none”.

  • db (TimelinkDatase, optional) – The TimelinkDatase object. Defaults to None. Either db or session must be provided.

  • session (object, optional) – The session object for the database connection. Defaults to None.

Raises:

None –

Topology the generated network:

  • If mode = “cliques” all the entities with attribute will be nodes in the graph and edges will be created between the entities with the same value of the attribute.

  • If mode = “value-node” a node will be created for each different value of the attribute a and edges will be created linking that node to the entities which have that value in attribute a.

  • In both cases entities with several values for the attribute contribute to the overall connectivity of the graph, by linking clusters of same value entities.

Returns:

The generated network as a networkx Graph object.

Return type:

networkx.classes.graph.Graph

Each node will have an associated dictionary of attributes:

  • “id”: (entity id)

  • “type”: “value_node” or entity class in the database.

  • “desc”: a description of the node, automatically fetched

    from the database (names of persons for instance)

  • “is_real”: a flag stating if the “id” key refers to a real

    entity or to an occurrence.

  • “url”: a link to the entity information in the database

    in the format http://localhost:8080/mhk/database/id/entityID

Each edge will have associated the following key-value pairs:

  • “date1”: date of the atribute in the left most node

  • “date2”: date of the attribute in the right most node

  • “attribute”: the type of the attribute

  • “value”: the value of the attribute

Examples:

Generate a network of people that graduated in the same place

G = network_from_attribute("graduated_at", mode="value-node")