booldog.io.interaction_networks

Read standard networks formats/objects (networkx, igraph, graphml, SIF) and convert to Boolean network.

Attributes

_IGRAPH_AVAILABLE

logger

Functions

networkx2interactions(g[, edge_type_key])

Convert a NetworkX Graph into a list of interactions.

igraph2interactions(g[, node_id_key, edge_type_key])

Convert an igraph Graph into a list of interactions.

read_interactions(interactions_input[, node_names])

Create Network from interactions (list)

read_sif(file[, delim, source_col, target_col, ...])

Reads in a SIF file of interactions

read_igraph(g[, node_id_key, edge_type_key, node_name_key])

Create BooleanNetwork from a igraph Graph object.

read_networkx(g[, edge_type_key, node_name_key])

Create BooleanNetwork from a networkx (Di)Graph object.

read_graphml(file[, edge_type_key, node_id_key, ...])

Extract relevant parts for a Boolean network from a graphml file.

Module Contents

booldog.io.interaction_networks._IGRAPH_AVAILABLE = True
booldog.io.interaction_networks.logger
booldog.io.interaction_networks.networkx2interactions(g, edge_type_key='interaction', **_)

Convert a NetworkX Graph into a list of interactions.

Parameters:

  • g (networkx.(Di)Graph) – The input graph object from the Networkx library. (Assumed to be directed.)

  • edge_type_key (str, optional) – The edge attribute key to use for interaction values (e.g., weight, type). Default is “interaction”.

  • **_ (dict) – Additional keyword arguments (ignored).

Returns:

A list of (source, target, symbol) tuples, one per edge of g, where symbol is the value of the edge_type_key edge attribute.

Return type:

list of tuple

booldog.io.interaction_networks.igraph2interactions(g, node_id_key='name', edge_type_key='interaction')

Convert an igraph Graph into a list of interactions.

Parameters:

  • g (igraph.Graph) – The input graph object from the igraph library.

  • node_id_key (str, optional) – The vertex attribute key that contains the primary identifier to use for node identifiers. Default is “name”.

  • edge_type_key (str, optional) – The edge attribute key to use for interaction type (e.g., weight, type). Default is “interaction”.

Returns:

A list of (source, target, symbol) tuples, one per edge of g, where source/target are the node_id_key vertex attribute values of the edge’s endpoints, and symbol is the value of the edge_type_key edge attribute.

Return type:

list of tuple

booldog.io.interaction_networks.read_interactions(interactions_input, node_names=None, **kwargs)

Create Network from interactions (list)

Parameters:

  • interactions_input (list or str) – A list of interactions in the network. Each interaction should be a tuple of (source, target, sign), where source and target are node identifiers and sign is either activator_symbol or inhibitor_symbol. Or a path to a JSON file containing such a list of interactions.

  • activator_symbol (int, optional) – The value representing activation in the network. Default is 1.

  • inhibitor_symbol (int, optional) – The value representing inhibition in the network. Default is -1.

  • logic (LogicBuilder, optional) – An optional logic builder (see booldog.io.interaction_logic.LogicBuilder) used to construct each node’s update rule from its regulators. Default is SquadLogic.

  • **kwargs (dict) – Additional keyword arguments, forwarded to booldog.io.interaction_logic.interactions2rules.

Returns:

n – A BoolDogModel object representing the Boolean network.

Return type:

booldog.BoolDogModel

Notes

Uses SQUAD logic to obtain Boolean network.

booldog.io.interaction_networks.read_sif(file, delim='\t', source_col=0, target_col=1, interaction_col=2, header=True, activator_symbol='1', inhibitor_symbol='-1', **kwargs)

Reads in a SIF file of interactions

Parameters:

  • file (str) – Path to the SIF file

  • delim (str, optional) – Delimiter used in the SIF file (default=” “)

  • header (bool, optional) – If the first line of the file is a header (default=True)

  • source_col (int or str, optional) – Column index (if int) or column name (if str) of source node (default=0)

  • target_col (int or str, optional) – Column index (if int) or column name (if str) of target node (default=1)

  • interaction_col (int, optional) – Column index (if int) or column name (if str) of interaction type (symbol) (default=2)

  • activator_symbol (int or str, optional) – Symbol of activation edges in interaction_col, compared against the (string) values read from the file. Default is "1" (a string, matching the fact that values parsed from the file are always strings).

  • inhibitor_symbol (int or str, optional) – Symbol of inhibition edges in interaction_col. Default is "-1" (see activator_symbol).

  • logic (LogicBuilder, optional) – An optional logic builder (see booldog.io.interaction_logic.LogicBuilder) used to construct each node’s update rule from its regulators. Default is SquadLogic.

  • **kwargs (dict) – Additional keyword arguments, forwarded to booldog.io.interaction_logic.interactions2rules.

Returns:

n – A BoolDogModel object representing the Boolean network.

Return type:

booldog.BoolDogModel

Notes

Uses SQUAD logic by default (see logic) to obtain the Boolean network.

booldog.io.interaction_networks.read_igraph(g, node_id_key='name', edge_type_key='interaction', node_name_key='name', **kwargs)

Create BooleanNetwork from a igraph Graph object.

Parameters:

  • g (igraph.Graph) – The input graph object from the igraph library.

  • node_id_key (str, optional) – The vertex attribute key that contains the primary identifier used for each node. Default is “name”.

  • node_name_key (str, optional) – The vertex attribute key that contains the node name (e.g. display label). Default is “name”.

  • edge_type_key (str, optional) – The edge attribute key to use for interaction values (e.g., weight, type). Default is “interaction”.

  • activator_symbol (int or str, optional) – Value of activation edges in edge_type_key of g, compared against the raw attribute value. Default is 1 (int, from interactions2rules).

  • inhibitor_symbol (int or str, optional) – Value of inhibition edges in edge_type_key of g. Default is -1 (see activator_symbol).

  • logic (LogicBuilder, optional) – An optional logic builder (see booldog.io.interaction_logic.LogicBuilder) used to construct each node’s update rule from its regulators. Default is SquadLogic.

  • **kwargs (dict) – Additional keyword arguments, forwarded to booldog.io.interaction_logic.interactions2rules.

Returns:

n – A BoolDogModel object representing the Boolean network.

Return type:

booldog.BoolDogModel

Notes

Uses SQUAD logic by default (see logic) to obtain the Boolean network.

booldog.io.interaction_networks.read_networkx(g, edge_type_key='interaction', node_name_key='name', **kwargs)

Create BooleanNetwork from a networkx (Di)Graph object.

Parameters:

  • g (networkx.(Di)Graph) – The input graph object from the networkx library.

  • edge_type_key (str, optional) – The edge attribute key to use for interaction values (e.g., weight, type). Default is “interaction”.

  • activator_symbol (int or str, optional) – Value of activation edges in edge_type_key of g, compared against the raw attribute value. Default is 1 (int, from interactions2rules).

  • inhibitor_symbol (int or str, optional) – Value of inhibition edges in edge_type_key of g. Default is -1 (see activator_symbol).

  • node_name_key (str, optional) – The node attribute key that contains the node name (display label). Default is “name”.

  • logic (LogicBuilder, optional) – An optional logic builder (see booldog.io.interaction_logic.LogicBuilder) used to construct each node’s update rule from its regulators. Default is SquadLogic.

  • **kwargs (dict) – Additional keyword arguments, forwarded to both networkx2interactions (which ignores unrecognized keys) and booldog.io.interaction_logic.interactions2rules.

Returns:

n – A BoolDogModel object representing the Boolean network.

Return type:

booldog.BoolDogModel

Notes

Uses SQUAD logic by default (see logic) to obtain the Boolean network.

booldog.io.interaction_networks.read_graphml(file, edge_type_key='interaction', node_id_key='name', node_name_key='name', yEd_labels=False, yEd_arrow_head=False, use_labels=True, logic=None, **kwargs)

Extract relevant parts for a Boolean network from a graphml file.

Since graphml is not well defined for Boolean networks (or even standard for interaction networks), this read functionality has limited support.

Parameters:

  • file (str) – Path to the graphml file

  • edge_type_key (str, optional) – The edge attribute key to use for interaction type (e.g., weight, type). Default is “interaction”. Only used if yEd_arrow_head is False (yEd uses the arrow head symbol instead).

  • node_id_key (str, optional) – The vertex attribute key that contains the primary identifier used for each node. Default is “name”. If this attribute is missing from the file, the igraph-assigned “id” attribute is used instead (with a warning logged).

  • node_name_key (str, optional) – The vertex attribute key that contains the node name (e.g. display label). Default is “name”. If this attribute is missing (and differs from node_id_key), node names fall back to node_id_key (with a warning logged).

  • yEd_labels (bool, optional) – If graphml file originates as a yEd export, the node attribute “y:NodeLabel” is used to determine node names (default=False)

  • yEd_arrow_head (bool, optional) – If graphml file originates as a yEd export, the edge attribute “y:Arrows” is used to determine interaction type (default=False)

  • use_labels (bool, optional) – Accepted for future use but currently unused by this function. Default is True.

  • activator_symbol (int or str, optional) – Symbol or value of activation edges. If yEd_arrow_head is True, default is “standard” (the yEd arrow-head symbol for a plain arrow), else default is 1.

  • inhibitor_symbol (int or str, optional) – Symbol or value of inhibition edges. If yEd_arrow_head is True, default is “t_shape” (the yEd arrow-head symbol for a T-bar), else default is -1.

  • logic (LogicBuilder, optional) – An optional logic builder (see booldog.io.interaction_logic.LogicBuilder) used to construct each node’s update rule from its regulators. Default is SquadLogic.

  • **kwargs (dict) – Additional keyword arguments. Only activator_symbol and inhibitor_symbol are read from this dict (see above); any other keys are ignored, since, unlike the other read_* functions, read_graphml does not forward **kwargs on to interactions2rules.

Returns:

n – A BoolDogModel object representing the Boolean network.

Return type:

booldog.BoolDogModel

Notes

Uses igraph to parse the graphml file, and extracts node and edge attributes to determine interactions. Uses SQUAD logic by default (see logic) to build rules.

Edges whose interaction value is missing or matches neither activator_symbol nor inhibitor_symbol are dropped with a logged warning, except with yEd_arrow_head=True, where an edge with no arrow symbol at all raises a ValueError. Note that igraph converts values according to the GraphML attr.type of the key: a double/int key yields numbers (matching the default 1/-1), whereas a string key yields e.g. "1", requiring activator_symbol="1"/inhibitor_symbol="-1".

If yEd_label=True, it also parses the yEd-specific attributes to extract node labels (y:NodeLabel) for node names.

If yEd_arrow_head=True, it parses the yEd-specific edge arrow types (arrow head symbols of y:Arrows) to determine interaction types (activation/inhibition).