igraph#

igraph is a popular graph library that PyGraphistry can interface with. This allows you to leverage igraph’s layout algorithms, and optionally, algorithmic enrichments. It is CPU-based and can generally handle small/medium-sized graphs.

graphistry.plugins.igraph.compute_igraph(self, alg, out_col=None, directed=None, use_vids=False, params={}, stringify_rich_types=True)#

Enrich or replace graph using igraph methods

igraph is a CPU-only library. cuDF DataFrames are automatically converted to pandas before calling igraph, and the result is converted back. Polars frames (eager or lazy) also run on pandas and come back as eager Polars frames with their identifier dtypes. For a GPU-native alternative, see compute_cugraph().

Parameters:

alg (str) – Name of an igraph.Graph method like pagerank
out_col (Optional[str]) – For algorithms that generate a node attribute column, out_col is the desired output column name. When None, use the algorithm’s name. (default None)
directed (Optional[bool]) – During the to_igraph conversion, whether to be directed. If None, try directed and then undirected. (default None)
use_vids (bool) – During the to_igraph conversion, whether to interpret IDs as igraph vertex IDs (non-negative integers) or arbitrary values (False, default)
params (dict) – Any named parameters to pass to the underlying igraph method
stringify_rich_types (bool) – When rich types like igraph.Graph are returned, which may be problematic for downstream rendering, coerce them to strings
self (Plottable)

Returns:

Plotter

Return type:

Plotter

Example: Pagerank

import graphistry, pandas as pd
edges = pd.DataFrame({'s': ['a','b','c','d'], 'd': ['c','c','e','e']})
g = graphistry.edges(edges, 's', 'd').materialize_nodes()
g2 = g.compute_igraph('pagerank')
assert 'pagerank' in g2._nodes.columns

Example: Pagerank with custom name

import graphistry, pandas as pd
edges = pd.DataFrame({'s': ['a','b','c','d'], 'd': ['c','c','e','e']})
g = graphistry.edges(edges, 's', 'd').materialize_nodes()
g2 = g.compute_igraph('pagerank', out_col='my_pr')
assert 'my_pr' in g2._nodes.columns

Example: Pagerank on an undirected

import graphistry, pandas as pd
edges = pd.DataFrame({'s': ['a','b','c','d'], 'd': ['c','c','e','e']})
g = graphistry.edges(edges, 's', 'd').materialize_nodes()
g2 = g.compute_igraph('pagerank', directed=False)
assert 'pagerank' in g2._nodes.columns

Example: Pagerank with custom parameters

import graphistry, pandas as pd
edges = pd.DataFrame({'s': ['a','b','c','d'], 'd': ['c','c','e','e']})
g = graphistry.edges(edges, 's', 'd').materialize_nodes()
g2 = g.compute_igraph('pagerank', params={'damping': 0.85})
assert 'pagerank' in g2._nodes.columns

Example: Personalized Pagerank

import graphistry, pandas as pd
edges = pd.DataFrame({'s': ['a','b','c','d'], 'd': ['c','c','e','e']})
g = graphistry.edges(edges, 's', 'd').materialize_nodes()
g2 = g.compute_igraph('personalized_pagerank')
assert 'personalized_pagerank' in g2._nodes.columns

graphistry.plugins.igraph.ensure_pandas(df)#

Convert to pandas if not already (e.g. cuDF). No-op for pandas.

Delegates to graphistry.compute.engine_coercion.ensure_pandas(). Defined here to avoid a circular import at module load time.

Parameters:: df (Any)
Return type:: DataFrame

graphistry.plugins.igraph.from_igraph(self, ig, node_attributes=None, edge_attributes=None, load_nodes=True, load_edges=True, merge_if_existing=True)#

Convert igraph object into Plotter

If base g has _node, _source, _destination definitions, use them

When merge_if_existing with preexisting nodes/edges df and shapes match ig, combine attributes

For merge_if_existing to work with edges, must set g._edge and have corresponding edge index attribute in igraph.Graph

Parameters:

ig (igraph) – Source igraph object
node_attributes (Optional[List[str]]) – Subset of node attributes to load; None means all (default)
edge_attributes (Optional[List[str]]) – Subset of edge attributes to load; None means all (default)
load_nodes (bool) – Whether to load nodes dataframe (default True)
load_edges (bool) – Whether to load edges dataframe (default True)
merge_if_existing (bool) – Whether to merge with existing node/edge dataframes (default True)
merge_if_existing – bool

Returns:

Plotter

Return type:

Plottable

Example: Convert from igraph, including all node/edge properties

import graphistry, pandas as pd
edges = pd.DataFrame({'s': ['a', 'b', 'c', 'd'], 'd': ['b', 'c', 'd', 'e'], 'v': [101, 102, 103, 104]})
g = graphistry.edges(edges, 's', 'd').materialize_nodes().get_degrees()
assert 'degree' in g._nodes.columns
g2 = g.from_igraph(g.to_igraph())
assert len(g2._nodes.columns) == len(g._nodes.columns)

Example: Enrich from igraph, but only load in 1 node attribute

import graphistry, pandas as pd
edges = pd.DataFrame({'s': ['a', 'b', 'c', 'd'], 'd': ['b', 'c', 'd', 'e'], 'v': [101, 102, 103, 104]})
g = graphistry.edges(edges, 's', 'd').materialize_nodes().get_degree()
assert 'degree' in g._nodes
ig = g.to_igraph(include_nodes=False)
assert 'degree' not in ig.vs
ig.vs['pagerank'] = ig.pagerank()
g2 = g.from_igraph(ig, load_edges=False, node_attributes=[g._node, 'pagerank'])
assert 'pagerank' in g2._nodes
asssert 'degree' in g2._nodes

graphistry.plugins.igraph.layout_igraph(self, layout, directed=None, use_vids=False, bind_position=True, x_out_col='x', y_out_col='y', play=0, params={})#

Compute graph layout using igraph algorithm. For a list of layouts, see layout_algs or igraph documentation.

igraph is a CPU-only library. cuDF DataFrames are automatically converted to pandas before calling igraph, and the result is converted back. For a GPU-native alternative, see layout_cugraph().

Parameters:

layout (str) – Name of an igraph.Graph.layout method like sugiyama
directed (Optional[bool]) – During the to_igraph conversion, whether to be directed. If None, try directed and then undirected. (default None)
use_vids (bool) – Whether to use igraph vertex ids (non-negative integers) or arbitary node ids (False, default)
bind_position (bool) – Whether to call bind(point_x=, point_y=) (default True)
x_out_col (str) – Attribute to write x position to. (default ‘x’)
y_out_col (str) – Attribute to write x position to. (default ‘y’)
play (Optional[str]) – If defined, set settings(url_params={‘play’: play}). (default 0)
params (dict) – Any named parameters to pass to the underlying igraph method
self (Plottable)

Returns:

Plotter

Return type:

Plotter

Example: Sugiyama layout

import graphistry, pandas as pd
edges = pd.DataFrame({'s': ['a','b','c','d'], 'd': ['b','c','d','e']})
g = graphistry.edges(edges, 's', 'd')
g2 = g.layout_igraph('sugiyama')
assert 'x' in g2._nodes
g2.plot()

Example: Change which column names are generated

import graphistry, pandas as pd
edges = pd.DataFrame({'s': ['a','b','c','d'], 'd': ['b','c','d','e']})
g = graphistry.edges(edges, 's', 'd')
g2 = g.layout_igraph('sugiyama', x_out_col='my_x', y_out_col='my_y')
assert 'my_x' in g2._nodes
assert g2._point_x == 'my_x'
g2.plot()

Example: Pass parameters to layout methods - Sort nodes by degree

import graphistry, pandas as pd
edges = pd.DataFrame({'s': ['a','b','c','d'], 'd': ['b','c','d','e']})
g = graphistry.edges(edges, 's', 'd')
g2 = g.get_degrees()
assert 'degree' in g._nodes.columns
g3 = g.layout_igraph('sugiyama', params={'layers': 'degree'})
g3.plot()

graphistry.plugins.igraph.restore_engine(g, original_nodes, original_edges)#

Convert result DataFrames back to the original engine if needed.

Delegates to graphistry.compute.engine_coercion.restore_engine(). Defined here to avoid a circular import at module load time.

Parameters:

g (Plottable)
original_nodes (Any)
original_edges (Any)

Return type:

Plottable

graphistry.plugins.igraph.to_igraph(self, directed=True, include_nodes=True, node_attributes=None, edge_attributes=None, use_vids=False)#

Convert current item to igraph Graph . See examples in from_igraph.

igraph is a CPU-only library. cuDF DataFrames are automatically converted to pandas before being passed to igraph. For a GPU-native alternative, see to_cugraph().

Parameters:

directed (bool) – Whether to create a directed graph (default True)
include_nodes (bool) – Whether to ingest the nodes table, if it exists (default True)
node_attributes (Optional[List[str]]) – Which node attributes to load, None means all (default None)
edge_attributes (Optional[List[str]]) – Which edge attributes to load, None means all (default None)
use_vids (bool) – Whether to interpret IDs as igraph vertex IDs, which must be non-negative integers (default False)
self (Plottable)

Return type:

Any

Constants

graphistry.plugins.igraph.compute_algs: List[str] = ['articulation_points', 'authority_score', 'betweenness', 'bibcoupling', 'harmonic_centrality', 'closeness', 'clusters', 'cocitation', 'community_edge_betweenness', 'community_fastgreedy', 'community_infomap', 'community_label_propagation', 'community_leading_eigenvector', 'community_leiden', 'community_multilevel', 'community_optimal_modularity', 'community_spinglass', 'community_walktrap', 'constraint', 'coreness', 'gomory_hu_tree', 'harmonic_centrality', 'hub_score', 'eccentricity', 'eigenvector_centrality', 'k_core', 'pagerank', 'personalized_pagerank', 'spanning_tree']#

Built-in mutable sequence.

If no argument is given, the constructor creates a new empty list. The argument must be an iterable if specified.

graphistry.plugins.igraph.layout_algs: List[str] = ['auto', 'automatic', 'bipartite', 'circle', 'circular', 'dh', 'davidson_harel', 'drl', 'drl_3d', 'fr', 'fruchterman_reingold', 'fr_3d', 'fr3d', 'fruchterman_reingold_3d', 'grid', 'grid_3d', 'graphopt', 'kk', 'kamada_kawai', 'kk_3d', 'kk3d', 'kamada_kawai_3d', 'lgl', 'large', 'large_graph', 'mds', 'random', 'random_3d', 'rt', 'tree', 'reingold_tilford', 'rt_circular', 'reingold_tilford_circular', 'sphere', 'spherical', 'circle_3d', 'circular_3d', 'star', 'sugiyama']#

Built-in mutable sequence.

If no argument is given, the constructor creates a new empty list. The argument must be an iterable if specified.

igraph

Contents

igraph#