grakel.graph_from_pandas

grakel.graph_from_pandas(edge_df, node_df=None, directed=False, as_Graph=False)[source][source]

Produces a collection of Graph Objects from pandas dataframes.

A function for helping a user that has a bunch of graph csv files make a collection for grakel input.

Parameters
edge_dftuple(pandas.DataFrame, pct, tuple(pct, pct), pct or None, pct or list(pct) or None)

A tuple of many elements:

  1. A pandas dataframe containing all the edges of a collection of graphs

  2. The column name of the graph index

  3. A tuple containing source and destination column names

  4. The column name of weights column (None if non-existent)

  5. The column name pointing the column containing edge-labels or a list pointing the columns of attributes (None if edge-labels are non-existent).

If node_df exists, correspondance of graph indexes and node indexes must be exact.

node_dftuple(pandas.DataFrame, pct, pct or list(pct) or None) or None, default=None

A tuple of many elements:

  1. A pandas dataframe containing all the nodes of a collection of graphs

  2. The column name of the graph index

  3. The column name pointing the column containing node-labels or a list pointing the columns of attributes (None if node-labels are non-existent).

Node id must correspond to node number.

directedbool, default=False

A variable indicating with True if both directions of the graph have been inserted inside the dataframe. Otherwise the graph is considered as undirected.

as_Graphbool, default=False

Return each output as a grakel.Graph object.

Returns
grakel_graphsdict

Dictionary with keys te keys of the graphs inside the pandas dataframe and values the corresponding graph object