NetworkX + igraph

What it is

NetworkX is the Pythonic way to model and analyze graphs in memory. igraph is the C-fast alternative for big graphs. Both ship every algorithm you'll ever need.

How Vaaani uses it

• Building and pruning the knowledge graph before storing in Neo4j
• PageRank to score entity importance for retrieval
• Community detection to cluster topics in a corpus
• Shortest-path computation for explainable answers

Why it makes the cut

Sometimes you don't need a database — you need a script. NetworkX runs analytics on millions of edges in seconds.

Sample code

import networkx as nx

G = nx.DiGraph()
G.add_edges_from(triples)

scores = nx.pagerank(G)
top = sorted(scores.items(), key=lambda x: -x[1])[:10]