mosaic network

Analyses the citation graph of your cache: the most connected papers, clusters of papers that cite each other, and exports for Gephi, D3, Graphviz or Mermaid.

Usage

mosaic network --top 10                                  # the most connected papers
mosaic network --query "hall thruster" --depth 2         # the neighbourhood of matching papers
mosaic network --query "thruster" --cluster --top 3      # clusters, and the hubs of each
mosaic network --min-connections 3 -o graph.json         # node-link JSON (D3, Gephi)
mosaic network --query "plume" -o plume.gv               # Graphviz
mosaic network --query "plume" -o plume.md               # a Mermaid diagram

Behaviour

The edges come from mosaic index --enrich-citations; without them the graph is empty.

--query seeds the graph with the cached papers matching it and follows citations --depth hops (default 2); with no query the whole graph is used. --min-connections drops papers with fewer edges.

--cluster uses Louvain community detection when networkx is installed (the analysis extra) and connected components otherwise.

--output writes .json (node-link), .gv/.dot (Graphviz) or .md/.markdown (Mermaid); the format follows the extension.

See also: tutorial 10, citation network, citation graph.

Options

mosaic network [OPTIONS]

Option	Value	Default	Description
--query, -q	TEXT		Seed graph from cached papers matching this query
--depth	INTEGER	2	Citation hops to follow from seed papers
--min-connections	INTEGER	1	Exclude papers with fewer edges than this
--cluster			Group papers into topic clusters
--output, -o	PATH		Write graph to file (.json, .gv, .md)
--top	INTEGER	5	Most-connected papers to show per cluster (terminal)