clustules

A Go library for clustering analysis based on dissimilarity (distance) matrices.

clustules provides data structures for working with dissimilarity matrices, graphs, and vertex sets, with a focus on hierarchical clustering workflows.

• vertices — labeled vertex set (generic over string | int), with bidirectional label ↔ index lookup
• diss — symmetric dissimilarity matrix (Diss[T]); supports reading .mat files in lower-triangular, upper-triangular, or square formats, with or without row labels
• graph — undirected graph (Graph[T]) built on a vertex set
• modules — module/cluster structure grouping multiple graphs (work in progress)

import "github.com/FrancoisBrucker/clustules/diss"

data, _ := os.ReadFile("distances.mat")
d, err := diss.NewFromString(string(data))

NewFromString accepts six formats, auto-detected from the token counts per line.

Lower-triangular with labels (like henley.mat):

Ours   0
Chat   47.2  0
Vache  27.7  30.9  0

Square without labels:

0    47.2  27.7
47.2 0     30.9
27.7 30.9  0

Go 1.25+

See LICENSE.