tralda

A Python library for tree algorithms and data structures.

Installation and usage

The package requires Python 3.10 or higher. The tralda package is available on PyPI:

pip install tralda

See also the page Installation in the documentation.

The tralda documentation contains a user guide with example code as well as the API reference.

If you want to contribute to the project, please see the contributing guidelines. Please report any bugs and questions in the Issues section.

Overview

tralda provides a collection of efficient algorithms and data structures centered around rooted trees, with a focus on phylogenetics and combinatorial algorithms. It is designed to serve as a building block for research software in computational biology and related fields.

Tree data structure — The Tree and TreeNode classes in tralda.datastructures offer a flexible rooted-tree representation with support for traversals (preorder, postorder, level-order), subtree operations, tree generation, and utilities such as topology comparison and hierarchy extraction.

Lowest common ancestor (LCA) — An $O(n)$-time/space preprocessing structure for $O(1)$ LCA queries, based on the algorithm by Bender et al. (2005). Available as LCA in tralda.datastructures.

Supertree computation — The tralda.supertree subpackage implements several algorithms for constructing supertrees and consensus trees from a set of (partial) input trees:

• BUILD (Aho et al. 1981) — classic triple-based supertree construction.
• BuildST (Deng & Fernández-Baca 2016) — fast compatibility testing and supertree construction for rooted phylogenetic trees.
• LinCR (Schaller et al. 2021) — linear-time algorithm for the minimal common refinement of rooted phylogenetic trees on a common leaf set.
• Loose consensus tree (Jansson et al. 2016) — linear-time construction of the loose consensus tree for trees with the same leaf set.

Balanced binary search trees — tralda.datastructures.bst provides AVL trees and red-black trees (ordered sets and dictionaries) with $O(\log n)$ insertion, deletion, and lookup, as well as efficient split and join operations.

Dynamic graph connectivity — The HDTGraph class in tralda.datastructures.hdtgraph implements the poly-logarithmic dynamic graph structure described by Holm et al. (2001), supporting edge insertions and deletions with $O(\log^2 n)$ amortized cost while answering connectivity queries in $O(\log n)$.

Cograph detection and editing — tralda.cograph offers linear-time cograph recognition (Corneil et al. 1985) with cotree construction, as well as a heuristic for cograph editing (Crespelle 2021) that modifies a graph with a near-minimum number of edge insertions/deletions to make it a cograph.

Visualization

tralda.visualization renders Tree objects as static (matplotlib) or interactive (Plotly) figures, with horizontal, vertical, and circular layouts and a flexible per-node styling API. Optional extra: pip install tralda[viz].

from tralda.datastructures import LCA, Tree
from tralda.visualization import plot_tree
from tralda.visualization.style import NodeStyle, TreeStyle

T = Tree.parse_newick(
    "((((Human:6,Chimpanzee:6)HC:2,Gorilla:8)HCG:9,Orangutan:17)Hominidae:3,"
    "(Gibbon:17,Siamang:17)Hylobatidae:3)Hominoidea;"
)

# Compute and highlight the last common ancestor of Human and Gorilla.
lca = LCA(T)
lca_node = lca("Human", "Gorilla")

tree_style = TreeStyle(
    leaf_symbol="circle",
    node_overrides={lca_node: NodeStyle(symbol="star", symbol_size=13, symbol_color="gold")},
)

fig, ax = plot_tree(T, tree_style=tree_style)

See the visualization user guide for the full styling API and more examples.

Citation and references

If you use tralda in your project or code from it, please consider citing:

Schaller, D., Hellmuth, M., Stadler, P.F. (2021) A simpler linear-time algorithm for the common refinement of rooted phylogenetic trees on a common leaf set. Algorithms for Molecular Biology 16, 23 (2021). https://doi.org/10.1186/s13015-021-00202-8

Additional references to algorithms that were implemented are given in the source code.