🧪 Benchmark: bigtree vs anytree vs treelib

[kayjan]

I'm excited to release a new benchmark comparing bigtree, anytree, and treelib performance.

🔍 Benchmark details

• The benchmark focuses on construction from a Python dictionary, the only common input method supported across all three libraries
• To ensure fairness, identical nested dict dataset and envrionment were used for all tests
• bigtree has built-in checks which were turned off during benchmarking. anytree has checks as well, but they are turned off by default

📊 Results summary

• bigtree demonstrated faster construction time compared to anytree, and was significantly faster than treelib
• Complete results, python package details, test code are available in the documentation

🌳 Beyond the benchmark

While this benchmark covers only nested dictionary construction, bigtree supports a broader feature set beyond what is comparable:

• Construct trees from lists, flat dictionary, and pandas/polars DataFrames
• Export trees to string/stdout, dictionary, pandas/polars DataFrame, and even images
• Search, query, traversal, modification, plotting methods
• Other utility functions e.g., clone tree, get subtree, prune tree, get tree diff

🚀 Why this matters

• Less overhead: Faster construction when your tree source is a dict (or easily convertible to one)
• Common use case: Many workflows start from nested dictionaries (e.g., JSON), so this benchmark gives actionable insight into library choice
• Transparent benchmark methodology: Code is provided so you can extend the tests for your own use case

Do share feedback or results (from your own data) on GitHub Discussions or issues to help further refine real-world scenarios. Thank you for your continued support and suggestions to make bigtree better 😄