Formula decoder uses Hugging Face tokenizers 0.22.2 (Apache-2.0), pinned to the legacy Python tokenizer version.
Source: https://github.com/huggingface/tokenizers/tree/v0.22.2/tokenizers
Full license: TOKENIZERS.txt.

formula_html_entities.txt contains sorted html.entities.html5 keys from Python 3.14.5.
Generator: scripts/migration/generate_formula_entities.py.
Source: https://github.com/python/cpython/blob/v3.14.5/Lib/html/entities.py
Python PSF license: PYTHON-UNICODE.txt (also used for the existing Python Unicode tables).
These names only gate unsupported ftfy inputs; they do not implement HTML or ftfy decoding.
