# Changelog

## Score

- CAI 59 → 58 (-0.9)

## Lenses

- Code Health 73 → 73 (+0.0)
- Architecture 92 → 91 (-1.4)
- Maturity 63 → 63 (+0.0)
- Readiness 45 → 45 (+0.0)
- Security 82 → 72 (-9.9)
- Performance 85 → 85 (+0.0)

## Resolved (3)

- Ambiguous return type and naming. `Pretokenize.pretokens()` returns a `FastR50kPretokenizer` but the method name `pretokens` (plural) suggests it might return a collection or iterator, or it's a typo for `pretokenizer`. Meanwhile, `pretokenize.pretokenize_as_iter` returns a `FastR50kPretokenizer` (not an iterator) despite the name `as_iter`. This is a significant naming contradiction: `as_iter` implies an iterator is returned, but it returns a tokenizer struct. Additionally, `Pretokenize.pretokens()` is likely a typo for `pretokenizer` or `get_pretokenizer`.
- Naming inconsistency for singular vs plural inputs. One method takes a slice of strings (patterns) and uses the plural 'regexes', while the other takes a single string (pattern) and uses the singular 'regex'. While functionally distinct, the naming convention is slightly inconsistent with common Rust idioms where `from_*` often implies a single source, or `with_*` is used for configuration. However, the bigger issue is that `from_split_regexes` implies multiple patterns but takes a slice, whereas `from_split_regex` takes a single string. It is more consistent to name them `from_split_regex` (single) and `from_split_regexes` (multiple) OR `with_split_regex`/`with_split_regexes`. The current naming is acceptable but `from_split_regexes` for a slice is slightly less idiomatic than `from_split_regex` for a single item if the intent is construction. A more glaring inconsistency is the lack of a `from_split_regex` for a single string if `from_split_regexes` exists for a slice, or vice versa. Actually, looking closer, `from_split_regex` takes `str` and `from_split_regexes` takes `&[&str]`. This is consistent in intent (singular vs plural). No real inconsistency here.
- Outdated: pyo3

## New (19)

- Critical CVE: [GHSA redacted] (uv.lock)
- Critical vulnerability: [GHSA redacted] (uv.lock)
- High CVE: [GHSA redacted] (uv.lock)
- Medium CVE: RUSTSEC-2026-0381 (Cargo.lock)
- Medium CVE: RUSTSEC-2026-0382 (Cargo.lock)
- Medium advisory (unsound): RUSTSEC-2026-0395 (Cargo.lock)
- Medium advisory (unsound): RUSTSEC-2026-0396 (Cargo.lock)
- Medium advisory (unsound): RUSTSEC-2026-0397 (Cargo.lock)
- Medium advisory (unsound): RUSTSEC-2026-0399 (Cargo.lock)
- Medium advisory (unsound): RUSTSEC-2026-0400 (Cargo.lock)
- Medium advisory (unsound): RUSTSEC-2026-0401 (Cargo.lock)
- Medium advisory (unsound): RUSTSEC-2026-0402 (Cargo.lock)
- Medium advisory (unsound): RUSTSEC-2026-0404 (Cargo.lock)
- Medium advisory (unsound): RUSTSEC-2026-0405 (Cargo.lock)
- Medium advisory (unsound): RUSTSEC-2026-0406 (Cargo.lock)
- Medium vulnerability: RUSTSEC-2026-0391 (Cargo.lock)
- Medium vulnerability: RUSTSEC-2026-0398 (Cargo.lock)
- Medium vulnerability: RUSTSEC-2026-0410 (Cargo.lock)
- Outdated: memmap2

## Architecture

- Unchanged — 0 containers · 1 contexts · 0 edges
