Performance Improvements

[LZL0]

Summary by cubic

Improves runtime performance by removing per-call regex compilation in drift detection and switching to a slot-backed State with an O(1) token buffer for content assembly. This reduces CPU time and string allocations during streaming.

• Performance
  • Pre-compiled regex patterns across drift.py (tone, sentence split, markdown, hedging); removed repeated lowercasing via case-insensitive matches.
  • Converted State to a slot-backed class with O(1) buffered appends (append_content) and lazy flush via the content property.
  • Switched callback IDs to UUIDv7 for faster, time-ordered guardrail rule traces.

^{Written for commit 8df4f77. Summary will update automatically on new commits.}

[cubic-dev-ai]

1 issue found across 3 files

Prompt for AI agents (all 1 issues)

Check if these issues are valid — if so, understand the root cause of each and fix them.


<file name="src/l0/types.py">

<violation number="1" location="src/l0/types.py:271">
P1: Breaking API change: `content` is no longer a constructor parameter. Existing code using `State(content=&quot;...&quot;)` will fail. Consider using `__post_init__` to accept `content` as an init parameter while still using the internal buffer, or add `content: InitVar[str] = &quot;&quot;` to preserve backward compatibility.</violation>
</file>

_{Reply to cubic to teach it or ask questions. Re-run a review with @cubic-dev-ai review this PR}

[cubic-dev-ai]

P1: Breaking API change: content is no longer a constructor parameter. Existing code using State(content="...") will fail. Consider using __post_init__ to accept content as an init parameter while still using the internal buffer, or add content: InitVar[str] = "" to preserve backward compatibility.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At src/l0/types.py, line 271:

<comment>Breaking API change: `content` is no longer a constructor parameter. Existing code using `State(content=&quot;...&quot;)` will fail. Consider using `__post_init__` to accept `content` as an init parameter while still using the internal buffer, or add `content: InitVar[str] = &quot;&quot;` to preserve backward compatibility.</comment>

<file context>
@@ -262,9 +262,14 @@ class BackoffStrategy(str, Enum):
+    Access `content` property to get the accumulated string (flushes buffer automatically).
+    &quot;&quot;&quot;
+
+    _content: str = &quot;&quot;
+    _content_buffer: list[str] = field(default_factory=list)
     checkpoint: str = &quot;&quot;  # Last known good slice for continuation
</file context>