From 2269abfb868796622a185334f1dfbec66039af25 Mon Sep 17 00:00:00 2001
From: "google-labs-jules[bot]"
 <161369871+google-labs-jules[bot]@users.noreply.github.com>
Date: Sat, 22 Aug 2026 17:16:42 +0000
Subject: [PATCH] =?UTF-8?q?=E2=9A=A1=20Bolt:=20Memoize=20clinical=20contex?=
 =?UTF-8?q?t=20lexicon=20compilation?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Co-authored-by: zrt219 <199104500+zrt219@users.noreply.github.com>
---
 .jules/bolt.md                      | 3 +++
 openmed/openmed/clinical/context.py | 6 +++++-
 2 files changed, 8 insertions(+), 1 deletion(-)
 create mode 100644 .jules/bolt.md

diff --git a/.jules/bolt.md b/.jules/bolt.md
new file mode 100644
index 0000000..606c80b
--- /dev/null
+++ b/.jules/bolt.md
@@ -0,0 +1,3 @@
+## 2024-05-24 - Memoizing NLP Regex Compilation
+**Learning:** NLP and clinical text evaluation repetitively instantiates the same compiled regexes and lexicons if missing memoization, leading to heavy redundant object compilation overhead.
+**Action:** Always memoize deterministic lexicon object/regex instantiations (e.g. using @functools.lru_cache) and wrap return types in immutables to avoid performance bottlenecks in loops/text processing pipelines.
diff --git a/openmed/openmed/clinical/context.py b/openmed/openmed/clinical/context.py
index 9fd11df..b761125 100644
--- a/openmed/openmed/clinical/context.py
+++ b/openmed/openmed/clinical/context.py
@@ -35,7 +35,9 @@
 
 from __future__ import annotations
 
+import functools
 import re
+import types
 from collections.abc import Iterable, Iterator, Mapping, Sequence
 from dataclasses import dataclass, replace
 from datetime import date
@@ -154,7 +156,9 @@ class _CompiledContextLexicon:
     backward_context_cues: frozenset[str]
 
 
+@functools.lru_cache(maxsize=32)
 def _compiled_context_lexicon(language: str | None = None) -> _CompiledContextLexicon:
+    """Returns a memoized compiled lexicon to avoid repeated regex compilation."""
     lexicon = get_clinical_cue_lexicon(language)
     token_boundaries = lexicon.token_boundaries
     return _CompiledContextLexicon(
@@ -195,7 +199,7 @@ def _compiled_context_lexicon(language: str | None = None) -> _CompiledContextLe
             ),
             token_boundaries=token_boundaries,
         ),
-        category_by_text=_cue_category_lookup(lexicon),
+        category_by_text=types.MappingProxyType(_cue_category_lookup(lexicon)),
         backward_context_cues=frozenset(
             _normalize_cue_text(cue) for cue in lexicon.backward
         ),