From 4b0829fe433f37672d944db44a56407bfe48d481 Mon Sep 17 00:00:00 2001
From: dpiet-oai <326640584+dpiet-oai@users.noreply.github.com>
Date: Fri, 16 Oct 2026 15:46:46 +0000
Subject: [PATCH] fix(live): avoid contextual transcript casing scans

Castiron-Internal-PR: https://github.com/openai/openai-java-internal/pull/281
Castiron-Source-SHA: e8eed1e3442f6a92193179487d796912242cae6d
Castiron-Public-Base-SHA: 848ef53c94498e795587d571900202ea383601d4
---
 .castiron.stats.yml                             |  6 +++---
 .../com/openai/helpers/LiveTranscriptGrouper.kt | 12 ++++++++----
 .../openai/helpers/LiveTranscriptGrouperTest.kt | 17 +++++++++++++++++
 3 files changed, 28 insertions(+), 7 deletions(-)

diff --git a/.castiron.stats.yml b/.castiron.stats.yml
index d04f294a7..9efd98756 100644
--- a/.castiron.stats.yml
+++ b/.castiron.stats.yml
@@ -1,8 +1,8 @@
 schema_version: 1
-generation_id: b935ba17-9980-45bd-b8c7-1939d1b03aba
+generation_id: 0cdeb809-627b-4caa-9340-887e46d7979f
 openapi_spec_hash: e81fa0fd2133715dabef9f00a42557a9
 openapi_transformed_spec_hash: bea7a30d23deed3fc8f1c6413a38d354
 config_hash: dda57991526f80e9f99d8f9621460eaa
-codegen_sha: 9306aff2af8ad922f591ebbc1d29362cf9271bc8
+codegen_sha: ca58cce22b36f9f5586a16379bc977d20f193a4a
 codegen_hash: bda6b5fe61d5a094103b7cbc38597c82109f4c14bb23d4501d6c3e1569f75878
-public_codegen_sha: 43d374297ff620990fa45fdd39a32079c3905e30
+public_codegen_sha: 3a713a4b40be032d22b5aa19b393ee32178ea142
diff --git a/openai-java-core/src/main/kotlin/com/openai/helpers/LiveTranscriptGrouper.kt b/openai-java-core/src/main/kotlin/com/openai/helpers/LiveTranscriptGrouper.kt
index b7eb92bf8..b4c3d9b3e 100644
--- a/openai-java-core/src/main/kotlin/com/openai/helpers/LiveTranscriptGrouper.kt
+++ b/openai-java-core/src/main/kotlin/com/openai/helpers/LiveTranscriptGrouper.kt
@@ -191,9 +191,13 @@ class LiveTranscriptGrouper private constructor(builder: Builder) : AutoCloseabl
             )
 
         private fun normalize(value: String): String =
-            whitespace
-                .replace(value.lowercase(Locale.ROOT).replace('ς', 'σ').replace('-', ' '), " ")
-                .trim { it == ' ' || it in punctuation }
+            whitespace.replace(foldCase(value).replace('-', ' '), " ").trim {
+                it == ' ' || it in punctuation
+            }
+
+        // Sigma forms are equivalent here; fold before ROOT casing to avoid contextual scans.
+        private fun foldCase(value: String): String =
+            value.replace('\u03A3', '\u03C3').lowercase(Locale.ROOT).replace('ς', 'σ')
     }
 
     private class AckNode {
@@ -221,7 +225,7 @@ class LiveTranscriptGrouper private constructor(builder: Builder) : AutoCloseabl
             highSurrogate = text.last().takeIf(Char::isHighSurrogate)
             val end = text.length - if (highSurrogate == null) 0 else 1
             // Folding final sigma keeps ROOT lowercase independent of fragment boundaries.
-            val lower = text.substring(0, end).lowercase(Locale.ROOT).replace('ς', 'σ')
+            val lower = foldCase(text.substring(0, end))
             for (raw in lower) {
                 val char = if (raw == '-' || whitespace.matches(raw.toString())) ' ' else raw
                 val trim = char == ' ' || char in punctuation
diff --git a/openai-java-core/src/test/kotlin/com/openai/helpers/LiveTranscriptGrouperTest.kt b/openai-java-core/src/test/kotlin/com/openai/helpers/LiveTranscriptGrouperTest.kt
index 721c11dcc..dccf3ee31 100644
--- a/openai-java-core/src/test/kotlin/com/openai/helpers/LiveTranscriptGrouperTest.kt
+++ b/openai-java-core/src/test/kotlin/com/openai/helpers/LiveTranscriptGrouperTest.kt
@@ -22,6 +22,7 @@ import java.util.concurrent.TimeoutException
 import java.util.function.Consumer
 import org.assertj.core.api.Assertions.assertThat
 import org.assertj.core.api.Assertions.assertThatThrownBy
+import org.junit.jupiter.api.Assertions.assertTimeout
 import org.junit.jupiter.api.Test
 
 internal class LiveTranscriptGrouperTest {
@@ -336,6 +337,22 @@ internal class LiveTranscriptGrouperTest {
         assertThat(finalized(events).map { it.segment().text() }).containsExactly("hello")
     }
 
+    @Test
+    fun longSigmaTranscriptPreservesTextWithoutContextualCasingDelay() {
+        val text = "\u03A3".repeat(32_000)
+        val events = updates()
+        assertTimeout(Duration.ofSeconds(5)) {
+            LiveTranscriptGrouper.create { events.add(it) }
+                .use { grouper ->
+                    grouper.push(input("u1", "hello", 0, 100))
+                    grouper.push(output("a1", text, 0, 100))
+                    grouper.push(input("u2", " again", 100, 200))
+                }
+        }
+        assertThat(finalized(events).map { it.segment().text() })
+            .containsExactly("hello again", text)
+    }
+
     @Test
     fun acknowledgmentCaseMatchingDoesNotDependOnDeltaBoundaries() {
         // U+0130 expands on ROOT lowercasing; Greek sigma has a final and an ordinary form.