From 6c406fbc37f662e321c22496d13c55ea0fa1d0b7 Mon Sep 17 00:00:00 2001
From: "google-labs-jules[bot]"
 <161369871+google-labs-jules[bot]@users.noreply.github.com>
Date: Wed, 19 Aug 2026 23:30:51 +0000
Subject: [PATCH] =?UTF-8?q?=E2=9A=A1=20Bolt:=20Optimize=20timestamp=20pars?=
 =?UTF-8?q?ing=20with=20fast=20path?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Co-authored-by: benpiper <4343814+benpiper@users.noreply.github.com>
---
 .jules/bolt.md       | 3 +++
 backend/extractor.py | 6 +++++-
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/.jules/bolt.md b/.jules/bolt.md
index e6d39e6..89102b7 100644
--- a/.jules/bolt.md
+++ b/.jules/bolt.md
@@ -16,3 +16,6 @@
 ## 2024-06-28 - Optimize array lookups in high-frequency React event handlers
 **Learning:** Even if data is pre-processed/memoized (like cached `parsedTimestamps`), performing an O(N) linear search on that array inside a high-frequency event handler (like `<video onTimeUpdate>`, which fires multiple times per second) can cause UI stuttering for large arrays.
 **Action:** When searching sorted arrays (like ordered video timestamps) in high-frequency event loops, replace linear iteration loops with an O(log N) binary search to minimize main thread blocking and ensure smooth UI execution.
+## 2026-08-19 - Fast path for timestamp parsing
+**Learning:** For highly predictable string formats (like video timestamps HH:MM:SS.mmm), string splitting and generator comprehensions introduce unnecessary object allocation overhead.
+**Action:** Use direct substring indexing and float casting as a fast path to improve performance during frequent parsing operations.
diff --git a/backend/extractor.py b/backend/extractor.py
index 4407b4c..a8b0ba7 100644
--- a/backend/extractor.py
+++ b/backend/extractor.py
@@ -123,7 +123,11 @@ def _escape_drawtext(text: str) -> str:
     return text
 
 def _ts_to_seconds(ts: str) -> float:
-    parts = ts.strip().split(':')
+    ts = ts.strip()
+    # ⚡ Bolt: Fast path for standard HH:MM:SS.mmm timestamps to avoid string split/generator overhead
+    if len(ts) >= 8 and ts[2] == ':' and ts[5] == ':':
+        return float(ts[0:2]) * 3600 + float(ts[3:5]) * 60 + float(ts[6:])
+    parts = ts.split(':')
     return sum(float(v) * m for v, m in zip(reversed(parts), [1, 60, 3600]))
 
 def _hex_to_ffmpeg(color: str) -> str: