From 7bfd66564a94f029147605fbcdef56616a54c5e3 Mon Sep 17 00:00:00 2001
From: "google-labs-jules[bot]"
 <161369871+google-labs-jules[bot]@users.noreply.github.com>
Date: Tue, 31 Mar 2026 19:43:30 +0000
Subject: [PATCH] perf: optimize results filtering loops in fix_encoding.py

Co-authored-by: daggerstuff <261005129+daggerstuff@users.noreply.github.com>
---
 training/scripts/fix_encoding.py | 32 +++++++++++++++++++++-----------
 1 file changed, 21 insertions(+), 11 deletions(-)

diff --git a/training/scripts/fix_encoding.py b/training/scripts/fix_encoding.py
index 6c93f76d..12593dbb 100644
--- a/training/scripts/fix_encoding.py
+++ b/training/scripts/fix_encoding.py
@@ -30,9 +30,7 @@
     S3DatasetLoader,  # noqa: E402
 )
 
-logging.getLogger("ai.training.utils.s3_dataset_loader").setLevel(
-    logging.ERROR
-)
+logging.getLogger("ai.training.utils.s3_dataset_loader").setLevel(logging.ERROR)
 
 DEFAULT_S3_BUCKET = "pixel-data"
 MAX_RETRIES = 3
@@ -887,10 +885,16 @@ def print_results(results: list[dict[str, Any]], output: OutputHandler) -> None:
     output.header("📊 ENCODING FIX RESULTS")
     output.separator()
 
-    successful = [r for r in results if r.get("success")]
-    failed = [r for r in results if not r.get("success")]
-    skipped = [r for r in successful if r.get("skipped")]
-    fixed = [r for r in successful if not r.get("skipped")]
+    successful, failed, skipped, fixed = [], [], [], []
+    for r in results:
+        if r.get("success"):
+            successful.append(r)
+            if r.get("skipped"):
+                skipped.append(r)
+            else:
+                fixed.append(r)
+        else:
+            failed.append(r)
 
     output.info(f"\n✅ Fixed: {len(fixed)} files")
     output.info(f"⏭️  Skipped (already UTF-8): {len(skipped)} files")
@@ -919,10 +923,16 @@ def save_results(
     results: list[dict[str, Any]],
 ) -> Path:
     """Save encoding fix results to JSON file"""
-    successful = [r for r in results if r.get("success")]
-    failed = [r for r in results if not r.get("success")]
-    skipped = [r for r in successful if r.get("skipped")]
-    fixed = [r for r in successful if not r.get("skipped")]
+    successful, failed, skipped, fixed = [], [], [], []
+    for r in results:
+        if r.get("success"):
+            successful.append(r)
+            if r.get("skipped"):
+                skipped.append(r)
+            else:
+                fixed.append(r)
+        else:
+            failed.append(r)
 
     results_path = project_root / "ai/training_ready/data/encoding_fix_results.json"
     with open(results_path, "w") as f: