Fix cross-request data leakage from base64 image cache collision

vllm_mlx/models/mllm.py

@@ -465,8 +465,10 @@ def save_base64_image(base64_string: str) -> str:
     """Save base64 image to temp file and return path. Caches identical images."""
     import hashlib
 
-    # Hash the full base64 string to prevent collisions between images
-    # with identical headers (e.g. JPEG images sharing first 1000 chars)
+    # Hash the FULL base64 string — not just a prefix.
+    # Using only the first 1000 chars caused cache collisions between
+    # different images with identical JPEG headers (e.g. invoices from
+    # the same PDF renderer), returning a previous request's image.
     image_hash = hashlib.sha256(base64_string.encode()).hexdigest()
 
     # Return cached path if available and file still exists

vllm_mlx/vision_embedding_cache.py

@@ -106,9 +106,10 @@ def compute_image_hash(image_path: str) -> str:
     try:
         path = Path(image_path)
         if path.exists() and path.is_file():
-            # Hash file content (first 64KB for speed)
+            # Hash full file content (not truncated — truncation can
+            # cause collisions for images with identical headers)
             with open(path, "rb") as f:
-                content = f.read(65536)
+                content = f.read()
             return hashlib.sha256(content).hexdigest()[:16]
         else:
             # Hash the string (URL or base64)