Feature: accept any buffer instance when processing file data (#148)

ArEnSc · ncoghlan · web-flow · commit d12e757bf356 · 2025-08-19T02:04:22.000+10:00
- Change _get_file_details() to return Buffer instead of bytes - Add special handling for memoryview inputs to avoid unnecessary conversion - Support all buffer protocol objects (bytes, bytearray, memoryview, array.array, etc.) - Update type annotations throughout to use Buffer type from typing_extensions - Improve error messages to mention "data buffer" instead of just "bytes" This change reduces memory overhead when working with buffer protocol objects by preserving memoryview objects instead of converting them to bytes unnecessarily. Closes #46 Co-authored-by: Alyssa Coghlan <ncoghlan@gmail.com>
diff --git a/src/lmstudio/history.py b/src/lmstudio/history.py
@@ -25,6 +25,8 @@
 from typing_extensions import (
     # Native in 3.11+
     Self,
+    # Native in Python 3.12+
+    Buffer,
     # Native in 3.13+
     TypeIs,
 )
@@ -529,17 +531,24 @@ def add_tool_result(self, result: ToolCallResultInput) -> ToolResultMessage:
         return message
 
 
-LocalFileInput = BinaryIO | bytes | str | os.PathLike[str]
+LocalFileInput = BinaryIO | Buffer | str | os.PathLike[str]
 
 
 # Private until file handle caching support is part of the published SDK API
 
 
-def _get_file_details(src: LocalFileInput) -> Tuple[str, bytes]:
+def _get_file_details(src: LocalFileInput) -> Tuple[str, Buffer]:
     """Read file contents as binary data and generate a suitable default name."""
-    if isinstance(src, bytes):
-        # We process bytes as raw data, not a bytes filesystem path
-        data = src
+    data: Buffer
+    if isinstance(src, Buffer):
+        if isinstance(src, memoryview):
+            # If already a memoryview, just use it directly
+            data = src
+        else:
+            # Try to create a memoryview - this will work for any buffer protocol object
+            # including bytes, bytearray, array.array, numpy arrays, etc.
+            data = memoryview(src)
+        # Received raw file data without any name information
         name = str(uuid.uuid4())
     elif hasattr(src, "read"):
         try:
@@ -550,10 +559,13 @@ def _get_file_details(src: LocalFileInput) -> Tuple[str, bytes]:
             raise LMStudioOSError(err_msg) from None
         name = getattr(src, "name", str(uuid.uuid4()))
     else:
+        # At this point, src must be a path-like object
         try:
             src_path = Path(src)
         except Exception as exc:
-            err_msg = f"Expected file-like object, filesystem path, or bytes ({exc!r})"
+            err_msg = (
+                f"Expected file-like object, filesystem path, or data buffer ({exc!r})"
+            )
             raise LMStudioValueError(err_msg) from None
         try:
             data = src_path.read_bytes()
@@ -573,7 +585,7 @@ class _LocalFileData:
     """Local file data to be added to a chat history."""
 
     name: str
-    raw_data: bytes
+    raw_data: Buffer
 
     def __init__(self, src: LocalFileInput, name: str | None = None) -> None:
         default_name, raw_data = _get_file_details(src)
