sdk/python: object file max_resume per object file

Signed-off-by: Ryan Koo <[email protected]>

python/aistore/sdk/object_file.py

@@ -2,11 +2,11 @@
 # Copyright (c) 2024, NVIDIA CORPORATION. All rights reserved.
 #
 
-# pylint: disable=too-many-branches
-
 from io import BufferedIOBase
 from typing import Iterator
+
 import requests
+
 from aistore.sdk.object_reader import ObjectReader
 from aistore.sdk.utils import get_logger
 
@@ -84,17 +84,16 @@ def empty(self):
 
 class ObjectFile(BufferedIOBase):
     """
-    A file-like object for reading object data, with support for both reading a fixed size
-    of data and reading until the end of the stream (EOF). It provides the ability to resume
-    and continue reading from the last known position in the event of a streaming error, such
-    as a `ChunkedEncodingError`.
-
-    Data is fetched in chunks via the object reader and temporarily stored in an internal buffer.
-    The buffer is filled either to the required size or until EOF is reached. If a streaming
-    error occurs during this process, the object file automatically attempts to resume the buffer
-    filling process from the last known chunk position, up to a configurable number of resume
-    attempts (`max_resume`), raising a `ChunkedEncodingError` if the maximum number of attempts
-    is exceeded.
+    A file-like object for reading object data, with support for both reading a fixed size of data
+    and reading until the end of the stream (EOF). It provides the ability to resume and continue
+    reading from the last known position in the event of a ChunkedEncodingError.
+
+    Data is fetched in chunks via the object reader iterator and temporarily stored in an internal
+    buffer. The buffer is filled either to the required size or until EOF is reached. If a
+    `ChunkedEncodingError` occurs during this process, ObjectFile catches and automatically attempts
+    to resume the buffer filling process from the last known chunk position. The number of resume
+    attempts is tracked across the entire object file, and if the total number of attempts exceeds
+    the configurable `max_resume`, a `ChunkedEncodingError` is raised.
 
     Once the buffer is adequately filled, the `read()` method reads and returns the requested amount
     of data from the buffer.
@@ -111,6 +110,7 @@ def __init__(self, object_reader: ObjectReader, max_resume: int):
         self._closed = False
         self._buffer = SimpleBuffer()
         self._chunk_iterator = self._object_reader.iter_from_position(self._current_pos)
+        self._resume_total = 0
 
     def close(self) -> None:
         """
@@ -178,23 +178,21 @@ def read(self, size=-1):
         if size == 0:
             return b""
 
-        resume_attempts = 0
-
         while True:
             try:
                 # Fill the buffer with the requested size or to EOF
                 self._buffer.fill(self._chunk_iterator, size)
                 # If successfully filled the buffer, exit
                 break
             except requests.exceptions.ChunkedEncodingError as err:
-                resume_attempts += 1
-                if resume_attempts > self._max_resume:
+                self._resume_total += 1
+                if self._resume_total > self._max_resume:
                     logger.error("Max retries reached. Cannot resume read.")
                     raise err
                 logger.warning(
                     "Chunked encoding error (%s), retrying %d/%d",
                     err,
-                    resume_attempts,
+                    self._resume_total,
                     self._max_resume,
                 )