From 0c680df2f1ec8df44a1c55d45c9fb57ceb4c7a0f Mon Sep 17 00:00:00 2001
From: sjiang95 <51251025+sjiang95@users.noreply.github.com>
Date: Wed, 8 May 2024 16:17:10 +0800
Subject: [PATCH] setup_logger: add optional arg `encoding` for FileHandler
 (#3240)

* setup_logger: add optional arg encoding for FileHandler

By default, the encoding is `utf-8` for compatibility with CJK characters.

Signed-off-by: Shengjiang Quan <qsj287068067@126.com>

* Added a test and updated docstring

* open() the test log file with corresponding encoding

* add Japanese and Korean test words

* bypass encoding = None check on Windows

* Updated tests

---------

Signed-off-by: Shengjiang Quan <qsj287068067@126.com>
Co-authored-by: vfdev <vfdev.5@gmail.com>
---
 ignite/utils.py            |  7 ++++++-
 tests/ignite/test_utils.py | 24 ++++++++++++++++++++++++
 2 files changed, 30 insertions(+), 1 deletion(-)

diff --git a/ignite/utils.py b/ignite/utils.py
index 078e16663f6..6e5b2176d6a 100644
--- a/ignite/utils.py
+++ b/ignite/utils.py
@@ -163,6 +163,7 @@ def setup_logger(
     filepath: Optional[str] = None,
     distributed_rank: Optional[int] = None,
     reset: bool = False,
+    encoding: Optional[str] = "utf-8",
 ) -> logging.Logger:
     """Setups logger: name, level, format etc.
 
@@ -175,6 +176,7 @@ def setup_logger(
         distributed_rank: Optional, rank in distributed configuration to avoid logger setup for workers.
             If None, distributed_rank is initialized to the rank of process.
         reset: if True, reset an existing logger rather than keep format, handlers, and level.
+        encoding: open the file with the encoding. By default, 'utf-8'.
 
     Returns:
         logging.Logger
@@ -228,6 +230,9 @@ def setup_logger(
 
     .. versionchanged:: 0.4.5
         Added ``reset`` parameter.
+
+    .. versionchanged:: 0.5.1
+        Argument ``encoding`` added to correctly handle special characters in the file, default "utf-8".
     """
     # check if the logger already exists
     existing = name is None or name in logging.root.manager.loggerDict
@@ -265,7 +270,7 @@ def setup_logger(
         logger.addHandler(ch)
 
         if filepath is not None:
-            fh = logging.FileHandler(filepath)
+            fh = logging.FileHandler(filepath, encoding=encoding)
             fh.setLevel(level)
             fh.setFormatter(formatter)
             logger.addHandler(fh)
diff --git a/tests/ignite/test_utils.py b/tests/ignite/test_utils.py
index c4c65a29d69..828533ce201 100644
--- a/tests/ignite/test_utils.py
+++ b/tests/ignite/test_utils.py
@@ -1,4 +1,5 @@
 import logging
+import platform
 import sys
 from collections import namedtuple
 
@@ -174,6 +175,29 @@ def test_override_setup_logger(capsys):
     logging.shutdown()
 
 
+@pytest.mark.parametrize("encoding", [None, "utf-8"])
+def test_setup_logger_encoding(encoding, dirname):
+    fp = dirname / "log.txt"
+    logger = setup_logger(name="logger", filepath=fp, encoding=encoding, reset=True)
+    test_words = ["say hello", "say 你好", "say こんにちわ", "say 안녕하세요", "say привет"]
+    for w in test_words:
+        logger.info(w)
+    logging.shutdown()
+
+    with open(fp, "r", encoding=encoding) as h:
+        data = h.readlines()
+
+    if platform.system() == "Windows" and encoding is None:
+        flatten_data = "\n".join(data)
+        assert test_words[0] in flatten_data
+        for word in test_words[1:]:
+            assert word not in flatten_data
+    else:
+        assert len(data) == len(test_words)
+        for expected, output in zip(test_words, data):
+            assert expected in output
+
+
 def test_deprecated():
     # Test on function without docs, @deprecated without reasons
     @deprecated("0.4.2", "0.6.0")