From ebfe88433fcfe93193d40478ee4f6e5ae24d93e0 Mon Sep 17 00:00:00 2001
From: Jan Holthuis <jan.holthuis@ruhr-uni-bochum.de>
Date: Sun, 12 Jan 2025 23:18:00 +0100
Subject: [PATCH] fix(ofx): Use compliant `UNICODE` name for UTF-8 encoding

---
 src/ofxstatement/ofx.py                   | 22 +++++++++++++++++++---
 src/ofxstatement/tests/test_ofx.py        |  4 ++--
 src/ofxstatement/tests/test_ofx_invest.py |  2 +-
 src/ofxstatement/tool.py                  |  5 +++--
 4 files changed, 25 insertions(+), 8 deletions(-)

diff --git a/src/ofxstatement/ofx.py b/src/ofxstatement/ofx.py
index 53359ab..1448d19 100644
--- a/src/ofxstatement/ofx.py
+++ b/src/ofxstatement/ofx.py
@@ -1,3 +1,4 @@
+import codecs
 from typing import Optional, Union
 from datetime import datetime, date
 from decimal import Decimal
@@ -22,20 +23,35 @@ def __init__(self, statement: Statement) -> None:
         self.default_float_precision = 2
         self.invest_transactions_float_precision = 5
 
-    def toxml(self, pretty: bool = False) -> str:
+    def toxml(self, pretty: bool = False, encoding: str = "utf-8") -> str:
         et = self.buildDocument()
         xmlstring = etree.tostring(et.getroot(), "unicode")
         if pretty:
             dom = minidom.parseString(xmlstring)
             xmlstring = dom.toprettyxml(indent="  ", newl="\r\n")
             xmlstring = xmlstring.replace('<?xml version="1.0" ?>', "").lstrip()
+
+        codec = codecs.lookup(encoding)
+        if codec.name == "utf-8":
+            encoding_name = "UNICODE"
+            charset_name = "UTF-8"
+        elif codec.name.startswith("cp"):
+            encoding_name = "USASCII"
+            charset_name = codec.name[2:]
+        else:
+            # This is non-standard, because according to the OFX spec the
+            # CHARSET should be the codepage number. We handle this gracefully,
+            # since the only alternative is throwing an error here.
+            encoding_name = "USASCII"
+            charset_name = codec.name.upper()
+
         header = (
             "OFXHEADER:100\r\n"
             "DATA:OFXSGML\r\n"
             "VERSION:102\r\n"
             "SECURITY:NONE\r\n"
-            "ENCODING:UTF-8\r\n"
-            "CHARSET:NONE\r\n"
+            f"ENCODING:{encoding_name}\r\n"
+            f"CHARSET:{charset_name}\r\n"
             "COMPRESSION:NONE\r\n"
             "OLDFILEUID:NONE\r\n"
             "NEWFILEUID:NONE\r\n"
diff --git a/src/ofxstatement/tests/test_ofx.py b/src/ofxstatement/tests/test_ofx.py
index 8be1f7a..e707fac 100644
--- a/src/ofxstatement/tests/test_ofx.py
+++ b/src/ofxstatement/tests/test_ofx.py
@@ -12,7 +12,7 @@
 DATA:OFXSGML
 VERSION:102
 SECURITY:NONE
-ENCODING:UTF-8
+ENCODING:UNICODE
 CHARSET:NONE
 COMPRESSION:NONE
 OLDFILEUID:NONE
@@ -130,7 +130,7 @@ def test_ofxWriter_pretty(self) -> None:
             "DATA:OFXSGML",
             "VERSION:102",
             "SECURITY:NONE",
-            "ENCODING:UTF-8",
+            "ENCODING:UNICODE",
             "CHARSET:NONE",
             "COMPRESSION:NONE",
             "OLDFILEUID:NONE",
diff --git a/src/ofxstatement/tests/test_ofx_invest.py b/src/ofxstatement/tests/test_ofx_invest.py
index 34013fa..1b8afb6 100644
--- a/src/ofxstatement/tests/test_ofx_invest.py
+++ b/src/ofxstatement/tests/test_ofx_invest.py
@@ -12,7 +12,7 @@
 DATA:OFXSGML
 VERSION:102
 SECURITY:NONE
-ENCODING:UTF-8
+ENCODING:UNICODE
 CHARSET:NONE
 COMPRESSION:NONE
 OLDFILEUID:NONE
diff --git a/src/ofxstatement/tool.py b/src/ofxstatement/tool.py
index 4fdfc7d..f09956d 100644
--- a/src/ofxstatement/tool.py
+++ b/src/ofxstatement/tool.py
@@ -201,9 +201,10 @@ def convert(args: argparse.Namespace) -> int:
         log.error("Statement validation error: %s" % (e.message))
         return 2  # Validation error
 
-    with smart_open(args.output, settings.get("encoding", None)) as out:
+    encoding = settings.get("encoding", "utf-8")
+    with smart_open(args.output, encoding) as out:
         writer = ofx.OfxWriter(statement)
-        out.write(writer.toxml(pretty=args.pretty))
+        out.write(writer.toxml(pretty=args.pretty, encoding=encoding))
 
     n_lines = len(statement.lines)
     log.info(