airbyte-source-s3 4.14.4__tar.gz → 4.14.5__tar.gz

{airbyte_source_s3-4.14.4 → airbyte_source_s3-4.14.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: airbyte-source-s3
-Version: 4.14.4
+Version: 4.14.5
 Summary: Source implementation for S3.
 Home-page: https://airbyte.com
 License: ELv2
@@ -15,7 +15,7 @@ Requires-Dist: airbyte-cdk[file-based] (>=7.0.4,<8.0.0)
 Requires-Dist: dill (>=0.3.4,<0.4.0)
 Requires-Dist: pendulum (>=3.0.0,<4.0.0)
 Requires-Dist: pytz (>=2024.2,<2025.0)
-Requires-Dist: smart-open[s3] (==4.14.4)
+Requires-Dist: smart-open[s3] (==4.14.5)
 Requires-Dist: transformers (>=4.38.2,<5.0.0)
 Requires-Dist: urllib3 (<2)
 Requires-Dist: wcmatch (>=10.0,<11.0)

{airbyte_source_s3-4.14.4 → airbyte_source_s3-4.14.5}/pyproject.toml

@@ -3,7 +3,7 @@ requires = [ "poetry-core>=1.0.0",]
 build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
-version = "4.14.4"
+version = "4.14.5"
 name = "airbyte-source-s3"
 description = "Source implementation for S3."
 authors = [ "Airbyte <contact@airbyte.io>",]
@@ -30,7 +30,7 @@ source-s3 = "source_s3.run:run"
 
 [tool.poetry.dependencies.smart-open]
 extras = [ "s3",]
-version = "4.14.4"
+version = "4.14.5"
 
 [tool.poetry.group.dev.dependencies]
 pytest = "^8.0.0"

{airbyte_source_s3-4.14.4 → airbyte_source_s3-4.14.5}/source_s3/v4/zip_reader.py

@@ -366,7 +366,21 @@ class ZipContentReader:
         data = self.buffer[:size]
         self.buffer = self.buffer[size:]
 
-        return data.decode(self.encoding) if self.encoding else bytes(data)
+        try:
+            return data.decode(self.encoding) if self.encoding else bytes(data)
+        except UnicodeDecodeError:
+            if self.encoding == "utf_8_sig":
+                # utf_8_sig considers `\xef\xbb\xbf` as a single character and therefore calling `bytearray(b'\xef').decode("utf_8_sig") will
+                # cause an exception to be raised.
+                number_of_bytes_to_add = size - 1
+                if data.endswith(bytearray(b"\xef")):
+                    number_of_bytes_to_add += 2
+                elif data.endswith(bytearray(b"\xbb")):
+                    number_of_bytes_to_add += 1
+                data = data + self.buffer[:number_of_bytes_to_add]
+                self.buffer = self.buffer[number_of_bytes_to_add:]
+                return data.decode(self.encoding) if self.encoding else bytes(data)
+            raise
 
     def seek(self, offset: int, whence: int = io.SEEK_SET) -> int:
         """