From fca1d61ae0b4646b6dee0ed077a134e2ebfe4205 Mon Sep 17 00:00:00 2001
From: Greg Brandt <brandt.greg@gmail.com>
Date: Sat, 14 Feb 2026 15:53:36 -0800
Subject: [PATCH] Bugfixes

- Fix size 0 for empty file on GCS
- Download/seek to end in append mode
- Add version variable for pydoc
---
 avrokit/__init__.py   | 4 ++++
 avrokit/url/google.py | 8 +++++---
 avrokit/url/s3.py     | 6 ++++--
 pyproject.toml        | 2 +-
 4 files changed, 14 insertions(+), 6 deletions(-)

diff --git a/avrokit/__init__.py b/avrokit/__init__.py
index 77d7356..4aefcc0 100644
--- a/avrokit/__init__.py
+++ b/avrokit/__init__.py
@@ -77,6 +77,10 @@
     ...             writer.roll()  # Create a new partition file
 """
 
+from importlib.metadata import version
+
+__version__ = version("avrokit")
+
 from .url import URL, FileURL, parse_url, create_url_mapping, flatten_urls
 from .io import (
     Appendable,
diff --git a/avrokit/url/google.py b/avrokit/url/google.py
index de2eee7..ef317fd 100644
--- a/avrokit/url/google.py
+++ b/avrokit/url/google.py
@@ -107,7 +107,7 @@ def size(self) -> int:
             if not blob.exists():
                 return 0
             blob.reload()  # N.b. loads metadata including size
-            return blob.size
+            return blob.size if blob.size is not None else 0
 
     @override
     def open(self) -> IO[Any]:
@@ -119,11 +119,13 @@ def open(self) -> IO[Any]:
         self._current_remote = blob
         self._current_local = tmpfile
         self._current_local_stream = self._current_local
-        # Download to file if r or "rb" mode
-        if "r" in self.mode:
+        # Download to file if r/rb mode, or if append mode (to preserve existing content on failure)
+        if "r" in self.mode or "a" in self.mode:
             # N.b. always writes in binary mode
             blob.download_to_file(tmpfile)
             tmpfile.seek(0)
+            if "a" in self.mode:
+                tmpfile.seek(0, 2)  # Seek to end for append mode
             if "b" not in self.mode:
                 # So if the user wants to read text, we need to decode it
                 self._current_local_stream = io.TextIOWrapper(tmpfile, encoding="utf-8")
diff --git a/avrokit/url/s3.py b/avrokit/url/s3.py
index 16cd157..2edeee4 100644
--- a/avrokit/url/s3.py
+++ b/avrokit/url/s3.py
@@ -94,11 +94,13 @@ def open(self) -> IO[Any]:
         self._current_client = client
         self._current_local = tmpfile
         self._current_local_stream = self._current_local
-        # Download to file if "r" or "rb" mode
-        if "r" in self.mode:
+        # Download to file if "r"/"rb" mode, or if append mode (to preserve existing content on failure)
+        if "r" in self.mode or "a" in self.mode:
             # N.b. always writes in binary mode
             client.download_fileobj(self.bucket, self.path, tmpfile)
             tmpfile.seek(0)
+            if "a" in self.mode:
+                tmpfile.seek(0, 2)  # Seek to end for append mode
             if "b" not in self.mode:
                 # So if the user wants to read text, we need to decode it
                 self._current_local_stream = io.TextIOWrapper(tmpfile, encoding="utf-8")
diff --git a/pyproject.toml b/pyproject.toml
index 124f850..84ce178 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -4,7 +4,7 @@
 
 [tool.poetry]
 name = "avrokit"
-version = "0.0.2"
+version = "0.0.3"
 description = "Python utilities for working with Avro data files"
 authors = ["Greg Brandt <brandt.greg@gmail.com>"]
 license = "Apache-2.0"