From acfbbaa745ff7ae6895fc5aa94fc2e0e322a4255 Mon Sep 17 00:00:00 2001
From: Yun Zheng Hu <hu@fox-it.com>
Date: Mon, 1 Dec 2025 12:12:49 +0100
Subject: [PATCH 1/3] Switch zstandard library to stdlib compression.zstd or
 backports.zstd

fixes #197
---
 flow/record/base.py          | 23 ++++++++++++++---------
 pyproject.toml               |  4 ++--
 tests/record/test_adapter.py |  4 +---
 3 files changed, 17 insertions(+), 14 deletions(-)

diff --git a/flow/record/base.py b/flow/record/base.py
index 0573af7d..dc8ef3e1 100644
--- a/flow/record/base.py
+++ b/flow/record/base.py
@@ -27,25 +27,33 @@
 from flow.record.exceptions import RecordAdapterNotFound, RecordDescriptorError
 from flow.record.utils import get_stdin, get_stdout
 
+# lz4
 try:
     import lz4.frame as lz4
 
     HAS_LZ4 = True
 except ImportError:
     HAS_LZ4 = False
+
+# bzip2
 try:
     import bz2
 
     HAS_BZ2 = True
 except ImportError:
     HAS_BZ2 = False
-try:
-    import zstandard as zstd
 
+# zstandard
+try:
+    if sys.version_info >= (3, 14):
+        from compression import zstd  # novermin
+    else:
+        from backports import zstd
     HAS_ZSTD = True
 except ImportError:
     HAS_ZSTD = False
 
+# fastavro
 try:
     import fastavro as avro  # noqa
 
@@ -727,8 +735,7 @@ def open_stream(fp: BinaryIO, mode: str) -> BinaryIO:
     elif HAS_LZ4 and peek_data[:4] == LZ4_MAGIC:
         fp = lz4.open(fp, mode=mode)
     elif HAS_ZSTD and peek_data[:4] == ZSTD_MAGIC:
-        dctx = zstd.ZstdDecompressor()
-        fp = dctx.stream_reader(fp)
+        fp = zstd.ZstdFile(fp, mode=mode)
 
     return fp
 
@@ -804,13 +811,11 @@ def open_path(path: str, mode: str, clobber: bool = True) -> IO:
             fp = lz4.open(path, mode)
         elif path.endswith((".zstd", ".zst")):
             if not HAS_ZSTD:
-                raise RuntimeError("zstandard python module not available")
+                raise RuntimeError("zstd python module not available")
             if not out:
-                dctx = zstd.ZstdDecompressor()
-                fp = dctx.stream_reader(pathobj.open("rb"))
+                fp = zstd.ZstdFile(pathobj.open("rb"), mode="rb")
             else:
-                cctx = zstd.ZstdCompressor()
-                fp = cctx.stream_writer(pathobj.open("wb"))
+                fp = zstd.ZstdFile(pathobj.open("wb"), mode="wb")
 
     # normal file or stdio for reading or writing
     if not fp:
diff --git a/pyproject.toml b/pyproject.toml
index 2b75e7f2..af59dd77 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -37,7 +37,7 @@ repository = "https://github.com/fox-it/flow.record"
 # Note: these compression libraries do not work well with pypy
 compression = [
     "lz4",
-    "zstandard; platform_python_implementation != 'PyPy'",
+    "backports.zstd; python_version < '3.14'",
 ]
 elastic = [
     "elasticsearch",
@@ -68,7 +68,7 @@ full = [
 [dependency-groups]
 compression = [
     "lz4",
-    "zstandard; platform_python_implementation != 'PyPy'",
+    "backports.zstd; python_version < '3.14'",
 ]
 elastic = [
     "elasticsearch",
diff --git a/tests/record/test_adapter.py b/tests/record/test_adapter.py
index 45eb80ad..a11c8470 100644
--- a/tests/record/test_adapter.py
+++ b/tests/record/test_adapter.py
@@ -85,12 +85,10 @@ def test_compressed_writer_reader(tmp_path: Path, compression: str) -> None:
     if compression == "lz4" and not HAS_LZ4:
         pytest.skip("lz4 module not installed")
     if compression == "zstd" and not HAS_ZSTD:
-        pytest.skip("zstandard module not installed")
+        pytest.skip("backports.zstd module not installed")
 
     if compression == "lz4" and platform.python_implementation() == "PyPy":
         pytest.skip("lz4 module not supported on PyPy")
-    if compression == "zstd" and platform.python_implementation() == "PyPy":
-        pytest.skip("zstandard module not supported on PyPy")
 
     p = tmp_path.joinpath(f"{compression}-test")
     p.mkdir()

From 676432f490a8fea21fd5707b696baba03fb0d757 Mon Sep 17 00:00:00 2001
From: Yun Zheng Hu <hu@fox-it.com>
Date: Mon, 1 Dec 2025 12:24:17 +0100
Subject: [PATCH 2/3] Apply code suggestions

---
 flow/record/base.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/flow/record/base.py b/flow/record/base.py
index dc8ef3e1..bfa4a2db 100644
--- a/flow/record/base.py
+++ b/flow/record/base.py
@@ -813,9 +813,9 @@ def open_path(path: str, mode: str, clobber: bool = True) -> IO:
             if not HAS_ZSTD:
                 raise RuntimeError("zstd python module not available")
             if not out:
-                fp = zstd.ZstdFile(pathobj.open("rb"), mode="rb")
+                fp = zstd.ZstdFile(path, mode)
             else:
-                fp = zstd.ZstdFile(pathobj.open("wb"), mode="wb")
+                fp = zstd.ZstdFile(path, mode)
 
     # normal file or stdio for reading or writing
     if not fp:

From 3c9940a718e93a7be94dfb3d615269e20c080fc9 Mon Sep 17 00:00:00 2001
From: Yun Zheng Hu <hu@fox-it.com>
Date: Mon, 1 Dec 2025 12:36:52 +0100
Subject: [PATCH 3/3] Fixes

---
 flow/record/base.py | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/flow/record/base.py b/flow/record/base.py
index bfa4a2db..7b574ee4 100644
--- a/flow/record/base.py
+++ b/flow/record/base.py
@@ -811,11 +811,8 @@ def open_path(path: str, mode: str, clobber: bool = True) -> IO:
             fp = lz4.open(path, mode)
         elif path.endswith((".zstd", ".zst")):
             if not HAS_ZSTD:
-                raise RuntimeError("zstd python module not available")
-            if not out:
-                fp = zstd.ZstdFile(path, mode)
-            else:
-                fp = zstd.ZstdFile(path, mode)
+                raise RuntimeError("backports.zstd python module not available")
+            fp = zstd.ZstdFile(path, mode)
 
     # normal file or stdio for reading or writing
     if not fp: