ygg 0.1.23__tar.gz → 0.1.24__tar.gz

{ygg-0.1.23 → ygg-0.1.24}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ygg
-Version: 0.1.23
+Version: 0.1.24
 Summary: Type-friendly utilities for moving data between Python objects, Arrow, Polars, Pandas, Spark, and Databricks
 Author: Yggdrasil contributors
 License:                                  Apache License

{ygg-0.1.23 → ygg-0.1.24}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "ygg"
-version = "0.1.23"
+version = "0.1.24"
 description = "Type-friendly utilities for moving data between Python objects, Arrow, Polars, Pandas, Spark, and Databricks"
 readme = { file = "README.md", content-type = "text/markdown" }
 license = { file = "LICENSE" }

{ygg-0.1.23 → ygg-0.1.24}/src/ygg.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ygg
-Version: 0.1.23
+Version: 0.1.24
 Summary: Type-friendly utilities for moving data between Python objects, Arrow, Polars, Pandas, Spark, and Databricks
 Author: Yggdrasil contributors
 License:                                  Apache License

{ygg-0.1.23 → ygg-0.1.24}/src/yggdrasil/pyutils/python_env.py

@@ -17,7 +17,7 @@ import zipfile
 from contextlib import contextmanager
 from dataclasses import dataclass
 from pathlib import Path, PurePosixPath
-from typing import Any, Iterable, Iterator, Mapping, MutableMapping, Optional, Union
+from typing import Any, Iterable, Iterator, Mapping, MutableMapping, Optional, Union, List
 
 log = logging.getLogger(__name__)
 
@@ -40,9 +40,55 @@ _LOCKS_GUARD: threading.RLock = threading.RLock()
 
 # Installed into newly created envs (and on create() upsert when env already exists)
 DEFAULT_CREATE_PACKAGES: tuple[str, ...] = ("uv", "ygg")
-EXCLUDED_PACKAGES = {
-    "python-apt"
-}
+_NON_PIPABLE_NAMES = (
+    "python-apt",
+    "python3-apt",
+    "unattended-upgrades",
+    "apt",
+    "apt-utils",
+    "dpkg",
+    "adduser",
+    "lsb-release",
+    "software-properties-common",
+    "systemd",
+    "udev",
+    "dbus",
+)
+
+# Matches a requirement line that starts with one of those names, followed by:
+#  - end of line
+#  - a version specifier (==,>=, etc)
+#  - extras ([...])
+#  - an environment marker (; ...)
+_NON_PIPABLE_RE = re.compile(
+    r"^\s*(?:"
+    + "|".join(re.escape(n) for n in _NON_PIPABLE_NAMES)
+    + r")(?=\s*(?:$|==|~=|!=|<=|>=|<|>|\[|;))",
+    re.IGNORECASE,
+)
+
+
+
+def _filter_non_pipable_linux_packages(requirements: Iterable[str]) -> List[str]:
+    """
+    Remove Linux OS-level packages that aren't realistically installable via pip/uv on Databricks.
+    Keeps comments/blank lines as-is.
+    """
+    out: List[str] = []
+    for line in requirements:
+        s = line.strip()
+
+        # keep empty lines and comments
+        if not s or s.startswith("#"):
+            out.append(line)
+            continue
+
+        if _NON_PIPABLE_RE.match(s):
+            continue
+
+        out.append(line)
+
+    return out
 
 
 def _is_windows() -> bool:
@@ -882,10 +928,7 @@ print("RESULT:" + json.dumps(top_level))""".strip()
                     if not isinstance(top_level, list) or not all(isinstance(x, str) for x in top_level):
                         raise PythonEnvError(f"Unexpected top-level requirements payload: {top_level!r}")
 
-                    # exclude python-apt (yanked/unmaintained; not usable on Databricks)
-                    _pyapt_re = re.compile(r"^\s*python-apt(\s*(?:==|~=|!=|<=|>=|<|>|\[)|\s*(?:;|$))", re.IGNORECASE)
-
-                    filtered = [line for line in top_level if not _pyapt_re.match(line)]
+                    filtered = _filter_non_pipable_linux_packages(top_level)
 
                     req_in_text = "\n".join(filtered) + "\n"
                     req_in_path.write_text(req_in_text, encoding="utf-8")