This is an automated email from the ASF dual-hosted git repository.
potiuk pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/airflow.git
The following commit(s) were added to refs/heads/main by this push:
new 33cb64afbd2 fix airflowignore negation does not work in subfolders
(#58740)
33cb64afbd2 is described below
commit 33cb64afbd214499a3da28182d853e087da14186
Author: Henry Chen <[email protected]>
AuthorDate: Sun Nov 30 07:22:50 2025 +0800
fix airflowignore negation does not work in subfolders (#58740)
---
airflow-core/src/airflow/utils/file.py | 6 +-
.../tests/unit/plugins/test_plugin_ignore.py | 11 ++--
airflow-core/tests/unit/utils/test_file.py | 65 ++++++++++++++++++++++
3 files changed, 76 insertions(+), 6 deletions(-)
diff --git a/airflow-core/src/airflow/utils/file.py
b/airflow-core/src/airflow/utils/file.py
index d546c0ac352..76cdf846fe0 100644
--- a/airflow-core/src/airflow/utils/file.py
+++ b/airflow-core/src/airflow/utils/file.py
@@ -111,7 +111,11 @@ class _GlobIgnoreRule(NamedTuple):
for rule in rules:
if not isinstance(rule, _GlobIgnoreRule):
raise ValueError(f"_GlobIgnoreRule cannot match rules of type:
{type(rule)}")
- rel_path = str(path.relative_to(rule.relative_to) if
rule.relative_to else path.name)
+ rel_obj = path.relative_to(rule.relative_to) if rule.relative_to
else Path(path.name)
+ if path.is_dir():
+ rel_path = f"{rel_obj.as_posix()}/"
+ else:
+ rel_path = rel_obj.as_posix()
if (
rule.wild_match_pattern.include is not None
and rule.wild_match_pattern.match_file(rel_path) is not None
diff --git a/airflow-core/tests/unit/plugins/test_plugin_ignore.py
b/airflow-core/tests/unit/plugins/test_plugin_ignore.py
index 8d8bd397b77..671cb71dacc 100644
--- a/airflow-core/tests/unit/plugins/test_plugin_ignore.py
+++ b/airflow-core/tests/unit/plugins/test_plugin_ignore.py
@@ -94,20 +94,21 @@ class TestIgnorePluginFile:
should_ignore_files = {
"test_notload.py",
"test_notload_sub.py",
- "test_noneload_sub1.py",
+ "subdir1/test_noneload_sub1.py",
+ "subdir2/test_shouldignore.py",
+ "subdir3/test_notload_sub3.py",
}
should_not_ignore_files = {
"test_load.py",
- "test_load_sub1.py",
- "test_shouldignore.py", # moved to here because it should not
ignore, as we do not ignore all
- # things from subdir 2
+ "subdir1/test_load_sub1.py",
}
ignore_list_file = ".airflowignore_glob"
print("-" * 20)
for raw_file_path in find_path_from_directory(plugin_folder_path,
ignore_list_file, "glob"):
file_path = Path(raw_file_path)
if file_path.is_file() and file_path.suffix == ".py":
- detected_files.add(file_path.name)
+ rel_path = file_path.relative_to(plugin_folder_path).as_posix()
+ detected_files.add(rel_path)
print(file_path)
print("-" * 20)
diff --git a/airflow-core/tests/unit/utils/test_file.py
b/airflow-core/tests/unit/utils/test_file.py
index bfdff826754..fc46a3a6560 100644
--- a/airflow-core/tests/unit/utils/test_file.py
+++ b/airflow-core/tests/unit/utils/test_file.py
@@ -189,6 +189,71 @@ class TestListPyFilesPath:
assert file_utils.might_contain_dag(file_path=file_path_with_dag,
safe_mode=True)
+ def test_airflowignore_negation_unignore_subfolder_file_glob(self,
tmp_path):
+ """Ensure negation rules can unignore a subfolder and a file inside it
when using glob syntax.
+
+ Patterns:
+ * -> ignore everything
+ !subfolder/ -> unignore the subfolder (must match
directory rule)
+ !subfolder/keep.py -> unignore a specific file inside the
subfolder
+ """
+ dags_root = tmp_path / "dags"
+ (dags_root / "subfolder").mkdir(parents=True)
+ # files
+ (dags_root / "drop.py").write_text("raise Exception('ignored')\n")
+ (dags_root / "subfolder" / "keep.py").write_text("# should be
discovered\n")
+ (dags_root / "subfolder" / "drop.py").write_text("raise
Exception('ignored')\n")
+
+ (dags_root / ".airflowignore").write_text(
+ "\n".join(
+ [
+ "*",
+ "!subfolder/",
+ "!subfolder/keep.py",
+ ]
+ )
+ )
+
+ detected = set()
+ for raw in find_path_from_directory(dags_root, ".airflowignore",
"glob"):
+ p = Path(raw)
+ if p.is_file() and p.suffix == ".py":
+ detected.add(p.relative_to(dags_root).as_posix())
+
+ assert detected == {"subfolder/keep.py"}
+
+ def test_airflowignore_negation_nested_with_globstar(self, tmp_path):
+ """Negation with ** should work for nested subfolders."""
+ dags_root = tmp_path / "dags"
+ nested = dags_root / "a" / "b" / "subfolder"
+ nested.mkdir(parents=True)
+
+ # files
+ (dags_root / "ignore_top.py").write_text("raise
Exception('ignored')\n")
+ (nested / "keep.py").write_text("# should be discovered\n")
+ (nested / "drop.py").write_text("raise Exception('ignored')\n")
+
+ (dags_root / ".airflowignore").write_text(
+ "\n".join(
+ [
+ "*",
+ "!a/",
+ "!a/b/",
+ "!**/subfolder/",
+ "!**/subfolder/keep.py",
+ "drop.py",
+ ]
+ )
+ )
+
+ detected = set()
+ for raw in find_path_from_directory(dags_root, ".airflowignore",
"glob"):
+ p = Path(raw)
+ if p.is_file() and p.suffix == ".py":
+ detected.add(p.relative_to(dags_root).as_posix())
+
+ assert detected == {"a/b/subfolder/keep.py"}
+
@conf_vars({("core", "might_contain_dag_callable"):
"unit.utils.test_file.might_contain_dag"})
def test_might_contain_dag(self):
"""Test might_contain_dag_callable"""
