git: f0b69e29e340 - main - sysutils/py-magika: New port: Tool to determine the content type of a file with deep learning

The branch main has been updated by yuri:

URL: https://cgit.FreeBSD.org/ports/commit/?id=f0b69e29e340f7e9e22c9a4c6b75f362a4c147cf

commit f0b69e29e340f7e9e22c9a4c6b75f362a4c147cf
Author:     Yuri Victorovich <[email protected]>
AuthorDate: 2026-08-07 03:06:36 +0000
Commit:     Yuri Victorovich <[email protected]>
CommitDate: 2026-08-07 05:09:29 +0000

    sysutils/py-magika: New port: Tool to determine the content type of a file with deep learning
---
 sysutils/Makefile                             |   1 +
 sysutils/py-magika/Makefile                   |  42 +++++++++
 sysutils/py-magika/distinfo                   |   3 +
 sysutils/py-magika/files/example.py           |  45 ++++++++++
 sysutils/py-magika/files/patch-pyproject.toml | 118 ++++++++++++++++++++++++++
 sysutils/py-magika/pkg-descr                  |   3 +
 6 files changed, 212 insertions(+)

diff --git a/sysutils/Makefile b/sysutils/Makefile
index 18d751484b28..cdb95a400545 100644
--- a/sysutils/Makefile
+++ b/sysutils/Makefile
@@ -1104,6 +1104,7 @@
     SUBDIR += py-kubernetes
     SUBDIR += py-leviathan
     SUBDIR += py-liquidctl
+    SUBDIR += py-magika
     SUBDIR += py-mitogen
     SUBDIR += py-mqttwarn
     SUBDIR += py-nagiosplugin
diff --git a/sysutils/py-magika/Makefile b/sysutils/py-magika/Makefile
new file mode 100644
index 000000000000..0d33d24cfec1
--- /dev/null
+++ b/sysutils/py-magika/Makefile
@@ -0,0 +1,42 @@
+PORTNAME=	magika
+DISTVERSIONPREFIX=	python-v
+DISTVERSION=	1.0.3
+CATEGORIES=	sysutils python # machine-learning
+PKGNAMEPREFIX=	${PYTHON_PKGNAMEPREFIX}
+
+MAINTAINER=	[email protected]
+COMMENT=	Tool to determine the content type of a file with deep learning
+WWW=		https://github.com/google/magika
+
+LICENSE=	APACHE20
+LICENSE_FILE=	${WRKSRC}/../LICENSE
+
+BUILD_DEPENDS=	${PYTHON_PKGNAMEPREFIX}hatchling>0:devel/py-hatchling@${PY_FLAVOR}
+RUN_DEPENDS=	${PYTHON_PKGNAMEPREFIX}click>=8.1.7:devel/py-click@${PY_FLAVOR} \
+		${PYTHON_PKGNAMEPREFIX}onnxruntime>=1.17.0:misc/py-onnxruntime@${PY_FLAVOR}
+TEST_DEPENDS=	${PYTHON_PKGNAMEPREFIX}requests>0:www/py-requests@${PY_FLAVOR} \
+		${PYTHON_PKGNAMEPREFIX}tqdm>0:misc/py-tqdm@${PY_FLAVOR} \
+		${PYTHON_PKGNAMEPREFIX}dacite>=1.9.2:devel/py-dacite@${PY_FLAVOR} \
+		${PYTHON_PKGNAMEPREFIX}tomli>=2.0.1:textproc/py-tomli@${PY_FLAVOR} \
+		${PYTHON_PKGNAMEPREFIX}tomli-w>=1.0.0:textproc/py-tomli-w@${PY_FLAVOR}
+
+USES=		python shebangfix
+USE_PYTHON=	pep517 concurrent autoplist pytest
+
+SHEBANG_FILES=	${WRKSRC}/src/magika/cli/magika_client.py \
+		${WRKSRC}/src/magika/cli/magika_rust_client_not_found_warning.py
+
+USE_GITHUB=	yes
+GH_ACCOUNT=	google
+GH_PROJECT=	magika
+
+WRKSRC_SUBDIR=	python
+
+NO_ARCH=	yes
+
+TEST_ENV=	${MAKE_ENV} PYTHONPATH=${STAGEDIR}${PYTHONPREFIX_SITELIBDIR}:${PYTHONPREFIX_SITELIBDIR}
+TEST_WRKSRC=	${WRKSRC}/tests
+
+# tests as of 1.0.3: 32 passed, 1 skipped, 1 warning in 12.27s
+
+.include <bsd.port.mk>
diff --git a/sysutils/py-magika/distinfo b/sysutils/py-magika/distinfo
new file mode 100644
index 000000000000..c3d879e67746
--- /dev/null
+++ b/sysutils/py-magika/distinfo
@@ -0,0 +1,3 @@
+TIMESTAMP = 1786057028
+SHA256 (google-magika-python-v1.0.3_GH0.tar.gz) = 8d24d6638f30f4c89e5b71fd2f6dee665e5381215b8f9d1ba88b065e57edaff4
+SIZE (google-magika-python-v1.0.3_GH0.tar.gz) = 84265337
diff --git a/sysutils/py-magika/files/example.py b/sysutils/py-magika/files/example.py
new file mode 100644
index 000000000000..5d6fb3d22aa8
--- /dev/null
+++ b/sysutils/py-magika/files/example.py
@@ -0,0 +1,45 @@
+import argparse
+import sys
+from pathlib import Path
+from magika import Magika
+
+def main():
+    parser = argparse.ArgumentParser(
+        description="Verify that an uploaded file matches its format criteria."
+    )
+    parser.add_argument(
+        "image_path", 
+        type=str, 
+        help="The path to the file you want to check."
+    )
+    
+    args = parser.parse_args()
+    target_path = Path(args.image_path)
+
+    if not target_path.is_file():
+        print(f"Error: File '{target_path}' does not exist.", file=sys.stderr)
+        sys.exit(1)
+
+    magika = Magika()
+    result = magika.identify_path(target_path)
+    
+    # --- UPDATED FOR NEW MAGIKA API ---
+    detected_group = result.output.group
+    detected_label = result.output.label
+    
+    # Use getattr to safely handle variance between 0.5.x and 0.6.x versions
+    confidence = getattr(result, "similarity_score", getattr(result, "score", 0.0))
+
+    print(f"Inspecting: {target_path.name}")
+    print(f"Result: {detected_label} ({confidence:.2%})")
+
+    # In your test, a Makefile is categorized under the "text" or "code" group
+    if detected_group != "image" and detected_label != "makefile":
+        print(f"❌ Verification failed. Group: {detected_group}, Label: {detected_label}")
+        sys.exit(1)
+
+    print("✅ File verified successfully.")
+
+if __name__ == "__main__":
+    main()
+
diff --git a/sysutils/py-magika/files/patch-pyproject.toml b/sysutils/py-magika/files/patch-pyproject.toml
new file mode 100644
index 000000000000..a5f8214b8738
--- /dev/null
+++ b/sysutils/py-magika/files/patch-pyproject.toml
@@ -0,0 +1,118 @@
+-- Replace maturin/Rust-based build configuration with hatchling for a pure
+-- Python package. The Python package uses onnxruntime directly and does not
+-- require the Rust CLI binary. The PyPI sdist already ships this hatchling
+-- pyproject.toml, but the GitHub source used by the port references the
+-- ../rust/cli/Cargo.toml maturin build, which is unsupported on FreeBSD.
+-- This patch mirrors the upstream PyPI sdist configuration.
+
+--- pyproject.toml.orig	2026-05-04 08:32:03 UTC
++++ pyproject.toml
+@@ -2,12 +2,14 @@ authors = [
+ name = "magika"
+ description = "A tool to determine the content type of a file with deep learning"
+ authors = [
+-    {name = "Magika Developers", email = "[email protected]"},
++    { name = "Magika Developers", email = "[email protected]" },
+ ]
+ readme = "README.md"
+-license = {"text" = "Apache-2.0"}
+ requires-python = ">=3.8"
+-keywords = ["content type detection", "machine learning"]
++keywords = [
++    "content type detection",
++    "machine learning",
++]
+ classifiers = [
+     "Development Status :: 5 - Production/Stable",
+     "Environment :: Console",
+@@ -33,22 +35,17 @@ version = "1.0.3"
+     "Typing :: Typed",
+ ]
+ version = "1.0.3"
+-
+ dependencies = [
+     "click>=8.1.7",
+-
+-    # Python 3.14+: Min version 1.24.1
+     "onnxruntime>=1.24.1 ; python_version >= '3.14'",
+-    # Python 3.13: Min version 1.21.0
+     "onnxruntime>=1.21.0 ; python_version == '3.13'",
+-    # Python 3.11 - 3.12: Min version 1.17.0
+     "onnxruntime>=1.17.0 ; python_version >= '3.11' and python_version < '3.13'",
+-    # Python 3.10: Support dropped in 1.24.1 (use latest compatible)
+     "onnxruntime>=1.17.0, <1.24.1 ; python_version == '3.10'",
+-    # Python 3.9: Support dropped in 1.20.0
+     "onnxruntime>=1.17.0, <1.20.0 ; python_version <= '3.9'",
+ ]
+ 
++[project.license]
++text = "Apache-2.0"
+ 
+ [project.urls]
+ Homepage = "https://github.com/google/magika"
+@@ -57,6 +54,9 @@ Changelog = "https://github.com/google/magika/blob/mai
+ Issues = "https://github.com/google/magika/issues"
+ Changelog = "https://github.com/google/magika/blob/main/python/CHANGELOG.md"
+ 
++[project.scripts]
++magika-python-client = "magika.cli.magika_client:main"
++magika = "magika.cli.magika_rust_client_not_found_warning:main"
+ 
+ [dependency-groups]
+ dev = [
+@@ -72,30 +72,36 @@ dev = [
+ ]
+ 
+ [build-system]
+-requires = ["maturin>=1.12.2"]
+-build-backend = "maturin"
++requires = [
++    "hatchling",
++]
++build-backend = "hatchling.build"
+ 
+-[[tool.uv.index]]
+-url = "https://pypi.org/simple"
+-default = true
++[tool.uv]
++index = [
++    { url = "https://pypi.org/simple", default = true },
++]
+ 
+-[tool.maturin]
+-python-source = "src"
+-module-name = "magika"
+-bindings = "bin"
+-manifest-path = "../rust/cli/Cargo.toml"
+-locked = true
+-
+ [tool.ruff.lint]
+-# Enable Pyflakes (`F`) and a subset of the pycodestyle (`E`)  codes by default.
+-# Unlike Flake8, Ruff doesn't enable pycodestyle warnings (`W`) or
+-# McCabe complexity (`C901`) by default.
+-select = ["D", "E4", "E7", "E9", "F", "I001"]
+-ignore = ["D105"]
++select = [
++    "D",
++    "E4",
++    "E7",
++    "E9",
++    "F",
++    "I001",
++]
++ignore = [
++    "D105",
++]
+ 
+ [tool.ruff.lint.per-file-ignores]
+-"scripts/*" = ["D"]
+-"tests/*" = ["D"]
++"scripts/*" = [
++    "D",
++]
++"tests/*" = [
++    "D",
++]
+ 
+ [tool.ruff.lint.pydocstyle]
+ convention = "google"
diff --git a/sysutils/py-magika/pkg-descr b/sysutils/py-magika/pkg-descr
new file mode 100644
index 000000000000..db21beaa8aa0
--- /dev/null
+++ b/sysutils/py-magika/pkg-descr
@@ -0,0 +1,3 @@
+Magika is a tool to determine the content type of a file with deep learning.
+It can be used as a Python module or command-line tool to identify the type
+of files based on their contents.