From 40af32033e6c23403bbff82f75cb28d08f1eee67 Mon Sep 17 00:00:00 2001
From: Yuri Victorovich <yuri@FreeBSD.org>
Date: Wed, 8 Jul 2026 22:17:23 -0700
Subject: [PATCH] textproc/py-semchunk: New port: Python library for splitting
 text into semantically coherent chunks

---
 textproc/Makefile              |  1 +
 textproc/py-semchunk/Makefile  | 23 +++++++++++++++++++++++
 textproc/py-semchunk/distinfo  |  3 +++
 textproc/py-semchunk/pkg-descr |  4 ++++
 4 files changed, 31 insertions(+)
 create mode 100644 textproc/py-semchunk/Makefile
 create mode 100644 textproc/py-semchunk/distinfo
 create mode 100644 textproc/py-semchunk/pkg-descr

diff --git a/textproc/Makefile b/textproc/Makefile
index da9dea7a9239..b36e823ae5a2 100644
--- a/textproc/Makefile
+++ b/textproc/Makefile
@@ -1653,6 +1653,7 @@
     SUBDIR += py-scour
     SUBDIR += py-segments
     SUBDIR += py-segno
+    SUBDIR += py-semchunk
     SUBDIR += py-sense2vec
     SUBDIR += py-sentencepiece
     SUBDIR += py-simplebayes
diff --git a/textproc/py-semchunk/Makefile b/textproc/py-semchunk/Makefile
new file mode 100644
index 000000000000..c2b3db78e480
--- /dev/null
+++ b/textproc/py-semchunk/Makefile
@@ -0,0 +1,23 @@
+PORTNAME=	semchunk
+DISTVERSION=	4.1.1
+CATEGORIES=	textproc python
+MASTER_SITES=	PYPI
+PKGNAMEPREFIX=	${PYTHON_PKGNAMEPREFIX}
+
+MAINTAINER=	yuri@FreeBSD.org
+COMMENT=	Python library for splitting text into semantically coherent chunks
+WWW=		https://github.com/isaacus-dev/semchunk
+
+LICENSE=	MIT
+
+BUILD_DEPENDS=	${PYTHON_PKGNAMEPREFIX}hatchling>0:devel/py-hatchling@${PY_FLAVOR}
+RUN_DEPENDS=	${PYTHON_PKGNAMEPREFIX}tqdm>0:misc/py-tqdm@${PY_FLAVOR}
+
+USES=		python
+USE_PYTHON=	autoplist concurrent pep517
+NO_ARCH=	yes
+
+post-extract:
+	@${REINPLACE_CMD} -e 's/^license = "\([^"]*\)"/license = {text = "\1"}/' -e '/^license-files/d' ${WRKSRC}/pyproject.toml
+
+.include <bsd.port.mk>
diff --git a/textproc/py-semchunk/distinfo b/textproc/py-semchunk/distinfo
new file mode 100644
index 000000000000..608b941540ab
--- /dev/null
+++ b/textproc/py-semchunk/distinfo
@@ -0,0 +1,3 @@
+TIMESTAMP = 1783063884
+SHA256 (semchunk-4.1.1.tar.gz) = f27dc85716a0e9509a3f99c78b961cc476b3120efe711ee9a9a950ed1bd0f8ba
+SIZE (semchunk-4.1.1.tar.gz) = 24910
diff --git a/textproc/py-semchunk/pkg-descr b/textproc/py-semchunk/pkg-descr
new file mode 100644
index 000000000000..bffa2872bf3b
--- /dev/null
+++ b/textproc/py-semchunk/pkg-descr
@@ -0,0 +1,4 @@
+Semchunk splits text into smaller chunks while preserving as much
+local semantic context as possible. It is designed for use in RAG
+and LLM applications where text must be divided into segments that
+maintain coherent meaning for embedding and retrieval.