chunklet-py 2.3.0__tar.gz → 2.3.2__tar.gz

{chunklet_py-2.3.0 → chunklet_py-2.3.2}/LICENSE

@@ -18,4 +18,4 @@ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
+SOFTWARE.

{chunklet_py-2.3.0/src/chunklet_py.egg-info → chunklet_py-2.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: chunklet-py
-Version: 2.3.0
+Version: 2.3.2
 Summary: High-fidelity context-aware chunking and interactive visualization for RAG. Advanced segmentation for code and documents, because your LLM is only as smart as the fragments you feed it.
 Author-email: speedyk_005 <speedy40115719@gmail.com>
 License-Expression: MIT
@@ -28,12 +28,11 @@ License-File: LICENSE
 Requires-Dist: loguru<1.0,>=0.7.3
 Requires-Dist: pysbd<1.0,>=0.3.4
 Requires-Dist: sentsplit<2.0,>=1.0.8
-Requires-Dist: sentencex<2.0,>=1.0.0; platform_system != "Android"
-Requires-Dist: sentencex<=0.6.1; platform_system == "Android"
+Requires-Dist: sentencex<=0.6.1; sys_platform == "linux" and (platform_machine == "aarch64" or platform_machine == "armv7l")
+Requires-Dist: sentencex<2.0,>=1.0.0; sys_platform != "linux" or (platform_machine != "aarch64" and platform_machine != "armv7l")
 Requires-Dist: indic-nlp-library<1.0,>=0.92
 Requires-Dist: py3langid<1.0,>=0.3.0
 Requires-Dist: mpire<3.0,>=2.10.2
-Requires-Dist: dotdict3<2.0,>=1.1.0
 Requires-Dist: more-itertools<12.0,>=10.6.0
 Requires-Dist: regex>=2025.7.29
 Requires-Dist: pydantic<3.0,>=2.11.0

{chunklet_py-2.3.0 → chunklet_py-2.3.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "chunklet-py"
-version = "2.3.0"
+version = "2.3.2"
 description = "High-fidelity context-aware chunking and interactive visualization for RAG. Advanced segmentation for code and documents, because your LLM is only as smart as the fragments you feed it."
 authors = [
   { name = "speedyk_005", email = "speedy40115719@gmail.com" }
@@ -38,14 +38,13 @@ dependencies = [
   "pysbd>=0.3.4,<1.0",
   "sentsplit>=1.0.8,<2.0",
 
-  # Rust binding (1.0+) for faster performance on supported platforms, legacy for Android.
-  "sentencex>=1.0.0,<2.0; platform_system != 'Android'",
-  "sentencex<=0.6.1; platform_system == 'Android'",
+  # Rust binding (1.0+) for faster performance on supported platforms, legacy for Android temporarily.
+  "sentencex<=0.6.1; sys_platform == 'linux' and (platform_machine == 'aarch64' or platform_machine == 'armv7l')",
+  "sentencex>=1.0.0,<2.0; sys_platform != 'linux' or (platform_machine != 'aarch64' and platform_machine != 'armv7l')",
 
   "indic-nlp-library>=0.92,<1.0",
   "py3langid>=0.3.0,<1.0",
   "mpire>=2.10.2,<3.0",
-  "dotdict3>=1.1.0,<2.0",
   "more-itertools>=10.6.0,<12.0",
   "regex>=2025.7.29",
   "pydantic>=2.11.0,<3.0",

{chunklet_py-2.3.0 → chunklet_py-2.3.2}/src/chunklet/base_chunker.py

@@ -7,7 +7,7 @@ Defines the interface for chunkers.
 from abc import ABC, abstractmethod
 from collections.abc import Generator
 
-from dotdict3 import DotDict
+from chunklet.common.dotdict import DotDict
 from loguru import logger
 
 

{chunklet_py-2.3.0 → chunklet_py-2.3.2}/src/chunklet/code_chunker/code_chunker.py

@@ -29,7 +29,7 @@ from itertools import chain
 from pathlib import Path
 from typing import Annotated, Any, Callable, Generator, Literal
 
-from dotdict3 import DotDict
+from chunklet.common.dotdict import DotDict
 from more_itertools import unique_everseen
 from pydantic import Field
 

chunklet_py-2.3.2/src/chunklet/common/dotdict.py

@@ -0,0 +1,106 @@
+"""
+DotDict/DotList with Box-compatible serialization.
+
+This module embeds code originally from ``dotdict3`` (MIT License).
+Co-authored by:
+  - speedyk-005
+  - Patrick Elmer <patrick@elmer.ws> (https://github.com/dotdict/dotdict3)
+
+Modifications: added ``to_dict()``, ``to_json()``, ``to_yaml()``,
+``to_msgpack()``, ``to_toml()``, ``to_csv()`` for backward compatibility
+with the python-box API.
+"""
+
+import json
+
+
+class DotDict(dict):
+    def __init__(self, data=None):
+        if data is not None:
+            for key, value in data.items():
+                self[key] = value
+
+    def __setitem__(self, key, value):
+        return super().__setitem__(key, _convert(value))
+
+    # Redirect attribute operations to dictionary methods
+    __delattr__ = dict.__delitem__
+    __getattr__ = dict.__getitem__
+    __setattr__ = __setitem__
+
+    def to_dict(self):
+        return {k: _to_plain(v) for k, v in self.items()}
+
+    def to_json(self, filename=None, **kwargs):
+        if filename:
+            with open(filename, "w", encoding="utf-8") as f:
+                json.dump(self.to_dict(), f, ensure_ascii=False, **kwargs)
+        else:
+            return json.dumps(self.to_dict(), ensure_ascii=False, **kwargs)
+
+    def to_yaml(self, filename=None, default_flow_style=False, **kwargs):
+        try:
+            import yaml
+        except ImportError:
+            raise ImportError("pyyaml required: pip install pyyaml") from None
+        if filename:
+            with open(filename, "w", encoding="utf-8") as f:
+                yaml.dump(self.to_dict(), f, default_flow_style=default_flow_style, **kwargs)
+        else:
+            return yaml.dump(self.to_dict(), default_flow_style=default_flow_style, **kwargs)
+
+    def to_toml(self, filename=None):
+        try:
+            import toml
+        except ImportError:
+            raise ImportError("toml required: pip install toml") from None
+        if filename:
+            with open(filename, "w", encoding="utf-8") as f:
+                toml.dump(self.to_dict(), f)
+        else:
+            return toml.dumps(self.to_dict())
+
+    def to_msgpack(self, filename=None, **kwargs):
+        try:
+            import msgpack
+        except ImportError:
+            raise ImportError("msgpack required: pip install msgpack") from None
+        if filename:
+            with open(filename, "wb") as f:
+                msgpack.pack(self.to_dict(), f, **kwargs)
+        else:
+            return msgpack.packb(self.to_dict(), **kwargs)
+
+
+class DotList(list):
+    def __init__(self, items=None):
+        if items is not None:
+            for item in items:
+                self.append(item)
+
+    def append(self, items):
+        return super().append(_convert(items))
+
+    def insert(self, index, items):
+        return super().insert(index, _convert(items))
+
+    def to_dict(self):
+        return [_to_plain(v) for v in self]
+
+
+def _convert(obj):
+    """Recursively converts dicts/lists to DotDict/DotList if not already converted."""
+    if isinstance(obj, dict) and not isinstance(obj, DotDict):
+        return DotDict(obj)
+    if isinstance(obj, list) and not isinstance(obj, DotList):
+        return DotList(obj)
+    return obj
+
+
+def _to_plain(obj):
+    """Recursively converts DotDict/DotList back to plain dict/list."""
+    if isinstance(obj, DotDict):
+        return {k: _to_plain(v) for k, v in obj.items()}
+    if isinstance(obj, (DotList, tuple)):
+        return [_to_plain(v) for v in obj]
+    return obj

{chunklet_py-2.3.0 → chunklet_py-2.3.2}/src/chunklet/document_chunker/_plain_text_chunker.py

@@ -5,7 +5,7 @@ from collections.abc import Iterable
 from functools import partial
 from typing import Annotated, Any, Callable, Generator, Literal
 
-from dotdict3 import DotDict
+from chunklet.common.dotdict import DotDict
 from loguru import logger
 from pydantic import Field
 
@@ -120,7 +120,7 @@ class PlainTextChunker:
         """
         chunks_out = []
         for i, chunk_str in enumerate(chunks, start=1):
-            chunk = DotDict()
+            chunk = DotDict({})
             chunk.content = chunk_str.strip()
             chunk.metadata = copy.deepcopy(base_metadata)
             chunk.metadata["chunk_num"] = i

{chunklet_py-2.3.0 → chunklet_py-2.3.2}/src/chunklet/document_chunker/document_chunker.py

@@ -3,7 +3,7 @@ from itertools import chain, tee
 from pathlib import Path
 from typing import Annotated, Any, Callable, Generator, Iterable, Literal
 
-from dotdict3 import DotDict
+from chunklet.common.dotdict import DotDict
 from loguru import logger
 from more_itertools import ilen, split_at
 from pydantic import Field
@@ -607,7 +607,7 @@ class DocumentChunker(BaseChunker):
             sections_per_path[curr_path] -= 1
 
     @deprecated_callable(
-        use_instead="chunk_file", deprecated_in="2.2.0", removed_in="3.0.0"
+        use_instead="chunk_file or chunk_text", deprecated_in="2.2.0", removed_in="3.0.0"
     )
     def chunk(  # pragma: no cover
         self,
@@ -632,7 +632,7 @@ class DocumentChunker(BaseChunker):
         return self.chunk_file(**params)
 
     @deprecated_callable(
-        use_instead="chunk_files", deprecated_in="2.2.0", removed_in="3.0.0"
+        use_instead="chunk_files or chunk_texts", deprecated_in="2.2.0", removed_in="3.0.0"
     )
     def batch_chunk(  # pragma: no cover
         self,

{chunklet_py-2.3.0 → chunklet_py-2.3.2}/src/chunklet/sentence_splitter/_universal_splitter.py

@@ -36,6 +36,11 @@ class UniversalSplitter:
         self.numbered_list_pattern = re.compile(r"[\n:]\s*\p{N}\.")
 
         # Core sentence split regex
+        # NOTE: Acronyms like "U.S.A" are protected primarily by the lookahead (?=\s+...).
+        # Since "U.S.A," has no space after it (just punctuation), the lookahead fails
+        # and no split occurs. The negative lookbehind handles other abbreviations like "Dr."
+        # This means acronym protection is *not* dependent on masking—it's explicit in the
+        # lookahead requirement for whitespace or newline before the next uppercase letter.
         self.sentence_end_pattern = re.compile(
             rf"""
             (?<!\b(\p{{Lu}}\p{{Ll}}{{1, 4}}\.)*)   # Latin-only abbreviation

{chunklet_py-2.3.0 → chunklet_py-2.3.2/src/chunklet_py.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: chunklet-py
-Version: 2.3.0
+Version: 2.3.2
 Summary: High-fidelity context-aware chunking and interactive visualization for RAG. Advanced segmentation for code and documents, because your LLM is only as smart as the fragments you feed it.
 Author-email: speedyk_005 <speedy40115719@gmail.com>
 License-Expression: MIT
@@ -28,12 +28,11 @@ License-File: LICENSE
 Requires-Dist: loguru<1.0,>=0.7.3
 Requires-Dist: pysbd<1.0,>=0.3.4
 Requires-Dist: sentsplit<2.0,>=1.0.8
-Requires-Dist: sentencex<2.0,>=1.0.0; platform_system != "Android"
-Requires-Dist: sentencex<=0.6.1; platform_system == "Android"
+Requires-Dist: sentencex<=0.6.1; sys_platform == "linux" and (platform_machine == "aarch64" or platform_machine == "armv7l")
+Requires-Dist: sentencex<2.0,>=1.0.0; sys_platform != "linux" or (platform_machine != "aarch64" and platform_machine != "armv7l")
 Requires-Dist: indic-nlp-library<1.0,>=0.92
 Requires-Dist: py3langid<1.0,>=0.3.0
 Requires-Dist: mpire<3.0,>=2.10.2
-Requires-Dist: dotdict3<2.0,>=1.1.0
 Requires-Dist: more-itertools<12.0,>=10.6.0
 Requires-Dist: regex>=2025.7.29
 Requires-Dist: pydantic<3.0,>=2.11.0

{chunklet_py-2.3.0 → chunklet_py-2.3.2}/src/chunklet_py.egg-info/SOURCES.txt

@@ -13,6 +13,7 @@ src/chunklet/code_chunker/utils.py
 src/chunklet/common/__init__.py
 src/chunklet/common/batch_runner.py
 src/chunklet/common/deprecation.py
+src/chunklet/common/dotdict.py
 src/chunklet/common/logging_utils.py
 src/chunklet/common/path_utils.py
 src/chunklet/common/token_utils.py

{chunklet_py-2.3.0 → chunklet_py-2.3.2}/src/chunklet_py.egg-info/requires.txt

@@ -4,7 +4,6 @@ sentsplit<2.0,>=1.0.8
 indic-nlp-library<1.0,>=0.92
 py3langid<1.0,>=0.3.0
 mpire<3.0,>=2.10.2
-dotdict3<2.0,>=1.1.0
 more-itertools<12.0,>=10.6.0
 regex>=2025.7.29
 pydantic<3.0,>=2.11.0
@@ -12,10 +11,10 @@ typer<1.0,>=0.19.0
 charset-normalizer<4.0,>=3.4.2
 setuptools<81
 
-[:platform_system != "Android"]
+[:sys_platform != "linux" or (platform_machine != "aarch64" and platform_machine != "armv7l")]
 sentencex<2.0,>=1.0.0
 
-[:platform_system == "Android"]
+[:sys_platform == "linux" and (platform_machine == "aarch64" or platform_machine == "armv7l")]
 sentencex<=0.6.1
 
 [all]