diffmonkey 1.1.0__tar.gz → 1.1.1__tar.gz

{diffmonkey-1.1.0/src/diffmonkey.egg-info → diffmonkey-1.1.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: diffmonkey
-Version: 1.1.0
+Version: 1.1.1
 Summary: Type-aware, key-based structural diffing of tabular datasets with human- and machine-readable reports.
 Author-email: RexBytes <pythonic@rexbytes.com>
 License: MIT License
@@ -40,9 +40,9 @@ Classifier: Topic :: Utilities
 Requires-Python: >=3.11
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: cleanmonkey
-Requires-Dist: typemonkey
-Requires-Dist: datemonkey
+Requires-Dist: cleanmonkey>=0.2.0
+Requires-Dist: typemonkey>=1.2.0
+Requires-Dist: datemonkey>=0.2.0
 Provides-Extra: excel
 Requires-Dist: openpyxl>=3.0; extra == "excel"
 Provides-Extra: dsv

{diffmonkey-1.1.0 → diffmonkey-1.1.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "diffmonkey"
-version = "1.1.0"
+version = "1.1.1"
 description = "Type-aware, key-based structural diffing of tabular datasets with human- and machine-readable reports."
 readme = "README.md"
 license = { file = "LICENSE" }
@@ -23,9 +23,9 @@ classifiers = [
     "Topic :: Utilities",
 ]
 dependencies = [
-    "cleanmonkey",
-    "typemonkey",
-    "datemonkey",
+    "cleanmonkey>=0.2.0",
+    "typemonkey>=1.2.0",
+    "datemonkey>=0.2.0",
 ]
 
 [project.optional-dependencies]

{diffmonkey-1.1.0 → diffmonkey-1.1.1}/src/diffmonkey/__init__.py

@@ -31,7 +31,7 @@ from .models import (
 )
 from .readers import read_csv, read_excel, read_table
 
-__version__ = "1.1.0"
+__version__ = "1.1.1"
 
 __all__ = [
     "compare",

{diffmonkey-1.1.0 → diffmonkey-1.1.1}/src/diffmonkey/comparators.py

@@ -76,7 +76,8 @@ def _clean(value: Any, settings: CompareSettings) -> str:
     text = value if isinstance(value, str) else str(value)
     if settings.normalize_whitespace:
         # 'minimal' would skip whitespace collapsing; 'default' strips and
-        # collapses runs and removes invisible characters — exactly the
+        # collapses runs, removes invisible characters, and folds typographic
+        # variants (smart quotes→straight, dashes→'-', '…'→'...') — exactly the
         # false-diff sources we want gone. strip=True is safe for cell values.
         return cleanmonkey.clean(text, profile="default")
     return text

{diffmonkey-1.1.0 → diffmonkey-1.1.1}/src/diffmonkey/compare.py

@@ -130,9 +130,11 @@ def compare(
             (``math.isclose``). Default ``1e-9``.
         abs_tol: Absolute floating-point tolerance for numeric columns. Default
             ``0.0`` — set this to compare values near zero.
-        normalize_whitespace: Collapse/strip whitespace and drop invisible
-            characters before comparing strings (via ``cleanmonkey``). Default
-            True.
+        normalize_whitespace: Collapse/strip whitespace, drop invisible
+            characters, and fold typographic variants (smart quotes→straight,
+            dashes/minus→``-``, ``…``→``...``) before comparing strings (via
+            ``cleanmonkey``'s ``default`` profile), so cosmetic punctuation
+            differences are not reported as changes. Default True.
         null_equivalent: Treat all null spellings (``None``, ``""``,
             whitespace-only, ``"na"``, …) as one value, so two different nulls
             are equal and null↔value is a change. Default True.

{diffmonkey-1.1.0 → diffmonkey-1.1.1/src/diffmonkey.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: diffmonkey
-Version: 1.1.0
+Version: 1.1.1
 Summary: Type-aware, key-based structural diffing of tabular datasets with human- and machine-readable reports.
 Author-email: RexBytes <pythonic@rexbytes.com>
 License: MIT License
@@ -40,9 +40,9 @@ Classifier: Topic :: Utilities
 Requires-Python: >=3.11
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: cleanmonkey
-Requires-Dist: typemonkey
-Requires-Dist: datemonkey
+Requires-Dist: cleanmonkey>=0.2.0
+Requires-Dist: typemonkey>=1.2.0
+Requires-Dist: datemonkey>=0.2.0
 Provides-Extra: excel
 Requires-Dist: openpyxl>=3.0; extra == "excel"
 Provides-Extra: dsv

{diffmonkey-1.1.0 → diffmonkey-1.1.1}/src/diffmonkey.egg-info/requires.txt

@@ -1,6 +1,6 @@
-cleanmonkey
-typemonkey
-datemonkey
+cleanmonkey>=0.2.0
+typemonkey>=1.2.0
+datemonkey>=0.2.0
 
 [dev]
 pytest>=7.0

{diffmonkey-1.1.0 → diffmonkey-1.1.1}/tests/test_review_fixes.py

@@ -524,3 +524,41 @@ def test_fractional_bare_number_not_a_date():
     assert c._as_date("1.2") is UNPARSEABLE
     # 8-digit compact ISO still parses (integer part >= 8 digits).
     assert c._as_date("20240301") is not UNPARSEABLE
+
+
+# =====================================================================
+
+# Panel 6 (dependency-upgrade verification: cleanmonkey 0.2.0 / typemonkey 1.2.0
+# / datemonkey 0.2.0). sonnet flagged that cleanmonkey's "default" profile folds
+# typographic variants (smart quotes, em/en-dashes, ellipsis) to ASCII, so e.g.
+# "Best—seller" compares equal to "Best-seller". Adjudication: NOT a regression
+# (cleanmonkey 0.1.0's "default" profile is byte-identical on these inputs) and
+# the equivalence is intended false-diff removal (SKILL.md documents smart-quote
+# equivalence); only the docstring/comment under-described it. These tests pin
+# the now-documented contract and its opt-out so a future cleanmonkey change or
+# a doc drift fails loudly.
+
+
+def test_typographic_variants_fold_when_normalising():
+    # normalize_whitespace=True (default): cosmetic punctuation differences are
+    # not reported as changes.
+    old = [
+        {"id": "1", "t": "Best-seller"},     # ASCII hyphen
+        {"id": "2", "t": 'say "hi"'},        # straight quotes
+        {"id": "3", "t": "a...b"},           # ASCII ellipsis
+    ]
+    new = [
+        {"id": "1", "t": "Best—seller"},  # em-dash
+        {"id": "2", "t": "say “hi”"},  # smart quotes
+        {"id": "3", "t": "a…b"},            # ellipsis char
+    ]
+    assert compare(old, new, key="id").summary.changed == 0
+
+
+def test_typographic_variants_preserved_when_not_normalising():
+    # normalize_whitespace=False: the raw bytes are compared, so the same pairs
+    # are reported as changes (intent-preserving opt-out).
+    old = [{"id": "1", "t": "Best-seller"}]
+    new = [{"id": "1", "t": "Best—seller"}]
+    summary = compare(old, new, key="id", normalize_whitespace=False).summary
+    assert summary.changed == 1