Merge branch 'master' into release/v1.3.0

Author: derek73

docs/customize.rst

@@ -53,7 +53,7 @@ Editable attributes of nameparser.config.CONSTANTS
 * :py:data:`~nameparser.config.CAPITALIZATION_EXCEPTIONS` - Dictionary of pieces that do not capitalize the first letter, e.g. "Ph.D".
 * :py:data:`~nameparser.config.REGEXES` - Regular expressions used to find words, initials, nicknames, etc.
 
-Each set-valued constant comes with :py:func:`~nameparser.config.SetManager.add` and :py:func:`~nameparser.config.SetManager.remove` methods for tuning
+Each set-valued constant comes with :py:func:`~nameparser.config.SetManager.add`, :py:func:`~nameparser.config.SetManager.remove`, and :py:func:`~nameparser.config.SetManager.discard` methods for tuning
 the constants for your project. These methods automatically lower case and
 remove punctuation to normalize them for comparison. The two dict-valued
 constants (``CAPITALIZATION_EXCEPTIONS`` and ``REGEXES``) are edited with

docs/release_log.rst

@@ -5,6 +5,9 @@ Release Log
     **Breaking Changes & Deprecations**
 
     - Deprecate ``HumanName.__eq__`` and ``__hash__`` for removal in 2.0 (#223): the current design's three promises — case-insensitive equality, equality with plain strings, and hashability — are mutually inconsistent (equal objects can hash differently), equality depends on ``string_format``, and ``maiden`` is invisible to it. Both now emit ``DeprecationWarning`` naming the replacement; behavior is otherwise unchanged until 2.0 (closes #224)
+    - Deprecate ``bytes`` input for removal in 2.0 (#245): passing ``bytes`` to ``HumanName``/``full_name`` or to ``SetManager.add()``/``add_with_encoding()`` now emits ``DeprecationWarning`` — decode first, e.g. ``value.decode('utf-8')``. The ``encoding`` constructor argument is deprecated with it
+    - Deprecate ``SetManager.__call__`` for removal in 2.0 (#243): calling a manager returns the raw underlying set, so mutating the result bypasses normalization and cache invalidation; iterate the manager or copy with ``set(manager)`` instead
+    - Add ``SetManager.discard()``, and deprecate ``remove()`` of a *missing* member (#243): it currently does nothing but will raise ``KeyError`` in 2.0, matching ``set.remove``; use ``discard()`` for intentional ignore-missing removal. Removing present members is unchanged and does not warn
     - Fix ``HumanName`` acting as its own iterator with a stored cursor: breaking out of a loop, iterating in nested loops, or calling ``len(name)`` mid-loop corrupted subsequent iteration; ``iter(name)`` now returns a fresh independent iterator each time. ``next(name)`` on the instance itself (undocumented) now raises ``TypeError`` — call ``next(iter(name))`` instead (closes #225)
     - Remove the vestigial ``unparsable`` attribute: the guard that was meant to set it has been unreachable since 2013 (v0.2.9), so it has reported ``False`` for every parsed name for over a decade; check ``len(name) == 0`` to detect an empty parse
     - Remove ``__ne__``; Python 3 derives ``!=`` from ``__eq__`` automatically

nameparser/config/__init__.py

@@ -28,6 +28,7 @@
 """
 import re
 import sys
+import warnings
 from collections.abc import Callable, Iterable, Iterator, Mapping, Set
 from typing import Any, TypeVar, overload
 
@@ -137,6 +138,20 @@ def __init__(self, elements: Iterable[str]) -> None:
         self._on_change = None
 
     def __call__(self) -> Set[str]:
+        """
+        .. deprecated:: 1.3.0
+            Removed in 2.0 (see issue #243). Returns the raw underlying set,
+            so mutating it bypasses normalization and cache invalidation;
+            iterate the manager or copy with ``set(manager)`` instead.
+        """
+        warnings.warn(
+            "Calling a SetManager to get the raw underlying set is "
+            "deprecated and will be removed in 2.0; iterate the manager or "
+            "copy it with set(manager) instead. See "
+            "https://github.com/derek73/python-nameparser/issues/243",
+            DeprecationWarning,
+            stacklevel=2,
+        )
         return self.elements
 
     def __repr__(self) -> str:
@@ -193,38 +208,90 @@ def __xor__(self, other: Iterable[str]) -> 'SetManager':  # type: ignore[overrid
 
     __rxor__ = __xor__
 
-    def add_with_encoding(self, s: str | bytes, encoding: str | None = None) -> None:
-        """
-        Add the lowercased, leading/trailing-periods-stripped version of the string to the set. Pass an
-        explicit `encoding` parameter to specify the encoding of binary strings that
-        are not DEFAULT_ENCODING (UTF-8).
-        """
+    def _add_normalized(self, s: str | bytes, encoding: str | None, *, stacklevel: int) -> None:
+        # Shared by add() and add_with_encoding() so each can call it
+        # directly with a stacklevel that attributes the warning to *its own*
+        # caller -- add() delegating to add_with_encoding() would otherwise
+        # add a frame and misattribute the warning to this module.
         stdin_encoding = None
         if sys.stdin:
             stdin_encoding = sys.stdin.encoding
         encoding = encoding or stdin_encoding or DEFAULT_ENCODING
         if isinstance(s, bytes):
+            warnings.warn(
+                "Passing bytes to SetManager.add()/add_with_encoding() is "
+                "deprecated and will raise TypeError in 2.0; decode it "
+                "first, e.g. value.decode('utf-8'). See "
+                "https://github.com/derek73/python-nameparser/issues/245",
+                DeprecationWarning,
+                stacklevel=stacklevel,
+            )
             s = s.decode(encoding)
         normalized = lc(s)
         if normalized not in self.elements:
             self.elements.add(normalized)
             if self._on_change:
                 self._on_change()
 
+    def add_with_encoding(self, s: str | bytes, encoding: str | None = None) -> None:
+        """
+        Add the lowercased, leading/trailing-periods-stripped version of the string to the set. Pass an
+        explicit `encoding` parameter to specify the encoding of binary strings that
+        are not DEFAULT_ENCODING (UTF-8).
+
+        .. deprecated:: 1.3.0
+            ``bytes`` arguments will raise ``TypeError`` in 2.0 (see issue
+            #245); decode before adding.
+        """
+        self._add_normalized(s, encoding, stacklevel=3)
+
     def add(self, *strings: str) -> Self:
         """
         Add the lowercased, leading/trailing-periods-stripped version of the string arguments to the set.
         Returns ``self`` for chaining.
+
+        .. deprecated:: 1.3.0
+            ``bytes`` arguments will raise ``TypeError`` in 2.0 (see issue
+            #245); decode before adding.
         """
         for s in strings:
-            self.add_with_encoding(s)
+            self._add_normalized(s, None, stacklevel=3)
 
         return self
 
     def remove(self, *strings: str) -> Self:
         """
         Remove the lower case and no-period version of the string arguments from the set.
         Returns ``self`` for chaining.
+
+        .. deprecated:: 1.3.0
+            Removing a *missing* member currently does nothing but will
+            raise ``KeyError`` in 2.0, matching ``set.remove`` (see issue
+            #243); use :py:func:`discard` to ignore missing members.
+        """
+        changed = False
+        for s in strings:
+            if (lower := lc(s)) in self.elements:
+                self.elements.remove(lower)
+                changed = True
+            else:
+                warnings.warn(
+                    "SetManager.remove() of a missing member currently does "
+                    "nothing, but will raise KeyError in 2.0; use discard() "
+                    "to ignore missing members. See "
+                    "https://github.com/derek73/python-nameparser/issues/243",
+                    DeprecationWarning,
+                    stacklevel=2,
+                )
+        if changed and self._on_change:
+            self._on_change()
+        return self
+
+    def discard(self, *strings: str) -> Self:
+        """
+        Remove the lower case and no-period version of the string arguments
+        from the set if present; missing members are ignored, like
+        ``set.discard``. Returns ``self`` for chaining.
         """
         changed = False
         for s in strings:

nameparser/parser.py

@@ -54,6 +54,8 @@ class HumanName:
         honored). Pass ``None`` for `per-instance config <customize.html>`_.
         Anything else raises ``TypeError``.
     :param str encoding: string representing the encoding of your input
+        (deprecated with ``bytes`` input, removal in 2.0 — decode before
+        passing; see issue #245)
     :param str string_format: python string formatting
     :param str initials_format: python initials string formatting
     :param str initials_delimter: string delimiter for initials
@@ -135,8 +137,10 @@ def __init__(
             self.nickname = nickname
             self.maiden = maiden
         else:
-            # full_name setter triggers the parse
-            self.full_name = full_name
+            # calls _apply_full_name directly (not the setter) so the
+            # deprecation warning below attributes to this constructor's
+            # caller rather than to the setter
+            self._apply_full_name(full_name, stacklevel=3)
 
     @staticmethod
     def _validate_constants(constants: 'Constants | None') -> 'Constants':
@@ -886,9 +890,25 @@ def full_name(self) -> str:
 
     @full_name.setter
     def full_name(self, value: str | bytes) -> None:
+        self._apply_full_name(value, stacklevel=3)
+
+    def _apply_full_name(self, value: str | bytes, *, stacklevel: int) -> None:
+        # Shared by the setter and the constructor so each can call it
+        # directly with a stacklevel that attributes the warning to *its own*
+        # caller -- the constructor going through the setter would otherwise
+        # add a frame and misattribute the warning to this module.
         self.original = value
 
         if isinstance(value, bytes):
+            # deprecated 1.3.0, raises TypeError in 2.0 (#245)
+            warnings.warn(
+                "Passing bytes to HumanName is deprecated and will raise "
+                "TypeError in 2.0; decode it first, e.g. "
+                "value.decode('utf-8'). See "
+                "https://github.com/derek73/python-nameparser/issues/245",
+                DeprecationWarning,
+                stacklevel=stacklevel,
+            )
             self._full_name = value.decode(self.encoding)
         else:
             self._full_name = value

tests/test_constants.py

@@ -2,6 +2,7 @@
 import pickle
 import re
 import timeit
+import warnings
 from typing import Any
 
 import pytest
@@ -332,10 +333,79 @@ def test_none_empty_attribute_string_formatting(self) -> None:
         self.assertEqual('', str(hn), hn)
 
     def test_add_constant_with_explicit_encoding(self) -> None:
+        # bytes input is deprecated (#245), still supported until 2.0
         c = Constants()
-        c.titles.add_with_encoding(b'b\351ck', encoding='latin_1')
+        with pytest.deprecated_call():
+            c.titles.add_with_encoding(b'b\351ck', encoding='latin_1')
         self.assertIn('béck', c.titles)
 
+    def test_set_manager_add_bytes_emits_deprecation_warning(self) -> None:
+        # bytes elements will be removed in 2.0 (#245); the caller should decode
+        sm = SetManager(['dr'])
+        with pytest.deprecated_call(match="decode"):
+            sm.add(b'esq')  # type: ignore[arg-type]  # deliberately deprecated input
+        self.assertIn('esq', sm)
+
+    def test_set_manager_add_str_does_not_warn(self) -> None:
+        sm = SetManager(['dr'])
+        with warnings.catch_warnings():
+            warnings.simplefilter("error")
+            sm.add('esq')
+        self.assertIn('esq', sm)
+
+    def test_set_manager_call_emits_deprecation_warning(self) -> None:
+        # __call__ hands out the raw underlying set, bypassing normalization
+        # and cache invalidation; will be removed in 2.0 (#243)
+        sm = SetManager(['dr'])
+        with pytest.deprecated_call(match="raw underlying set"):
+            elements = sm()
+        self.assertEqual(elements, {'dr'})
+
+    def test_set_manager_discard_ignores_missing_without_warning(self) -> None:
+        sm = SetManager(['dr', 'mr'])
+        with warnings.catch_warnings():
+            warnings.simplefilter("error")
+            result = sm.discard('nope').discard('Dr.')  # normalizes like remove()
+        self.assertIs(result, sm)
+        self.assertEqual(set(sm), {'mr'})
+
+    def test_set_manager_discard_invalidates_cached_union(self) -> None:
+        c = Constants()
+        self.assertIn('hon', c.suffixes_prefixes_titles)  # prime the cache
+        c.titles.discard('hon')
+        self.assertNotIn('hon', c.suffixes_prefixes_titles)
+
+    def test_set_manager_remove_missing_member_emits_deprecation_warning(self) -> None:
+        # ignore-missing remove() becomes KeyError in 2.0 (#243); discard()
+        # is the intentional ignore-missing spelling
+        sm = SetManager(['dr'])
+        with pytest.deprecated_call(match="discard"):
+            sm.remove('nope')
+        self.assertEqual(set(sm), {'dr'})
+        # removing a present member stays silent
+        with warnings.catch_warnings():
+            warnings.simplefilter("error")
+            sm.remove('dr')
+        self.assertEqual(len(sm), 0)
+
+    def test_set_manager_remove_mixed_present_and_missing_in_one_call(self) -> None:
+        # a single call mixing a present and a missing member must still warn
+        # (for the missing one) and still invalidate the cache (for the
+        # present one) -- not short-circuit either behavior
+        c = Constants()
+        self.assertIn('hon', c.suffixes_prefixes_titles)  # prime the cache
+        with pytest.deprecated_call(match="discard"):
+            c.titles.remove('hon', 'nope')
+        self.assertNotIn('hon', c.suffixes_prefixes_titles)
+
+    def test_set_manager_discard_mixed_present_and_missing_in_one_call(self) -> None:
+        sm = SetManager(['dr', 'mr'])
+        with warnings.catch_warnings():
+            warnings.simplefilter("error")
+            result = sm.discard('dr', 'nope')
+        self.assertIs(result, sm)
+        self.assertEqual(set(sm), {'mr'})
+
     def test_pickle_roundtrip_preserves_customizations(self) -> None:
         """A pickled Constants must restore its customized collections.
 
@@ -622,7 +692,8 @@ def test_suffixes_prefixes_titles_reflects_add_with_encoding(self) -> None:
         """add_with_encoding must invalidate the cache like add()/remove() do."""
         c = Constants()
         _ = c.suffixes_prefixes_titles  # prime the cache
-        c.titles.add_with_encoding(b'b\351ck', encoding='latin_1')
+        with pytest.deprecated_call():  # bytes input deprecated (#245)
+            c.titles.add_with_encoding(b'b\351ck', encoding='latin_1')
         self.assertIn('béck', c.suffixes_prefixes_titles)
 
     def test_suffixes_prefixes_titles_reflects_replaced_manager(self) -> None:

tests/test_python_api.py

@@ -28,10 +28,29 @@ def test_string_output(self) -> None:
         self.m(str(hn), "Jüan de la Véña", hn)
 
     def test_escaped_utf8_bytes(self) -> None:
-        hn = HumanName(b'B\xc3\xb6ck, Gerald')
+        # bytes input is deprecated (#245), still supported until 2.0
+        with pytest.deprecated_call():
+            hn = HumanName(b'B\xc3\xb6ck, Gerald')
         self.m(hn.first, "Gerald", hn)
         self.m(hn.last, "Böck", hn)
 
+    def test_bytes_full_name_emits_deprecation_warning(self) -> None:
+        # bytes input will be removed in 2.0 (#245); the caller should decode
+        with pytest.deprecated_call(match="decode"):
+            hn = HumanName(b'John Smith')
+        self.m(hn.first, "John", hn)
+        hn2 = HumanName("Jane Doe")
+        with pytest.deprecated_call(match="decode"):
+            hn2.full_name = b'John Smith'
+        self.m(hn2.first, "John", hn2)
+
+    def test_str_full_name_does_not_warn(self) -> None:
+        with warnings.catch_warnings():
+            warnings.simplefilter("error")
+            hn = HumanName("John Smith")
+            hn.full_name = "Jane Doe"
+        self.m(hn.first, "Jane", hn)
+
     def test_len(self) -> None:
         hn = HumanName("Doe-Ray, Dr. John P., CLU, CFP, LUTC")
         self.m(len(hn), 5, hn)