nomenklatura-mpt 4.1.12__tar.gz → 4.1.14__tar.gz

{nomenklatura_mpt-4.1.12 → nomenklatura_mpt-4.1.14}/PKG-INFO

@@ -1,39 +1,19 @@
 Metadata-Version: 2.4
 Name: nomenklatura_mpt
-Version: 4.1.12
+Version: 4.1.14
 Summary: Make record linkages in followthemoney data.
 Project-URL: Documentation, https://github.com/opensanctions/nomenklatura/
 Project-URL: Repository, https://github.com/opensanctions/nomenklatura.git
 Project-URL: Issues, https://github.com/opensanctions/nomenklatura/issues
 Author-email: OpenSanctions <info@opensanctions.org>
-License: Copyright (c) 2013-2022, Friedrich Lindenberg
-        Copyright (c) 2023-2025, OpenSanctions Datenbanken GmbH
-        
-        Permission is hereby granted, free of charge, to any person obtaining a
-        copy of this software and associated documentation files (the
-        "Software"), to deal in the Software without restriction, including
-        without limitation the rights to use, copy, modify, merge, publish,
-        distribute, sublicense, and/or sell copies of the Software, and to
-        permit persons to whom the Software is furnished to do so, subject to
-        the following conditions:
-        
-        The above copyright notice and this permission notice shall be included
-        in all copies or substantial portions of the Software.
-        
-        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
-        OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
-        MERCHANTABILITY, 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.
+License: MIT
 License-File: LICENSE
 Classifier: Intended Audience :: Developers
 Classifier: License :: OSI Approved :: MIT License
 Classifier: Operating System :: OS Independent
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
-Requires-Python: >=3.9
+Requires-Python: <3.14,>=3.11
 Requires-Dist: click
 Requires-Dist: fingerprints
 Requires-Dist: followthemoney

{nomenklatura_mpt-4.1.12 → nomenklatura_mpt-4.1.14}/nomenklatura/matching/logic_v2/names/match.py

@@ -5,7 +5,7 @@ from rigour.names import remove_obj_prefixes
 from followthemoney.proxy import E, EntityProxy
 from followthemoney import model
 from followthemoney.types import registry
-from followthemoney.names import schema_type_tag
+from rigour.names import schema_type_tag
 
 from nomenklatura.matching.logic_v2.names.analysis import entity_names
 from nomenklatura.matching.logic_v2.names.magic import weight_extra_match
@@ -198,15 +198,74 @@ def name_match(query: E, result: E, config: ScoringConfig) -> FtResult:
 
 
 def name_match_levenshtein(query: E, result: E, config: ScoringConfig) -> FtResult:
-    """Match two entities by analyzing and comparing their names."""
+    """Match two entities using Levenshtein distance on their normalized names.
+
+    This function compares entities by computing the Levenshtein distance between
+    all pairs of their names and returns the best match score.
+
+    Args:
+        query: The query entity to match
+        result: The result entity to match against
+        config: Scoring configuration parameters
+
+    Returns:
+        FtResult with score (0.0-1.0) and detail string explaining the match
+    """
     schema = model.common_schema(query.schema, result.schema)
     type_tag = schema_type_tag(schema)
     best = FtResult(score=0.0, detail=None)
+
     if type_tag == NameTypeTag.UNK:
         # Name matching is not supported for entities that are not listed
         # as a person, organization, or a thing.
         best.detail = "Unsuited for name matching: %s" % schema.name
         return best
-    
-    return match_object_names(query, result, config)
+
+    # For object types (vessels, assets, etc.), use the existing object name matching
+    if type_tag == NameTypeTag.OBJ:
+        return match_object_names(query, result, config)
+
+    # Get all names for both entities
+    query_names = entity_names(type_tag, query, is_query=True)
+    result_names = entity_names(type_tag, result)
+
+    if not query_names or not result_names:
+        best.detail = "No names available for matching"
+        return best
+
+    # Check for literal matches first (early return for efficiency)
+    query_comparable = {name.comparable: name for name in query_names}
+    result_comparable = {name.comparable: name for name in result_names}
+    common = set(query_comparable).intersection(result_comparable)
+    if len(common) > 0:
+        longest = max(common, key=len)
+        best.detail = f"[{longest!r} literalMatch]"
+        return FtResult(score=1.0, detail=best.detail)
+
+    # Consolidate names (remove short names contained in longer names)
+    query_names = Name.consolidate_names(query_names)
+    result_names = Name.consolidate_names(result_names)
+
+    # Compare all pairs of names using Levenshtein distance
+    for query_name in query_names:
+        for result_name in result_names:
+            # Get the comparable forms (normalized strings)
+            query_str = query_name.comparable
+            result_str = result_name.comparable
+
+            # Use strict_levenshtein which already implements a good scoring mechanism
+            # max_rate=4 means we allow up to len/4 edits
+            score = strict_levenshtein(query_str, result_str, max_rate=4)
+
+            if score > best.score:
+                best.score = score
+                if score == 1.0:
+                    best.detail = f"[{query_str!r}≡{result_str!r}, exactMatch]"
+                else:
+                    best.detail = f"[{query_str!r}≈{result_str!r}, levenshteinScore: {score:.4f}]"
+
+    if best.detail is None:
+        best.detail = "No sufficient Levenshtein match found"
+
+    return best
    

{nomenklatura_mpt-4.1.12 → nomenklatura_mpt-4.1.14}/nomenklatura/matching/logic_v4/model.py

@@ -1,7 +1,7 @@
 from typing import Dict, List
 
-from nomenklatura.matching.logic_v1.phonetic import name_soundex_match, person_name_phonetic_match
-from nomenklatura.matching.logic_v4.phonetic import name_metaphone_match
+from nomenklatura.matching.logic_v1.identifiers import orgid_disjoint
+from nomenklatura.matching.logic_v3.multi import numbers_mismatch
 from nomenklatura.matching.types import Feature, FtResult, HeuristicAlgorithm
 from nomenklatura.matching.types import ConfigVar, ConfigVarType
 from nomenklatura.matching.compare.countries import country_mismatch
@@ -9,17 +9,14 @@ from nomenklatura.matching.compare.gender import gender_mismatch
 from nomenklatura.matching.compare.identifiers import crypto_wallet_address
 from nomenklatura.matching.compare.identifiers import identifier_match
 from nomenklatura.matching.compare.dates import dob_day_disjoint, dob_year_disjoint
-from nomenklatura.matching.compare.names import weak_alias_match
+from nomenklatura.matching.compare.names import last_name_mismatch, weak_alias_match
 from nomenklatura.matching.compare.addresses import address_entity_match
-from nomenklatura.matching.compare.addresses import address_prop_match
 from nomenklatura.matching.logic_v2.names.match import name_match_levenshtein
 from nomenklatura.matching.logic_v2.identifiers import bic_code_match
 from nomenklatura.matching.logic_v2.identifiers import inn_code_match, ogrn_code_match
 from nomenklatura.matching.logic_v2.identifiers import isin_security_match
 from nomenklatura.matching.logic_v2.identifiers import lei_code_match
 from nomenklatura.matching.logic_v2.identifiers import vessel_imo_mmsi_match
-from nomenklatura.matching.logic_v2.identifiers import uei_code_match
-from nomenklatura.matching.logic_v2.identifiers import npi_code_match
 from nomenklatura.matching.util import FNUL
 
 
@@ -34,10 +31,6 @@ class LogicV4(HeuristicAlgorithm):
     NAME = "logic-v4"
     features = [
         Feature(func=name_match_levenshtein, weight=1.0),
-        Feature(func=FtResult.wrap(person_name_phonetic_match), weight=0.9),
-        # These are there so they can be enabled using custom weights:
-        Feature(func=FtResult.wrap(name_metaphone_match), weight=FNUL),
-        Feature(func=FtResult.wrap(name_soundex_match), weight=FNUL),
         Feature(func=address_entity_match, weight=0.98),
         Feature(func=crypto_wallet_address, weight=0.98),
         Feature(func=isin_security_match, weight=0.98),
@@ -46,15 +39,15 @@ class LogicV4(HeuristicAlgorithm):
         Feature(func=vessel_imo_mmsi_match, weight=0.95),
         Feature(func=inn_code_match, weight=0.95),
         Feature(func=bic_code_match, weight=0.95),
-        Feature(func=uei_code_match, weight=0.95),
-        Feature(func=npi_code_match, weight=0.95),
         Feature(func=identifier_match, weight=0.85),
         Feature(func=weak_alias_match, weight=0.8),
-        Feature(func=address_prop_match, weight=0.2, qualifier=True),
         Feature(func=country_mismatch, weight=-0.2, qualifier=True),
+        Feature(func=FtResult.wrap(last_name_mismatch), weight=-0.2, qualifier=True),
         Feature(func=dob_year_disjoint, weight=-0.15, qualifier=True),
-        Feature(func=dob_day_disjoint, weight=-0.25, qualifier=True),
+        Feature(func=dob_day_disjoint, weight=-0.2, qualifier=True),
         Feature(func=gender_mismatch, weight=-0.2, qualifier=True),
+        Feature(func=orgid_disjoint, weight=-0.2, qualifier=True),
+        Feature(func=numbers_mismatch, weight=-0.1, qualifier=True),
     ]
     CONFIG = {
         "nm_number_mismatch": ConfigVar(

{nomenklatura_mpt-4.1.12 → nomenklatura_mpt-4.1.14}/pyproject.toml

@@ -4,10 +4,10 @@ build-backend = "hatchling.build"
 
 [project]
 name = "nomenklatura_mpt"
-version = "4.1.12"
+version = "4.1.14"
 description = "Make record linkages in followthemoney data."
 readme = "README.md"
-license = { file = "LICENSE" }
+license = { text = "MIT" }
 authors = [{ name = "OpenSanctions", email = "info@opensanctions.org" }]
 classifiers = [
     "Intended Audience :: Developers",
@@ -16,7 +16,7 @@ classifiers = [
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
 ]
-requires-python = ">= 3.9"
+requires-python = ">=3.11,<3.14"
 dependencies = [
     "followthemoney",
     "rigour",
@@ -66,7 +66,7 @@ leveldb = ["plyvel < 2.0.0"]
 redis = ["redis > 5.0.0, < 7.0.0"]
 
 [tool.hatch.build.targets.sdist]
-only-include = ["nomenklatura", "LICENSE", "README.md"]
+only-include = ["nomenklatura", "README.md"]
 
 [tool.hatch.build.targets.wheel]
 packages = ["nomenklatura"]