followthemoney 4.3.0__py3-none-any.whl → 4.3.1__py3-none-any.whl

followthemoney/__init__.py

@@ -9,7 +9,7 @@ from followthemoney.statement import Statement, StatementEntity, SE
 from followthemoney.dataset import Dataset, DefaultDataset, DS
 from followthemoney.util import set_model_locale
 
-__version__ = "4.3.0"
+__version__ = "4.3.1"
 
 # Data model singleton
 model = Model.instance()

followthemoney/compare.py

@@ -71,12 +71,31 @@ def _compare(scores: Scores, weights: Weights, n_std: int = 1) -> float:
     return 1.0 / (1.0 + math.exp(-prob))
 
 
+def entity_is_same(left: EntityProxy, right: EntityProxy) -> bool:
+    """Check if two entities are the same apart from their ID."""
+    if left.schema != right.schema:
+        return False
+
+    props = set(left.properties.keys()).union(right.properties.keys())
+    if 0 == len(props):
+        return False
+
+    for prop in props:
+        left_vals = sorted(left.get(prop))
+        right_vals = sorted(right.get(prop))
+        if left_vals != right_vals:
+            return False
+    return True
+
+
 def compare(
     left: EntityProxy,
     right: EntityProxy,
     weights: Weights = COMPARE_WEIGHTS,
 ) -> float:
     """Compare two entities and return a match score."""
+    if entity_is_same(left, right):
+        return 1.0
     scores = compare_scores(left, right)
     return _compare(scores, weights)
 

followthemoney/property.py

@@ -1,5 +1,6 @@
 import re
 from banal import is_mapping, as_bool
+from rigour.ids import get_identifier_format
 from typing import TYPE_CHECKING, Any, List, Optional, TypedDict
 
 from followthemoney.exc import InvalidModel
@@ -157,6 +158,11 @@ class Property:
                     raise InvalidModel("Invalid reverse: %s" % self)
                 self.reverse = self.range._add_reverse(model, self._reverse, self)
 
+        if self.type == registry.identifier and self.format is not None:
+            format_ = get_identifier_format(self.format)
+            if format_ is None or format_.NAME != self.format:
+                raise InvalidModel("Invalid identifier format: %s" % self.format)
+
     @property
     def label(self) -> str:
         """User-facing title for this property."""

followthemoney/proxy.py

@@ -462,7 +462,7 @@ class EntityProxy(object):
         return self._size
 
     def __hash__(self) -> int:
-        if not self.id:
+        if self.id is None:
             raise RuntimeError("Cannot hash entity without an ID")
         return hash(self.id)
 

followthemoney/schema/Thing.yaml

@@ -55,7 +55,7 @@ Thing:
     wikidataId:
       label: Wikidata ID
       type: identifier
-      format: qid
+      format: wikidata
       maxLength: 32
     keywords:
       label: Keywords

followthemoney/types/identifier.py

@@ -1,6 +1,6 @@
 import re
 from typing import Optional, TYPE_CHECKING
-from rigour.ids import get_identifier_format_names, get_identifier_format
+from rigour.ids import get_identifier_format
 
 from followthemoney.types.common import PropertyType
 from followthemoney.util import dampen, shortest, longest
@@ -35,8 +35,8 @@ class IdentifierType(PropertyType):
         format: Optional[str] = None,
         proxy: Optional["EntityProxy"] = None,
     ) -> Optional[str]:
-        if format in get_identifier_format_names():
-            format_ = get_identifier_format(format)
+        format_ = get_identifier_format(format)
+        if format_ is not None:
             return format_.normalize(text)
         return text
 
@@ -61,7 +61,7 @@ class IdentifierType(PropertyType):
         return f"id:{value}"
 
     def caption(self, value: str, format: Optional[str] = None) -> str:
-        if format in get_identifier_format_names():
-            format_ = get_identifier_format(format)
+        format_ = get_identifier_format(format)
+        if format_ is not None:
             return format_.format(value)
         return value

{followthemoney-4.3.0.dist-info → followthemoney-4.3.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: followthemoney
-Version: 4.3.0
+Version: 4.3.1
 Summary: A data model for anti corruption data modeling and analysis.
 Project-URL: Documentation, https://followthemoney.tech/
 Project-URL: Repository, https://github.com/opensanctions/followthemoney.git
@@ -50,7 +50,7 @@ Requires-Dist: pytz>=2021.1
 Requires-Dist: pyyaml<7.0.0,>=5.0.0
 Requires-Dist: rdflib<7.3.0,>=6.2.0
 Requires-Dist: requests<3.0.0,>=2.21.0
-Requires-Dist: rigour<2.0.0,>=1.3.13
+Requires-Dist: rigour<2.0.0,>=1.4.0
 Requires-Dist: sqlalchemy[mypy]<3.0.0,>=2.0.0
 Provides-Extra: dev
 Requires-Dist: build; extra == 'dev'

{followthemoney-4.3.0.dist-info → followthemoney-4.3.1.dist-info}/RECORD

@@ -1,5 +1,5 @@
-followthemoney/__init__.py,sha256=SGtXmqy22FtLMX0kcGDSUMbWNGUvoA13F5SUobauMg8,856
-followthemoney/compare.py,sha256=bZlnj2VMoe67q4Lyq_VwS1a-EJnEK1kC8prbs8jyL9E,5774
+followthemoney/__init__.py,sha256=9UJzMQUcqjQ4kWQIDMF1GB0ji8ph-GQlXa3e_X2uuX4,856
+followthemoney/compare.py,sha256=frgumsDv4Ru9UkNof62jDjKCxxpCgV1Rusfu8s20uGA,6327
 followthemoney/entity.py,sha256=bBiX7hNquXemS3vYCUHKtWI_IqX43Z6i8RQDbZ7gXsg,3449
 followthemoney/exc.py,sha256=GyMgwY4QVm87hLevDfV7gM1MJsDqfNCi_UQw7F_A8X8,858
 followthemoney/graph.py,sha256=7X1CGHGvmktS2LSZqld2iXWzG7B831eCNYyBqamqEJ8,10921
@@ -9,8 +9,8 @@ followthemoney/model.py,sha256=bWFVNa-DhYzc8BdSXBZdG2ev6Nh9uHx6i4tin8DvEEU,7374
 followthemoney/names.py,sha256=LODQqExKEHdH4z6Mmbhlm0KeKRzGcptaSWzYXZ7lONI,1120
 followthemoney/namespace.py,sha256=utggu9IGA8bhgEYom3OUB1KxkAJR_TrMNbY5MUF_db8,4536
 followthemoney/ontology.py,sha256=WWY_PYQGl5Ket4zZBuZglzQxD2Bh9UqHok6GJNNX7GA,3001
-followthemoney/property.py,sha256=9qZ_o2iA-1llLMJ3O2hsW7c2XhkFU1YbvVqretGYUSA,7913
-followthemoney/proxy.py,sha256=5pMyP0JWWuBcvMIu3AXlW6Za7EEWWdFD2EZkD8UkbQw,19703
+followthemoney/property.py,sha256=6FoKoUloDoQVM4XMnjjolKArlV4yBq9KOYxv8Os8WGQ,8234
+followthemoney/proxy.py,sha256=nahd9lLZzum_-QEchqydinDa1Zg5_Ffl0fyo35BNncQ,19707
 followthemoney/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 followthemoney/schema.py,sha256=WYnPE4Lego0pJHlojECEv0aO9Miw_YIvEb35HoDo4Zk,18087
 followthemoney/util.py,sha256=LoCSp1iE6VwXjotCkBXFRppeQs55726GzOuNIu3CvRE,4409
@@ -104,7 +104,7 @@ followthemoney/schema/Similar.yaml,sha256=gD8rZEaPQWzU-rEfsKdn62uEucF3KxYBcPMoSd
 followthemoney/schema/Succession.yaml,sha256=RMJQqZ4Fv88N1RvWTAgjYg9BB5cELSj5CCAjM681Fpg,749
 followthemoney/schema/Table.yaml,sha256=GcsIAgSO9t2tvObA9zU2HhxlSqTe9CePmUnagu1Z0vI,641
 followthemoney/schema/TaxRoll.yaml,sha256=ugMzaaS7uyq2OLD50eGLcfvd6Cg0cSt65-T9GVqpRSA,746
-followthemoney/schema/Thing.yaml,sha256=hh1oMDQzWiSs1TamBNonmwEdlh2TVrNc3w9hWW8iSeY,2716
+followthemoney/schema/Thing.yaml,sha256=xqGtSzyfE_EWx5PrxETKftF-9NSlsTYIbXrNqhdx9jw,2721
 followthemoney/schema/Trip.yaml,sha256=nLQD_ApmVJ8D56Czl7K700hhNZjzFV9FOQ3NBSQDLiM,771
 followthemoney/schema/UnknownLink.yaml,sha256=lneS_HZNgeLyJxwzWnLx0ZoyY3MXt99I_K2X_o9z5g8,682
 followthemoney/schema/UserAccount.yaml,sha256=2bbPKNtt1R3zWSSkaq_SVzRPfFzX74kAxwtIxTymHA8,840
@@ -150,7 +150,7 @@ followthemoney/types/date.py,sha256=PjcaEyW6CBzf0-gHWKUsKjWIaD3AVBEl0zLSRQOVXxc,
 followthemoney/types/email.py,sha256=L3RTYrMABlNQF7hCynXGfzoj6YNEHW5JAY_BwuhoZdA,3375
 followthemoney/types/entity.py,sha256=oDxVEhuxyU1ScpOpebPpUm3o0I9j_p7Qrq-t5yNpluQ,2338
 followthemoney/types/gender.py,sha256=fi9iKLbjAUxDCLBtU1MxWidxv7KgCY2eH5746FYlEGk,1725
-followthemoney/types/identifier.py,sha256=hzD188FtwG0w3TcmbnDwnUMc8MZVcWgQJKGAvrwygc4,2296
+followthemoney/types/identifier.py,sha256=7YET9mdYAktMup_kTvnxLu9SJzEU7ZJgN3KImzjfT-Y,2219
 followthemoney/types/ip.py,sha256=mMFTODFiXAJROCUYJvoLAShyIiTIWVmMBh5zT_GquYM,1300
 followthemoney/types/json.py,sha256=V3qJD5RxJykNX51u3w1Nx9xqoNBnkulhzkJI9XMYKFo,1690
 followthemoney/types/language.py,sha256=SXgRRH-DyPmyyrqYurSyMiG6WHB8a0Gw81XxroEGD-c,2747
@@ -161,8 +161,8 @@ followthemoney/types/phone.py,sha256=r8uRqWinS0CYnYBTs405k5gO4jeatUDgjdzzijoMKJE
 followthemoney/types/string.py,sha256=fqyTauAm4mNnNaoH-yH087RBbNh-G5ZZUO3awTGQUUg,1230
 followthemoney/types/topic.py,sha256=Mi0Gx0m3bDeTmyuvM6jdRMqv81O03U4eI99R13KGu2Y,4503
 followthemoney/types/url.py,sha256=QFpS_JIV8unFHuh_uGv22SWUUkocBoOpzLsAJWom_gI,1455
-followthemoney-4.3.0.dist-info/METADATA,sha256=dyrwBXiefNsSpmwOcsP-TsEnJBqmn9DY9vJsMhFVwk0,6748
-followthemoney-4.3.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-followthemoney-4.3.0.dist-info/entry_points.txt,sha256=caoFTlf213jhg5sz3TNSofutjUTzaKtWATuSIdd9Cps,653
-followthemoney-4.3.0.dist-info/licenses/LICENSE,sha256=H6_EVXisnJC0-18CjXIaqrBSFq_VH3OnS7u3dccOv6g,1148
-followthemoney-4.3.0.dist-info/RECORD,,
+followthemoney-4.3.1.dist-info/METADATA,sha256=i41B2NtbZrqWsc0F26XIP9GCvbRtMZ1ZsWeRWvk4AAg,6747
+followthemoney-4.3.1.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+followthemoney-4.3.1.dist-info/entry_points.txt,sha256=caoFTlf213jhg5sz3TNSofutjUTzaKtWATuSIdd9Cps,653
+followthemoney-4.3.1.dist-info/licenses/LICENSE,sha256=H6_EVXisnJC0-18CjXIaqrBSFq_VH3OnS7u3dccOv6g,1148
+followthemoney-4.3.1.dist-info/RECORD,,