modernize with pyugprade

Author: fgregg

dedupe/api.py

@@ -867,7 +867,7 @@ def blocks(self, data):
         con.execute("ROLLBACK")
         con.close()
 
-    def score(self, blocks: Blocks) -> Generator[Scores, None, None]:
+    def score(self, blocks: Blocks) -> Generator[Scores]:
         """
         Scores groups of pairs of records. Yields structured numpy arrays
         representing pairs of records in the group and the associated

dedupe/blocking.py

@@ -57,7 +57,7 @@ def __init__(self, predicates: Iterable[dedupe.predicates.Predicate]) -> None:
 
     def __call__(
         self, records: Iterable[Record], target: bool = False
-    ) -> Generator[tuple[str, RecordID], None, None]:
+    ) -> Generator[tuple[str, RecordID]]:
         """
         Generate the predicates for records. Yields tuples of (predicate,
         record_id).

dedupe/clustering.py

@@ -17,9 +17,7 @@
 logger = logging.getLogger(__name__)
 
 
-def connected_components(
-    edgelist: Scores, max_components: int
-) -> Generator[Scores, None, None]:
+def connected_components(edgelist: Scores, max_components: int) -> Generator[Scores]:
     if len(edgelist) == 0:
         raise StopIteration()
 
@@ -51,9 +49,7 @@ def connected_components(
         edgelist._mmap.close()  # type: ignore
 
 
-def _connected_components(
-    edgelist: Scores, max_components: int
-) -> Generator[Scores, None, None]:
+def _connected_components(edgelist: Scores, max_components: int) -> Generator[Scores]:
     component_stops = union_find(edgelist)
 
     start = 0

dedupe/convenience.py

@@ -161,7 +161,7 @@ def console_label(deduper: dedupe.api.ActiveMatching) -> None:  # pragma: no cov
 
         for record in record_pair:
             for field in fields:
-                line = "{} : {}".format(field, record[field])
+                line = f"{field} : {record[field]}"
                 _print(line)
             _print()
         _print(f"{n_match}/10 positive, {n_distinct}/10 negative")

dedupe/core.py

@@ -218,7 +218,7 @@ def scoreGazette(
     featurizer: FeaturizerFunction,
     classifier: Classifier,
     num_cores: int = 1,
-) -> Generator[Scores, None, None]:
+) -> Generator[Scores]:
     first, record_pairs = peek(record_pairs)
     if first is None:
         return  # terminate iteration

dedupe/datamodel.py

@@ -86,7 +86,7 @@ def __len__(self) -> int:
     @property
     def _field_comparators(
         self,
-    ) -> Generator[tuple[str, Comparator, int, int], None, None]:
+    ) -> Generator[tuple[str, Comparator, int, int]]:
         start = 0
         stop = 0
         for var in self.field_variables:

dedupe/predicates.py

@@ -50,7 +50,7 @@ def __iter__(self):
         yield self
 
     def __repr__(self) -> str:
-        return "{}: {}".format(self.type, self.__name__)
+        return f"{self.type}: {self.__name__}"
 
     def __hash__(self) -> int:
         try:
@@ -83,7 +83,7 @@ class SimplePredicate(Predicate):
 
     def __init__(self, func: PredicateFunction, field: str):
         self.func = func
-        self.__name__ = "({}, {})".format(func.__name__, field)
+        self.__name__ = f"({func.__name__}, {field})"
         self.field = field
 
     def __call__(self, record: RecordDict, **kwargs) -> frozenset[str]:
@@ -107,7 +107,7 @@ class ExistsPredicate(Predicate):
     type = "ExistsPredicate"
 
     def __init__(self, field: str):
-        self.__name__ = "(Exists, {})".format(field)
+        self.__name__ = f"(Exists, {field})"
         self.field = field
 
     @staticmethod
@@ -129,7 +129,7 @@ class IndexPredicate(Predicate):
     _cache: dict[Any, frozenset[str]]
 
     def __init__(self, threshold: float, field: str):
-        self.__name__ = "({}, {})".format(threshold, field)
+        self.__name__ = f"({threshold}, {field})"
         self.field = field
         self.threshold = threshold
         self.index = None

dedupe/variables/base.py

@@ -44,7 +44,7 @@ class DerivedType(Variable):
     type = "Derived"
 
     def __init__(self, name: str, var_type: str, **kwargs):
-        self.name = "({}: {})".format(str(name), str(var_type))
+        self.name = f"({str(name)}: {str(var_type)})"
         super().__init__(**kwargs)
 
 
@@ -59,7 +59,7 @@ def __init__(self, field: str, name: str | None = None, has_missing: bool = Fals
         self.field = field
 
         if name is None:
-            self.name = "({}: {})".format(self.field, self.type)
+            self.name = f"({self.field}: {self.type})"
         else:
             self.name = name