slob.py: dedup alias targets instead of alias keys, #556

itkach/slob@5a3758f

pyglossary/slob.py

@@ -1305,18 +1305,21 @@ def read_key_frag(item: "Blob", default_fragment: str) -> "tuple[str, str]":
 				alias_writer.finalize()
 
 		with Slob(path) as resolved_aliases_reader:
-			previous_key = None
+			previous = None
+			targets = set()
+
 			for item in resolved_aliases_reader:
 				ref = pickle.loads(item.content)
-				if ref.key == previous_key:
-					continue
-				self._write_ref(
-					ref.key,
-					ref.bin_index,
-					ref.item_index,
-					ref.fragment,
-				)
-				previous_key = ref.key
+				if previous is not None and ref.key != previous.key:
+					for bin_index, item_index, fragment in targets:
+						self._write_ref(previous.key, bin_index, item_index, fragment)
+					targets.clear()
+				targets.add((ref.bin_index, ref.item_index, ref.fragment))
+				previous = ref
+
+			for bin_index, item_index, fragment in targets:
+				self._write_ref(previous.key, bin_index, item_index, fragment)
+
 		self._sort()
 		self._fire_event("end_resolve_aliases")
 

pyproject.toml

@@ -157,7 +157,9 @@ lint.dummy-variable-rgx = "^(_+|(_+[a-zA-Z0-9_]*[a-zA-Z0-9]+?))$"
     "PLR0904", # Too many public methods
     "C90",     # mccabe: C901: {name} is too complex ({complexity})
 ]
-"slob.py" = []
+"slob.py" = [
+    "C901", # `...` is too complex (14 > 13)
+]
 "html_dir.py" = [
     "C901", # `...` is too complex
 ]

tests/slob_test.py

@@ -467,6 +467,7 @@ def observer(event):
 			w.add(v.encode("ascii"), k)
 
 		w.add_alias("w", "u")
+		w.add_alias("small u", "u")
 		w.add_alias("y1", "y2")
 		w.add_alias("y2", "y3")
 		w.add_alias("y3", "z")
@@ -482,6 +483,10 @@ def observer(event):
 
 		w.add_alias("g1", "g")
 		w.add_alias("g2", ("g1", "g-frag1"))
+
+		w.add_alias("n or p", "n")
+		w.add_alias("n or p", "p")
+
 		w.finalize()
 
 		self.assertEqual(too_many_redirects, ["l1", "l2", "l3"])
@@ -494,6 +499,7 @@ def get(key):
 				return [item.content.decode("ascii") for item in d[key]]
 
 			self.assertEqual(get("w"), ["LATIN SMALL LETTER U"])
+			self.assertEqual(get("small u"), ["LATIN SMALL LETTER U"])
 			self.assertEqual(get("y1"), ["LATIN SMALL LETTER Z"])
 			self.assertEqual(get("y2"), ["LATIN SMALL LETTER Z"])
 			self.assertEqual(get("y3"), ["LATIN SMALL LETTER Z"])
@@ -502,6 +508,8 @@ def get(key):
 			self.assertEqual(get("l2"), [])
 			self.assertEqual(get("l3"), [])
 
+			self.assertEqual(len(list(d["n or p"])), 2)
+
 			item_a1 = cast(slob.Blob, next(d["a1"]))
 			self.assertEqual(item_a1.content, b"LATIN SMALL LETTER A")
 			self.assertEqual(item_a1.fragment, "a-frag1")