A little nicer

.gitignore

@@ -1,4 +1,5 @@
 __pycache__/
+.mypy_cache/
 env/
 build/
 develop-eggs/

pyproject.toml

@@ -8,11 +8,12 @@ version = "0.0.1"
 requires-python = ">= 3.10"
 dependencies = [
   "numpy>=1.19.5",
+  "nptyping>=2.5.0",
 ]
 authors = [
   {name = "Andrew DalPino", email = "[email protected]"},
 ]
-description = "PyBloomer is an Python implementation of the OkBloomer algorithm, an autoscaling Bloom filter with ultra-low memory footprint."
+description = "A Python implementation of the OkBloomer algorithm, an autoscaling Bloom filter with ultra-low memory footprint."
 readme = "README.md"
 license = {text = "MIT"}
 

src/pybloomer/bloom_filter.py → src/okbloomer/bloom_filter.py

@@ -1,4 +1,5 @@
 import numpy as np
+from nptyping import NDArray
 
 class BloomFilter(object):
     """
@@ -8,7 +9,8 @@ class BloomFilter(object):
 
     MAX_SLICE_SIZE = 2147483647
 
-    n = 0 # The number of bits currently stored in the filter
+    n = 0 # The number of bits currently stored in the filter.
+    m = 0 # The maximum number of bits that can be stored in the filter.
 
     def __init__(self,
                 max_false_positive_rate: float = 0.01,
@@ -33,8 +35,9 @@ def __init__(self,
         self.num_hashes = num_hashes
         self.layer_size = layer_size
         self.slice_size = slice_size
-        self.layers = [np.zeros(layer_size, dtype='bool')]
-        self.m = layer_size
+        self.layers: list[NDArray] = []
+
+        self._add_layer()
 
     @property
     def num_layers(self) -> int:
@@ -127,8 +130,7 @@ def exists_or_insert(self, token: str) -> bool:
 
     def _add_layer(self) -> None:
         """
-        Add another layer to the filter for maintaining the false positivity rate
-        below the threshold.
+        Add another layer to the filter for maintaining the false positivity rate below the threshold.
         """
         self.layers.append(np.zeros(self.layer_size, dtype='bool'))
 

tests/test_bloom_filter.py

@@ -2,11 +2,11 @@
 import random
 import string
 
-import pybloomer
+import okbloomer
 
 class TestBloomFilter(unittest.TestCase):
     def test_basic(self):
-        filter = pybloomer.BloomFilter()
+        filter = okbloomer.BloomFilter()
 
         self.assertEqual(filter.false_positive_rate, 0)
 
@@ -27,7 +27,7 @@ def test_basic(self):
         self.assertFalse(filter.exists('baz'))
 
     def test_exists_or_insert(self):
-        filter = pybloomer.BloomFilter()
+        filter = okbloomer.BloomFilter()
 
         self.assertFalse(filter.exists_or_insert('foo'))
 
@@ -42,10 +42,10 @@ def test_exists_or_insert(self):
         self.assertTrue(filter.exists_or_insert('baz'))
 
     def test_autoscaling(self):
-        random.seed(0)
+        random.seed(1)
 
-        filter = pybloomer.BloomFilter(
-            max_false_positive_rate=0.01,
+        filter = okbloomer.BloomFilter(
+            max_false_positive_rate=0.001,
             num_hashes=4,
             layer_size=320000,
         )
@@ -59,10 +59,11 @@ def test_autoscaling(self):
 
         filter.insert('bar')
 
-        self.assertEqual(filter.num_layers, 3)
-        self.assertLessEqual(filter.false_positive_rate, 0.01)
+        self.assertEqual(filter.num_layers, 6)
+        self.assertLessEqual(filter.false_positive_rate, 0.001)
         self.assertLessEqual(filter.utilization, 1.0)
+        self.assertGreater(filter.capacity, 0.0)
 
         self.assertTrue(filter.exists('foo'))
         self.assertTrue(filter.exists('bar'))
-        self.assertFalse(filter.exists('father'))
+        self.assertFalse(filter.exists('baz'))