practice with caches

Algorithms/datastructs/hash_table/cache.py

@@ -0,0 +1,100 @@
+import functools
+
+class Node(object):
+    def __init__(self, key, value):
+        self.key = key
+        self.value = value
+        self.prev = None
+        self.next = None
+
+class NodeList:
+    def __init__(self):
+        self.start = None
+        self.end = None
+    def add(self, key: int, value: int):
+        temp = Node(key, value)
+        if self.start:
+            if self.start == self.end:
+                self.start.next = temp
+                self.end = temp
+                self.end.prev = self.start
+            else:
+                temp.prev = self.end
+                self.end.next = temp
+                self.end = temp
+        else:
+            self.start = temp
+            self.end = self.start
+        return temp
+    def remove(self, node: Node):
+        prev_node = node.prev
+        next_node = node.next
+        if prev_node:
+            prev_node.next = next_node
+        else:
+            self.start = next_node
+        if next_node:
+            next_node.prev = prev_node
+        else:
+            self.end = prev_node
+    def move_to_end(self, node):
+        if node == self.end:
+            return
+        self.remove(node)
+        node.prev = self.end
+        node.next = None
+        self.end.next = node
+        self.end = node
+    def evict(self):
+        start = self.start
+        if not start:
+            return None
+        next_node = start.next
+        if next_node:
+            next_node.prev = None
+            self.start = next_node
+        else:
+            self.start = None
+            self.end = None
+        return start
+    def __str__(self):
+        head = self.start
+        res = ""
+        while head:
+            res += f"Node(key: {head.key}, value: {head.value}, prev: {head.prev}, next: {head.next})"
+            head = head.next
+        return res
+
+class LRUCache:
+    def __init__(self, capacity: int):
+        self.capacity = capacity
+        self.table = dict()
+        self.cache = NodeList()
+    def get(self, key: int) -> int:
+        if key in self.table:
+            node = self.table[key]
+            self.cache.move_to_end(node)
+            return node.value
+        else:
+            return -1
+    def put(self, key: int, value: int) -> None:
+        if key in self.table:
+            old_node = self.table[key]
+            old_node.value = value
+            self.cache.move_to_end(old_node)
+        else:
+            if self.capacity == 0:
+                return
+            if len(self.table) == self.capacity:
+                lru = self.cache.evict()
+                del self.table[lru.key]
+                node = self.cache.add(key, value)
+                self.table[key] = node
+            else:
+                node = self.cache.add(key, value)
+                self.table[key] = node
+
+# using the functools cache decorator
+@functools.lru_cache
+def factorial(n: int) -> int:
+    return n*factorial(n-1) if n > 1 else 1

tests/test_cache.py

@@ -0,0 +1,127 @@
+import time
+import pytest
+import functools
+from Algorithms.datastructs.hash_table.cache import LRUCache, factorial
+
+# -------------------------
+# Correctness tests: LRUCache
+# -------------------------
+
+def test_basic_put_get():
+    cache = LRUCache(2)
+    cache.put(1, 10)
+    cache.put(2, 20)
+
+    assert cache.get(1) == 10
+    assert cache.get(2) == 20
+    assert cache.get(3) == -1
+
+
+def test_eviction_order():
+    cache = LRUCache(2)
+    cache.put(1, 1)
+    cache.put(2, 2)
+
+    # Access key 1 -> makes key 2 LRU
+    assert cache.get(1) == 1
+
+    # This should evict key 2
+    cache.put(3, 3)
+
+    assert cache.get(2) == -1
+    assert cache.get(1) == 1
+    assert cache.get(3) == 3
+
+
+def test_update_existing_key():
+    cache = LRUCache(2)
+    cache.put(1, 1)
+    cache.put(2, 2)
+
+    cache.put(1, 100)  # update value + recency
+
+    assert cache.get(1) == 100
+
+    # key 2 should now be LRU
+    cache.put(3, 3)
+
+    assert cache.get(2) == -1
+    assert cache.get(1) == 100
+    assert cache.get(3) == 3
+
+
+def test_capacity_one():
+    cache = LRUCache(1)
+    cache.put(1, 1)
+    cache.put(2, 2)
+
+    assert cache.get(1) == -1
+    assert cache.get(2) == 2
+
+
+def test_capacity_zero():
+    cache = LRUCache(0)
+    cache.put(1, 1)
+
+    assert cache.get(1) == -1
+
+
+def test_repeated_get_does_not_break_order():
+    cache = LRUCache(2)
+    cache.put(1, 1)
+    cache.put(2, 2)
+
+    # Repeated gets
+    assert cache.get(1) == 1
+    assert cache.get(1) == 1
+
+    cache.put(3, 3)
+
+    # key 2 should still be evicted
+    assert cache.get(2) == -1
+    assert cache.get(1) == 1
+    assert cache.get(3) == 3
+
+
+# -------------------------
+# Timing tests: functools.lru_cache
+# -------------------------
+
+def uncached_factorial(n: int) -> int:
+    return n * uncached_factorial(n - 1) if n > 1 else 1
+
+
+def test_lru_cache_speedup():
+    """
+    This test checks that functools.lru_cache provides
+    a measurable speedup over uncached recursion.
+
+    We avoid flaky microbenchmarks by:
+    - warming the cache
+    - using multiple calls
+    - asserting relative (order-of-magnitude) improvement
+    """
+
+    n = 300
+
+    # Warm-up cached version
+    factorial.cache_clear()
+    factorial(n)
+
+    # Measure cached calls
+    start = time.perf_counter()
+    for _ in range(1000):
+        factorial(n)
+    cached_time = time.perf_counter() - start
+
+    # Measure uncached calls (much fewer iterations)
+    start = time.perf_counter()
+    for _ in range(10):
+        uncached_factorial(n)
+    uncached_time = time.perf_counter() - start
+
+    # Cached should be significantly faster per call
+    cached_per_call = cached_time / 1000
+    uncached_per_call = uncached_time / 10
+
+    assert cached_per_call < uncached_per_call / 50

tests/test_hash_table.py

@@ -1,4 +1,4 @@
-from Algorithms.datastructs.hash_table import RandomizedSet
+from Algorithms.datastructs.hash_table.hash_table import RandomizedSet
 
 def test_randomized_set():
     rs = RandomizedSet()