fix prefixless search

This commit is contained in:
Linus Björnstam 2026-02-11 20:59:55 +01:00
parent 280177a9cb
commit a9b6ed9161
12 changed files with 1054 additions and 44 deletions

View file

@ -0,0 +1,198 @@
using System.Collections.Immutable;
using BenchmarkDotNet.Attributes;
using BenchmarkDotNet.Running;
using LanguageExt;
using PersistentMap;
namespace AgainstLanguageExt;
// Mocking your library types for the sake of the example to ensure compilation.
// Replace these with your actual namespace imports.
[MemoryDiagnoser]
[HideColumns("Ratio", "RatioSD", "Alloc Ratio")]
public class MapBenchmarks
{
[Params(10, 100, 1000)]
public int KeyLength { get; set; } // Key length
[Params(1000, 100000, 1000000)]
public int CollectionSize { get; set; }
// Data Source
private KeyValuePair<string, int>[] _data;
private string[] _searchKeys;
private int _index = 0;
private int _mask;
// Comparison Targets (for Lookup)
private ImmutableDictionary<string, int> _sysDict;
private ImmutableSortedDictionary<string, int> _sysSorted;
private LanguageExt.HashMap<string, int> _langExtHash;
private LanguageExt.Map<string, int> _langExtSorted; // Map<K,V> is Sorted in LangExt
private PersistentMap<string, int, UnicodeStrategy> _persistentMap;
[GlobalSetup]
public void Setup()
{
var random = new Random(42);
var dict = new Dictionary<string, int>();
// 1. Generate Data
while (dict.Count < CollectionSize)
{
string key = GenerateRandomString(random, KeyLength);
if (!dict.ContainsKey(key))
{
dict[key] = random.Next();
}
}
_data = dict.ToArray();
_searchKeys = dict.Keys.ToArray();
// 2. Pre-build maps for the Lookup benchmarks
_sysDict = dict.ToImmutableDictionary();
_sysSorted = dict.ToImmutableSortedDictionary();
_langExtHash = Prelude.toHashMap(dict);
_langExtSorted = Prelude.toMap(dict);
// Build PersistentMap for lookup test
var trans = PersistentMap<string, int, UnicodeStrategy>.Empty(new UnicodeStrategy()).ToTransient();
foreach (var kvp in _data)
{
trans.Set(kvp.Key, kvp.Value);
}
_persistentMap = trans.ToPersistent();
}
// ==========================================
// BUILD BENCHMARKS
// ==========================================
[Benchmark(Description = "Lookup: PersistentMap (Cyclic)")]
public int Lookup_Persistent_Cyclic()
{
// Fast wrap-around
var key = _searchKeys[_index++ & _mask];
_persistentMap.TryGetValue(key, out var value);
return value;
}
[Benchmark(Description = "Lookup: Sys.Sorted (Cyclic)")]
public int Lookup_SysSorted_Cyclic()
{
var key = _searchKeys[_index++ & _mask];
_sysSorted.TryGetValue(key, out var value);
return value;
}
[Benchmark(Description = "Build: Sys.ImmutableDict")]
public ImmutableDictionary<string, int> Build_SysImmutable()
{
// Using CreateRange/ToImmutable is usually the standard 'bulk' build
return _data.ToImmutableDictionary();
}
[Benchmark(Description = "Build: LangExt.HashMap")]
public LanguageExt.HashMap<string, int> Build_LangExtHash()
{
return Prelude.toHashMap(_data);
}
[Benchmark(Description = "Build: LangExt.SortedMap")]
public LanguageExt.Map<string, int> Build_LangExtSorted()
{
return Prelude.toMap(_data);
}
[Benchmark(Description = "Build: PersistentMap (Iterative)")]
public PersistentMap<string, int, UnicodeStrategy> Build_Persistent_Iterative()
{
// Simulating naive immutable building (O(n log n) or worse due to copying)
var map = PersistentMap<string, int, UnicodeStrategy>.Empty(new UnicodeStrategy());
foreach (var item in _data)
{
map = map.Set(item.Key, item.Value);
}
return map;
}
[Benchmark(Description = "Build: PersistentMap (Transient)")]
public PersistentMap<string, int, UnicodeStrategy> Build_Persistent_Transient()
{
// Simulating efficient mutable build -> freeze
var trans = PersistentMap<string, int, UnicodeStrategy>.Empty(new UnicodeStrategy()).ToTransient();
foreach (var item in _data)
{
trans.Set(item.Key, item.Value);
}
return trans.ToPersistent();
}
// ==========================================
// LOOKUP BENCHMARKS
// ==========================================
[Benchmark(Baseline = true, Description = "Lookup: Sys.ImmutableDict")]
public int Lookup_SysImmutable()
{
var key = _searchKeys[CollectionSize / 2];
_sysDict.TryGetValue(key, out var value);
return value;
}
[Benchmark(Description = "Lookup: Sys.SortedDict")]
public int Lookup_SysSorted()
{
var key = _searchKeys[CollectionSize / 2];
_sysSorted.TryGetValue(key, out var value);
return value;
}
[Benchmark(Description = "Lookup: LangExt.HashMap")]
public int Lookup_LangExtHash()
{
var key = _searchKeys[CollectionSize / 2];
// LanguageExt often uses Find which returns Option, or [] operator
// Assuming TryGetValue-like behavior or using match for fairness
return _langExtHash.Find(key).IfNone(0);
}
[Benchmark(Description = "Lookup: LangExt.SortedMap")]
public int Lookup_LangExtSorted()
{
var key = _searchKeys[CollectionSize / 2];
return _langExtSorted.Find(key).IfNone(0);
}
[Benchmark(Description = "Lookup: PersistentMap")]
public int Lookup_PersistentMap()
{
var key = _searchKeys[CollectionSize / 2];
_persistentMap.TryGetValue(key, out var value);
return value;
}
// Helper
private string GenerateRandomString(Random rng, int length)
{
const string chars = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789";
var buffer = new char[length];
for (int i = 0; i < length; i++)
{
buffer[i] = chars[rng.Next(chars.Length)];
}
return new string(buffer);
}
}
public class Program
{
public static void Main(string[] args)
{
// This scans the assembly and lets the command line (args) decide what to run
BenchmarkSwitcher.FromAssembly(typeof(Program).Assembly).Run(args);
}
}

View file

@ -0,0 +1,112 @@
using System;
using System.Collections.Generic;
using System.Collections.Immutable;
using System.Linq;
using BenchmarkDotNet.Attributes;
using BenchmarkDotNet.Running;
using LanguageExt;
using static LanguageExt.Prelude;
using PersistentMap;
[MemoryDiagnoser]
[HideColumns("Ratio", "RatioSD", "Alloc Ratio")]
public class CyclicMapBenchmarks
{
// Powers of 2 for fast bitwise masking
[Params(1024, 131072)]
public int CollectionSize { get; set; }
[Params(10, 100, 1000)]
public int N { get; set; }
// Collections
private PersistentMap<string, int, UnicodeStrategy> _persistentMap;
private ImmutableSortedDictionary<string, int> _sysSorted;
private LanguageExt.HashMap<string, int> _langExtHash;
private LanguageExt.Map<string, int> _langExtSorted;
// Lookup scaffolding
private string[] _searchKeys;
private int _index = 0;
private int _mask;
[GlobalSetup]
public void Setup()
{
_mask = CollectionSize - 1;
var random = new Random(42);
var data = new Dictionary<string, int>();
// 1. Generate Data
while (data.Count < CollectionSize)
{
string key = GenerateRandomString(random, N);
if (!data.ContainsKey(key))
{
data[key] = random.Next();
}
}
// 2. Build Collections
// PersistentMap
var builder = PersistentMap<string, int, UnicodeStrategy>.Empty(new UnicodeStrategy()).ToTransient();
foreach (var kvp in data) builder.Set(kvp.Key, kvp.Value);
_persistentMap = builder.ToPersistent();
// System
_sysSorted = data.ToImmutableSortedDictionary();
// LanguageExt
_langExtHash = toHashMap(data);
_langExtSorted = toMap(data);
// 3. Setup Cyclic Keys
_searchKeys = data.Keys.ToArray();
// Shuffle to defeat branch prediction / cache pre-fetching
Random.Shared.Shuffle(_searchKeys);
}
[Benchmark(Description = "Cyclic: PersistentMap")]
public int Lookup_Persistent()
{
var key = _searchKeys[_index++ & _mask];
_persistentMap.TryGetValue(key, out var value);
return value;
}
[Benchmark(Description = "Cyclic: Sys.Sorted")]
public int Lookup_SysSorted()
{
var key = _searchKeys[_index++ & _mask];
_sysSorted.TryGetValue(key, out var value);
return value;
}
[Benchmark(Description = "Cyclic: LangExt.HashMap")]
public int Lookup_LangExtHash()
{
var key = _searchKeys[_index++ & _mask];
// Option<T> struct return, overhead is minimal but present
return _langExtHash.Find(key).IfNone(0);
}
[Benchmark(Description = "Cyclic: LangExt.Sorted")]
public int Lookup_LangExtSorted()
{
var key = _searchKeys[_index++ & _mask];
// AVL Tree traversal
return _langExtSorted.Find(key).IfNone(0);
}
private string GenerateRandomString(Random rng, int length)
{
const string chars = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789";
var buffer = new char[length];
for (int i = 0; i < length; i++)
{
buffer[i] = chars[rng.Next(chars.Length)];
}
return new string(buffer);
}
}