min_max 0.1.3 → 0.1.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 5b1020f3ac8e632e3b2b4d99318a78edf48a06d3099e5812637d8dfd5fdf1623
-  data.tar.gz: b774055ca657edf4a01712a71020cbb55df034e38f3db676b0c85484ae1f155e
+  metadata.gz: 6d739946dcb0939e2e05893222c927f4dea0eda65da5c0a00156402de3e49c29
+  data.tar.gz: 61b4453a94b93bb2ccefd0787a751cef5d3ef380ac0ed591c73c3b0afddc394b
 SHA512:
-  metadata.gz: 2c7a1e0c3893382526f5c94c488618a3d01b7e19aa3a6d1b3cfc0aa012ede494bfb033c1b9544224bcfa3f9147d3d96779d11c83f2723a384c36109c5f837dba
-  data.tar.gz: 4c60344475780f72a7369cac4fc0a22e156750fee690b326a2cbc9b6d1c492446f05dbdda5e39910e9f9549b848730ade850fda8e282d418ba1fa0e8e8c19793
+  metadata.gz: 70dd4402ca82a60cffbc451ea134ca9fc34391f28311d23726d22055bfc01db9ba62002f0d54c76282888993613f2bf5618dcfa7a964da39c6995747401f3a3f
+  data.tar.gz: 31ada4f8ed50624701e389c6a1d50b83b75a585ef703d2f18c6511498dc9373647fa1f88d959ccdcbf56336d38dc550e1ce96710644f65970f2341549f843d53

data/Cargo.lock

@@ -270,9 +270,9 @@ checksum = "24188a676b6ae68c3b2cb3a01be17fbf7240ce009799bb56d5b1409051e78fde"
 
 [[package]]
 name = "shlex"
-version = "1.1.0"
+version = "1.3.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "43b2853a4d09f215c24cc5489c992ce46052d359b5109343cbafbf26bc62f8a3"
+checksum = "0fda2ff0d084019ba4d7c6f371c95d8fd75ce3524c3cb8fb653a3023f6323e64"
 
 [[package]]
 name = "syn"

data/Cargo.toml

@@ -8,5 +8,6 @@ resolver = "2"
 
 [profile.release]
 opt-level = 3
-lto = "fat"
+lto = "thin"
 codegen-units = 1
+strip = "debuginfo"

data/README.md

@@ -1,7 +1,8 @@
-# MinMax
+# MinMax Heap
 
-The MinMax gem provides a high-performance minmax heap implementation for Ruby, written in Rust. 
-This gem allows for the creation of a min-max-heap and supporting operations like pushing and popping multiple items, iterating over heap items, and converting heaps to arrays. 
+The MinMax Heap gem provides a high-performance minmax heap implementation for Ruby, written in Rust. 
+The gem wraps the excellent [min-max-heap-rs](https://github.com/tov/min-max-heap-rs) Rust library.
+It allows for the creation of a min-max-heap and supporting operations like pushing and popping multiple items, iterating over heap items, and converting heaps to arrays. 
 
 ## Features
 
@@ -25,6 +26,7 @@ Add this line to your application's Gemfile:
 gem 'min_max'
 ```
 
+
 And then execute:
 
 ```bash
@@ -34,6 +36,10 @@ bundle install
 Or install it yourself as:
 ```bash
 gem install min_max
+
+# or manually specify target . E.g.
+
+CARGO_BUILD_TARGET=x86_64-apple-darwin gem install min_max
 ```
 
 ## Usage
@@ -84,4 +90,13 @@ heap.length # Alias for size
 heap.clear
 ```
 
+## Count number of times an item is in the heap
+```ruby
+heap.count(item)
+```
+
+## Check if item is contained in the heap
+```ruby
+heap.contains?(item)
+```
 

data/benchmarks/benchmarks.rb

@@ -0,0 +1,85 @@
+require "bundler/setup"
+require "min_max"
+
+require 'bundler/inline'
+
+gemfile do
+  source 'https://rubygems.org'
+  gem 'rb_heap'
+  gem 'algorithms'
+  gem 'ruby-heap'
+  gem 'pqueue'
+end
+
+require 'benchmark'
+require 'rb_heap'
+require 'min_max'
+
+
+data = 100_000.times.map{ Random.rand(0...1000_000_000) }
+rb_hp = Heap.new(:<)
+mm_hp = MinMax[]
+
+Benchmark.bm do |x|
+  x.report("push_rb_heap"){ data.each{|d| rb_hp.add(d) } }
+  x.report("push_mm_heap"){ data.each{|d| mm_hp.push(d) } }
+  x.report("push_mm_heap_batches"){ mm_hp = MinMax[]; data.each_slice(1000){|d| mm_hp.push(*d) } }
+end
+
+Benchmark.bm do |x|
+  x.report("pop_rb_heap"){ 100_000.times{|d| rb_hp.pop } }
+  x.report("pop_mm_heap"){ 100_000.times{|d| mm_hp.pop_min } }
+  x.report("pop_mm_heap_batches"){ 1000.times{|d| mm_hp.pop_min(100) } }
+end
+
+Object.send(:remove_const, :Heap)
+
+require 'Heap'
+puts "# ruby-heap vs min_max"
+ruby_hp = Heap::BinaryHeap::MinHeap.new
+mm_hp = MinMax[]
+
+Benchmark.bm do |x|
+  x.report("push_ruby_heap"){ data.each{|d| ruby_hp.add(d) } }
+  x.report("push_mm_heap"){ data.each{|d| mm_hp.push(d) } }
+  x.report("push_mm_heap_batches"){ mm_hp = MinMax[]; data.each_slice(1000){|d| mm_hp.push(*d) } }
+end
+
+Benchmark.bm do |x|
+  x.report("pop_ruby_heap"){ 100_000.times{|d| ruby_hp.extract_min! } }
+  x.report("pop_mm_heap"){ 100_000.times{|d| mm_hp.pop_min } }
+  x.report("pop_mm_heap_batches"){ 1000.times{|d| mm_hp.pop_min(100) } }
+end
+
+
+require 'algorithms'
+minheap = Containers::MinHeap.new()
+mm_hp = MinMax[]
+
+Benchmark.bm do |x|
+  x.report("push_algos_heap"){ data.each{|d| minheap.push(d) } }
+  x.report("push_mm_heap"){ data.each{|d| mm_hp.push(d) } }
+  x.report("push_mm_heap_batches"){ mm_hp = MinMax[]; data.each_slice(1000){|d| mm_hp.push(*d) } }
+end
+
+Benchmark.bm do |x|
+  x.report("pop_algos_heap"){ 100_000.times{|d| minheap.pop } }
+  x.report("pop_mm_heap"){ 100_000.times{|d| mm_hp.pop_min } }
+  x.report("pop_mm_heap_batches"){ 1000.times{|d| mm_hp.pop_min(100) } }
+end
+
+require 'pqueue'
+pqueue = PQueue.new()
+mm_hp = MinMax[]
+
+Benchmark.bm do |x|
+  x.report("push_algos_heap"){ data.each{|d| pqueue.push(d) } }
+  x.report("push_mm_heap"){ data.each{|d| mm_hp.push(d) } }
+  x.report("push_mm_heap_batches"){ mm_hp = MinMax[]; data.each_slice(1000){|d| mm_hp.push(*d) } }
+end
+
+Benchmark.bm do |x|
+  x.report("pop_algos_heap"){ 100_000.times{|d| minheap.pop } }
+  x.report("pop_mm_heap"){ 100_000.times{|d| mm_hp.pop_min } }
+  x.report("pop_mm_heap_batches"){ 1000.times{|d| mm_hp.pop_min(100) } }
+end

data/ext/min_max/src/lib.rs

@@ -1,5 +1,5 @@
 use magnus::scan_args::scan_args;
-use magnus::value::{BoxValue, ReprValue};
+use magnus::value::ReprValue;
 use magnus::{
     block::{block_given, Yield},
     define_class, function, method,
@@ -9,32 +9,11 @@ use magnus::{
 use magnus::{DataTypeFunctions, Integer, TypedData};
 use min_max_heap::MinMaxHeap;
 use std::cell::RefCell;
-use std::collections::HashMap;
-use std::sync::{Arc, RwLock};
+use std::rc::Rc;
 
 #[derive(Debug, Clone)]
 struct PriorityOrderableValue(i64, i64);
 
-#[macro_use]
-extern crate lazy_static;
-
-struct MEM(
-    RwLock<HashMap<i64, BoxValue<Value>>>,
-    RwLock<HashMap<i64, RefCell<usize>>>,
-);
-
-impl MEM {
-    fn new() -> Self {
-        MEM(RwLock::new(HashMap::new()), RwLock::new(HashMap::new()))
-    }
-}
-
-unsafe impl Sync for MEM {}
-
-lazy_static! {
-    static ref MEMORY: MEM = MEM::new();
-}
-
 impl PartialOrd for PriorityOrderableValue {
     fn partial_cmp(&self, other: &Self) -> Option<std::cmp::Ordering> {
         Some(self.cmp(other))
@@ -43,9 +22,8 @@ impl PartialOrd for PriorityOrderableValue {
 
 impl Ord for PriorityOrderableValue {
     fn cmp(&self, other: &Self) -> std::cmp::Ordering {
-        match (self, other) {
-            (PriorityOrderableValue(p1, _), PriorityOrderableValue(p2, _)) => p1.cmp(p2),
-        }
+        let (PriorityOrderableValue(p1, _), PriorityOrderableValue(p2, _)) = (self, other);
+        p1.cmp(p2)
     }
 }
 
@@ -53,11 +31,8 @@ impl Eq for PriorityOrderableValue {}
 
 impl PartialEq for PriorityOrderableValue {
     fn eq(&self, other: &Self) -> bool {
-        match (self, other) {
-            (PriorityOrderableValue(p1, v1), PriorityOrderableValue(p2, v2)) => {
-                p1.eq(p2) && (v1).eq(v2)
-            }
-        }
+        let (PriorityOrderableValue(p1, v1), PriorityOrderableValue(p2, v2)) = (self, other);
+        p1.eq(p2) && (v1).eq(v2)
     }
 }
 
@@ -65,21 +40,21 @@ unsafe impl Send for RubyMinMaxHeap {}
 #[derive(DataTypeFunctions, TypedData, Clone)]
 #[magnus(class = "MinMax", size, free_immediately, mark)]
 struct RubyMinMaxHeap {
-    heap: Arc<RefCell<MinMaxHeap<PriorityOrderableValue>>>,
+    heap: Rc<RefCell<MinMaxHeap<PriorityOrderableValue>>>,
 }
 
 impl RubyMinMaxHeap {
     fn new() -> Result<Self, Error> {
         Ok(RubyMinMaxHeap {
-            heap: Arc::new(RefCell::new(MinMaxHeap::new())),
+            heap: Rc::new(RefCell::new(MinMaxHeap::new())),
         })
     }
 
     fn push(&self, values: RArray) -> Result<(), Error> {
         let mut hp = self.heap.borrow_mut();
-        let values_vec = values.to_vec::<(i64, i64)>().unwrap();
-        values_vec.into_iter().for_each(|(priority, key)| {
-            hp.push(PriorityOrderableValue(priority, key));
+        let values_vec = values.to_vec::<(i64, i64)>()?;
+        values_vec.iter().for_each(|(priority, key)| {
+            hp.push(PriorityOrderableValue(*priority, *key));
         });
         Ok(())
     }
@@ -213,8 +188,8 @@ impl RubyMinMaxHeap {
         Ok(ary)
     }
 
-    fn clear(&self) -> Option<()> {
-        Some(self.heap.borrow_mut().clear())
+    fn clear(&self) {
+        self.heap.borrow_mut().clear()
     }
 }
 

data/lib/min_max/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 class MinMax
-  VERSION = "0.1.3"
+  VERSION = "0.1.5"
 end

data/lib/min_max.rb

@@ -6,46 +6,35 @@ require_relative "min_max/min_max"
 class MinMax
   class Error < StandardError; end
 
+  attr_reader :priority_blk, :storage, :mtx
+
   def self.[](*args, &blk)
     new(*args, &blk)
   end
 
   def self.new(*args, &blk)
     self._new.tap{|s|
-      s.priority_blk =(blk || proc{|x| (x.respond_to?(:priority) ? x.priority : x.to_i ) rescue 0})
+      s.instance_eval{
+        @priority_blk = (blk || proc{|x| x.respond_to?(:priority) ? x.priority : x.to_i })
+        @storage = Hash.new{|h,k| h[k] = [0, nil] }
+        @mtx  ||= Mutex.new
+      }
       s.push(*args)
     }
   end
 
-  def priority_blk=(bll)
-    @priority_blk = bll
-  end
-
-  def storage
-    @storage ||= Hash.new
-  end
-
-  def counts
-    @counts ||= Hash.new(0)
-  end
-
-  def mtx
-    @_mtx  ||= Mutex.new
-  end
-
   def push(*args)
-    mtx.synchronize do
-      mapped = args.map do |a|
-          counts[a.hash] += 1
-          storage[a.hash] = a
-          [
-            @priority_blk.call(a),
-            a.hash
-          ]
-      end
-       _push(mapped)
+    mapped = args.map do |a|
+      hash = a.hash
+      entry = self.storage[hash]
+      entry[0] += 1
+      entry[1] ||= a
+      [
+        (self.priority_blk.call(a) rescue 0),
+        hash
+      ]
     end
-    self
+     _push(mapped)
   end
 
   def add(*args)
@@ -53,29 +42,13 @@ class MinMax
   end
 
   def pop_max(*args)
-    mtx.synchronize {
-      popped = _pop_max(*args)
-      popped.kind_of?(Array) ? popped.map{|p| retrieve(p) } : retrieve(popped)
-    }
+    popped = _pop_max(*args)
+    popped.kind_of?(Array) ? popped.map{|p| retrieve(p) } : retrieve(popped)
   end
 
   def pop_min(*args)
-    mtx.synchronize {
-      popped = _pop_min(*args)
-      popped.kind_of?(Array) ? popped.map{|p| retrieve(p) } : retrieve(popped)
-    }
-  end
-
-  def retrieve(hash, remove=true)
-    if remove
-      if (counts[hash] -= 1) == 0
-        storage.delete(hash)
-      else
-        storage[hash]
-      end
-    else
-      storage[hash]
-    end
+    popped = _pop_min(*args)
+    popped.kind_of?(Array) ? popped.map{|p| retrieve(p) } : retrieve(popped)
   end
 
   def peek_min(*args)
@@ -100,6 +73,14 @@ class MinMax
     each.to_a
   end
 
+  def count(val)
+    counts.has_key?(val.hash) ? counts[val.hash] : 0
+  end
+
+  def contains?(val)
+    counts.has_key?(val.hash) && counts[val.hash] > 0
+  end
+
   def to_a_asc
     mtx.synchronize { _to_a_asc.map{|p| retrieve(p, false) } }
   end
@@ -109,7 +90,15 @@ class MinMax
   end
 
   def inspect
-    "MinMax[#{to_a_asc.first(10).map{|v| v.to_s}.join(", ")}#{size > 10 ? ", ..." : ""}]"
+    "MinMax[#{_each.first(10).map{|v| retrieve(v, false).to_s }.join(", ")}#{size > 10 ? ", ..." : ""}]"
+  end
+
+  private
+  def retrieve(hash, remove=true)
+    entry = self.storage[hash]
+    self.storage.delete(hash) if remove && (entry[0] -= 1) == 0
+    entry[1]
   end
+
 end
 

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: min_max
 version: !ruby/object:Gem::Version
-  version: 0.1.3
+  version: 0.1.5
 platform: ruby
 authors:
 - Wouter Coppieters
@@ -66,6 +66,7 @@ files:
 - LICENSE.txt
 - README.md
 - Rakefile
+- benchmarks/benchmarks.rb
 - ext/min_max/Cargo.toml
 - ext/min_max/extconf.rb
 - ext/min_max/src/lib.rs