contrek 1.0.6 → 1.0.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9a9bb4f4ff327f04f6877b024c533c8d6d7b6f0ecfc7ad0ae51e6ad05a84438c
-  data.tar.gz: 9bf8394de257d8a3832400936992d5f848e015d10e389ce986347668d4f53b23
+  metadata.gz: 862b863c602031251fedc2dc18a59c9bd270d91acff25761c24164e34e8952e9
+  data.tar.gz: 97f18e804accd2d6ae11f95223c95eebe198eafb07c745261d1110f5edf5eae4
 SHA512:
-  metadata.gz: c9640a27737904a3ab5eb4b4f1357262ba05dac579c743ba653ef9be0ed73a90648adc5af0d3c5ff9629f04aa12ca27b5f7e703855805488690d604dd7a82572
-  data.tar.gz: 656e58e008e3a6b912f5b99ed1b90dbe9b673dab08815177d14ec5d3de3a9380359b70d259517bdd2278b5f327db86bb448293d72555534205751e04a49378b1
+  metadata.gz: d3d9ce7c0b5ae1a3c62f91480d63856312d89334f181899ccc2ff304c3fef983578838484da18f865ce1d50ea3cb3c291ce49fef1b01dcfaf397ab688bed9e37
+  data.tar.gz: c8f7aade05577e132829e7ffcc6af5e637aa879bb6972f4ba8566c8db7d11b641862e32a170c39db4b8ab57666996a4639ffd3f516ba85712266b7ffd82ac276

data/CHANGELOG.md

@@ -18,4 +18,14 @@ All notable changes to this project will be documented in this file.
 ### Added
 - Added C++ multithreading supports
 - Optimized old just present C++ code
-- Removed Png++ dependency in place of libspn 
+- Removed Png++ dependency in place of libspn
+
+## [1.0.7] - 2026-01-10
+### Added
+- Optimized C++/Ruby data transfer: Implemented NumPy-compatible binary serialization for coordinate outputs. This reduces serialization overhead and significantly increases data throughput between the C++ core and the Ruby interface.
+
+## [1.0.8] - 2026-01-25
+### Changed
+- Fixed ARGB/RGBA format discrepancies between Ruby and C++.
+- Updated the multithreading-side algorithm for rejoining mono-connected orphan polygons to other orphan polygons. The algorithm can now identify them and defer their processing within the pipeline.
+- Improved, on the multithreading side, the polygon intersection detection mechanism; the geometric approach has been dropped in favor of a purely topological one.

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    contrek (1.0.5)
+    contrek (1.0.8)
       chunky_png (~> 1.4)
       concurrent-ruby (~> 1.3.5)
       rice (= 4.5.0)

data/README.md

@@ -1,5 +1,5 @@
 # Contrek
-Contrek is a Ruby library (C++ powered) to trace png bitmap areas polygonal contours. Manages png images usign libspng (version 0.7.4) library. May work multithreading.
+Contrek is a Ruby gem with a C++ core for fast contour tracing and edge detection in PNG images. Unlike standard geometric tracers, it employs a **topological approach** to extract polygonal contours, representing shapes as a connected graph of shared endpoints. This ensures perfect adjacency and structural integrity for shape analysis and raster-to-vector workflows, such as PNG to SVG conversion, managed via libspng (0.7.4) with multithreading support.
 
 ## About Contrek library
 Contrek (**con**tour **trek**king) simply scans your png bitmap and returns shape contour as close polygonal lines, both for the external and internal sides. It can compute the nesting level of the polygons found with a tree structure. It supports various levels and modes of compression and approximation of the found coordinates. It is capable of multithreaded processing, splitting the image into vertical strips and recombining the coordinates in pairs.
@@ -34,9 +34,10 @@ result = Contrek.contour!(
   }
 )
 ```
-The result reports information about the execution times (microseconds), the polygons found, their coordinates and the nesting tree.
+The resulting metadata information contains the execution times (microseconds), the count of polygons found and the nesting tree map. You can access polygons coordinates too (see later).
+
 ```ruby
-{:benchmarks=>{"build_tangs_sequence"=>0.129, "compress"=>0.037, "plot"=>0.198, "scan"=>0.114, "total"=>0.478}, :groups=>2, :named_sequence=>"", :polygons=>[...], :treemap=>[]}
+{:benchmarks=>{"build_tangs_sequence"=>0.129, "compress"=>0.037, "plot"=>0.198, "scan"=>0.114, "total"=>0.478}, :groups=>2, :named_sequence=>"", :treemap=>[]}
 
 ```
 
@@ -63,7 +64,7 @@ polygonfinder = CPPPolygonFinder.new(png_bitmap,
   {versus: :a, compress: {visvalingam: {tolerance: 1.5}}})
 result = polygonfinder.process_info
 # draws the polygons found
-Contrek::Bitmaps::Painting.direct_draw_polygons(result[:polygons], png_image)
+Contrek::Bitmaps::Painting.direct_draw_polygons(result.points, png_image)
 png_image.save('result.png') # => inspect the image to feedback the result
 ```
 
@@ -98,7 +99,7 @@ Regarding multithreading:
 
 - The treemap option is currently ignored (multithreaded treemap support will be introduced in upcoming revisions).
 
-By not declaring native option CPP Multithreading optimized code is used. In the above example a [105 MP image](spec/files/images/sample_10240x10240.png) is examined by 2 thread working on 2 tiles (total compute time about 2 secs).
+By not declaring native option CPP Multithreading optimized code is used. In the above example a [105 MP image](spec/files/images/sample_10240x10240.png) is examined by 2 threads working on 2 tiles (total compute time about 2 secs).
 
 ```ruby
 result = Contrek.contour!(
@@ -110,7 +111,7 @@ result = Contrek.contour!(
     finder: {number_of_tiles: 2, compress: {uniq: true}}
   }
 )
-puts result[:benchmarks].inspect
+puts result.metadata[:benchmarks].inspect
 
 {"compress"=>14.596786999999999,
   "init"=>2078.745861,
@@ -118,6 +119,97 @@ puts result[:benchmarks].inspect
   "outer"=>118.94489599999999,
   "total"=>2093.342648}
 ```
+
+## Result
+
+The result structure contains polygon coordinates and a set of metadata. Polygon coordinates can be accessed via:
+
+```ruby
+result.polygons
+
+[{:outer=>[11, 2, 11, 5, 6, 5, 6, 2],
+  :inner=>[[10, 3, 7, 3, 7, 4, 10, 4]]}]
+```
+
+For native classes (C/C++), coordinates are represented as an interleaved NumPy array (`ndarray`) in the form:
+
+```
+[x0, y0, x1, y1, ...]
+```
+
+A ruby side helper method could converts this data structure into point-based representations:
+
+```ruby
+result.points
+
+[{:outer=>[{:x=>11, :y=>2}, {:x=>11, :y=>5}, {:x=>6, :y=>5}, {:x=>6, :y=>2}],
+  :inner=>[[{:x=>10, :y=>3}, {:x=>7, :y=>3}, {:x=>7, :y=>4}, {:x=>10, :y=>4}]]}]
+```
+
+- `outer` represents a single polygonal sequence  
+- `inner` is represented as an array of polygonal sequences (holes)
+
+For pure Ruby implementations (`{ native: false }`), coordinates are always expressed as points.  
+In this case, both `polygons` and `points` return the same data, represented as hashes with `x` and `y` keys.
+
+## Metadata
+
+Metadata associated with the result can be accessed via:
+
+```ruby
+result.metadata
+
+{:benchmarks=>{"build_tangs_sequence"=>0.00577,
+               "compress"=>0.01239,
+               "plot"=>0.01919,
+               "scan"=>0.01473,
+               "total"=>0.05208},
+ :groups=>1,
+ :named_sequence=>"AFEDCBA",
+ :treemap=>[]}
+```
+
+## Treemap
+
+The treemap is a data structure that represents the containment hierarchy of polygons. For each polygon, it defines the parent polygon in which it is contained and its relative position among siblings. This structure allows reconstruction of the inclusion tree and determination of nesting relationships between geometries. Consider the above image
+
+```ruby
+ "AAAAAAAAAAAAAAAAAAAAAA" \
+ "A                    A" \
+ "A BBBBBBBBBBBBBBBBBB A" \
+ "A BBBBBBBBBBBBBBBBBB A" \
+ "A BBB    BBBBB   BBB A" \
+ "A BBB CC BBBBB D BBB A" \
+ "A BBB    BBBBB   BBB A" \
+ "A BBBBBBBBBBBBBBBBBB A" \
+ "A BBBBBBBBBBBBBBBBBB A" \
+ "A                    A" \
+ "AAAAAAAAAAAAAAAAAAAAAA"
+```
+
+```ruby
+result.metadata[:treemap]
+
+[[-1, -1],  # A
+ [0, 0],    # B
+ [1, 0],    # C
+ [1, 1]]    # D
+```
+
+There are four polygons (`A`, `B`, `C`, and `D`).  
+The order matches the one provided in `result.polygons`.
+
+Each entry has the structure:
+
+```
+[parent_index, position]
+```
+
+- Polygon **A** (index `0`) has no parent and is represented as `[-1, -1]`
+- Polygon **B** is contained in **A** and is its first child → `[0, 0]`
+- Polygon **C** is contained in **B** and is its first child → `[1, 0]`
+- Polygon **D** is also contained in **B** and is its second child → `[1, 1]`
+
 ## Multithreaded approach
 
 The multithreaded contour-tracing implementation operates as follows:
@@ -178,6 +270,10 @@ This project is licensed under the terms of the MIT license.
 
 See [LICENSE.md](LICENSE.md).
 
+## Changelog
+
+See [CHANGELOG.md](CHANGELOG.md) for a complete list of changes.
+
 ## History
 The algorithm was originally developed by me in 2018 when I was commissioned to create a Rails web application whose main objective was to census buildings from GoogleMAPS; the end user had to be able to select their home building by clicking its roof on the map which had to be identified as a clickable polygon. The solution was to configure GoogleMAPS to render buildings of a well-defined color (red), and at each refresh of the same to transform the div into an image (html2canvas) then process it server side returning the polygons to be superimposed again on the map. This required very fast polygons determination. Searching for a library for tracing the contours I was not able to find anything better except OpenCV which however seemed to me a very heavy dependency. So I decided to write my algorithm directly in the context of the ROR application. Once perfected, it was already usable but a bit slow in the higher image resolutions. So I decided to write the counterpart in C++, which came out much faster and which I then used as an extension on Ruby by means of Rice.
 

data/ext/cpp_polygon_finder/PolygonFinder/Makefile

@@ -2,15 +2,15 @@ CXX = g++
 CXXFLAGS = -Wall -Wextra -std=c++17 -Isrc -MMD -MP -pthread -g \
            -fno-omit-frame-pointer -fsanitize=address,undefined \
            -O0
-LDFLAGS = -lpng -pthread -lz -fsanitize=address,undefined
+LDFLAGS = -pthread -lz -fsanitize=address,undefined
 
 # valgrinf or perf setup
-#CXXFLAGS = -Wall -Wextra -std=c++17 -Isrc -MMD -MP -pthread -g -fno-omit-frame-pointer -march=native -DNDEBUG -O3
-#LDFLAGS = -lpng -pthread -lz
+# CXXFLAGS = -Wall -Wextra -std=c++17 -Isrc -MMD -MP -pthread -g -fno-omit-frame-pointer -march=native -DNDEBUG -O3
+# LDFLAGS = -lpng -pthread -lz
 
 # best performances
 CXXFLAGS = -Wall -Wextra -std=c++17 -Isrc -MMD -MP -pthread -g -march=native -DNDEBUG -Ofast -flto
-LDFLAGS = -lpng -pthread -lz -flto=auto
+LDFLAGS = -pthread -lz -flto=auto
 
 SRC = $(shell find src -name '*.cpp')
 C_SRC = $(shell find src -name 'spng.c')

data/ext/cpp_polygon_finder/PolygonFinder/src/Tests.cpp

@@ -107,7 +107,7 @@ void Tests::test_c()
   Point* p2 = new Point({2, 2});
   Point* p3 = new Point({3, 3});
 
-  Hub* hub = new Hub(4, 4);  // coordinate 0 -> 3 sia x che y, matrice di 4 * 4
+  Hub* hub = new Hub(4, 0, 3);
 
   Position* pos1 = new Position(hub, p1);
   Position* pos2 = new Position(hub, p2);
@@ -176,15 +176,16 @@ void Tests::test_e()
   cpu_timer.start();
   FastPngBitmap png_bitmap("images/sample_10240x10240.png");
   // FastPngBitmap png_bitmap("images/sample_1024x1024.png");
-  std::cout << "immagine =" << cpu_timer.stop() << std::endl;
+  std::cout << "image reading time =" << cpu_timer.stop() << std::endl;
 
-  int color = png_bitmap.value_at(0, 0);
-  std::cout << "color =" << color << std::endl;
+  int color = png_bitmap.rgb_value_at(0, 0);
+  std::cout << "color = " << color << std::endl;
   RGBNotMatcher not_matcher(color);
 
   std::vector<std::string> arguments = {"--versus=a", "--compress_uniq", "--number_of_tiles=2"};
   Finder pl(2, &png_bitmap, &not_matcher, &arguments);
   ProcessResult *o = pl.process_info();
   o->print_info();
+  std::cout << "polygons =" << o->groups << std::endl;
   delete o;
 }

data/ext/cpp_polygon_finder/PolygonFinder/src/polygon/bitmaps/FastPngBitmap.cpp

@@ -70,12 +70,10 @@ char FastPngBitmap::value_at(int x, int y) {
   return(0);
 }
 
+// source image format RGBA => returning uint ABGR
 unsigned int FastPngBitmap::rgb_value_at(int x, int y) {
-  int32_t index = ((y * width) + x) * 4;
-  unsigned int color;
-  unsigned char *red = &image[index];
-  std::memcpy(&color, red, 3);
-  return(color);
+  uint32_t index = (uint32_t(y) * width + x) * 4;
+  return *reinterpret_cast<const uint32_t*>(&image[index]);
 }
 
 const unsigned char* FastPngBitmap::get_row_ptr(int y) const {

data/ext/cpp_polygon_finder/PolygonFinder/src/polygon/finder/PolygonFinder.cpp

@@ -89,7 +89,7 @@ void PolygonFinder::scan() {
       if (rgb_m) run_loop(rgb_m, fetcher);
       else       run_loop(this->matcher, fetcher);
     } else {
-      auto fetcher = [](const unsigned char* p) { return p[0] | (p[1] << 8) | (p[2] << 16); };
+      auto fetcher = [](const unsigned char* p) { return *reinterpret_cast<const uint32_t*>(p); };
       if (rgb_m) run_loop(rgb_m, fetcher);
       else       run_loop(this->matcher, fetcher);
     }

data/ext/cpp_polygon_finder/PolygonFinder/src/polygon/finder/concurrent/Cluster.cpp

@@ -16,10 +16,10 @@
 #include "Cursor.h"
 #include "../../CpuTimer.h"
 
-Cluster::Cluster(Finder *finder, int height, int width)
+Cluster::Cluster(Finder *finder, int height, int start_x, int end_x)
   : finder(finder)
 { tiles_.reserve(2);  // only two (left|right)
-  this->hub_ = new Hub(height, width);
+  this->hub_ = new Hub(height, start_x, end_x);
 }
 
 Cluster::~Cluster() {

data/ext/cpp_polygon_finder/PolygonFinder/src/polygon/finder/concurrent/Cluster.h

@@ -20,7 +20,7 @@ class Cluster {
   std::vector<Tile*> tiles_;
   Hub *hub_ = nullptr;
  public:
-  Cluster(Finder *finder, int height, int width);
+  Cluster(Finder *finder, int height, int start_x, int end_x);
   virtual ~Cluster();
   void add(Tile* tile);
   Tile* merge_tiles();

data/ext/cpp_polygon_finder/PolygonFinder/src/polygon/finder/concurrent/Cursor.cpp

@@ -159,13 +159,36 @@ std::vector<Sequence*> Cursor::join_inners(Sequence* outer_seq) {
     }
   }
 
+  std::vector<Shape*> to_delay;
+  to_delay = connect_missings(missing_shapes);
+  while (!to_delay.empty()) {
+    to_delay = connect_missings(to_delay);
+  }
+
+  retme = collect_inner_sequences(outer_seq);
+  for (Polyline* polyline : polylines_sequence) {
+    polyline->turn_on(Polyline::TRACKED_INNER);
+  }
+  return(retme);
+}
+
+std::vector<Shape*> Cursor::connect_missings(std::vector<Shape*> missing_shapes) {
+  std::vector<Shape*> delay_shapes;
+
   for (Polyline *polyline : polylines_sequence)
   { for (Shape *missing_shape : missing_shapes)
     { Polyline* outer_polyline = missing_shape->outer_polyline;
-      if (!polyline->vert_intersect(*outer_polyline)) continue;
-      std::vector<Point*> intersection = polyline->intersection(outer_polyline);
+      if ( outer_polyline->is_on(Polyline::TRACKED_OUTER)  ||
+           !polyline->vert_intersect(*outer_polyline)) continue;
+      std::vector<std::pair<int, int>> intersection = polyline->intersection(outer_polyline);
       if (intersection.size() > 0)
-      { auto [inject_sequences_left, inject_sequences_right] = polyline->sew(intersection, outer_polyline);
+      { auto result = polyline->sew(intersection, outer_polyline);
+        if (!result) {
+          delay_shapes.push_back(missing_shape);
+          continue;
+        }
+        auto& inject_sequences_left = result->first;
+        auto& inject_sequences_right = result->second;
         auto combined = combine(inject_sequences_right, inject_sequences_left);
         for (auto& sewn_sequence : combined) {
           std::vector<Point*> unique;
@@ -205,11 +228,8 @@ std::vector<Sequence*> Cursor::join_inners(Sequence* outer_seq) {
       }
     }
   }
-  retme = collect_inner_sequences(outer_seq);
-  for (Polyline* polyline : polylines_sequence) {
-    polyline->turn_on(Polyline::TRACKED_INNER);
-  }
-  return(retme);
+
+  return(delay_shapes);
 }
 
 std::vector<Sequence*> Cursor::collect_inner_sequences(Sequence* outer_seq) {

data/ext/cpp_polygon_finder/PolygonFinder/src/polygon/finder/concurrent/Cursor.h

@@ -43,4 +43,5 @@ class Cursor {
   void traverse_inner(Part* act_part, std::vector<Part*> &all_parts, Bounds& bounds);
   std::vector<Point*> duplicates_intersection(const Part& part_a, const Part& part_b);
   std::vector<std::vector<Point*>> combine(std::vector<std::vector<Point*>>& seqa, std::vector<std::vector<Point*>>& seqb);
+  std::vector<Shape*> connect_missings(std::vector<Shape*> missing_shapes);
 };

data/ext/cpp_polygon_finder/PolygonFinder/src/polygon/finder/concurrent/Finder.cpp

@@ -86,7 +86,17 @@ void Finder::process_tiles() {
 
     if (it != arriving_tiles.end()) {
       Tile* twin_tile = *it;
-      Cluster *cluster = new Cluster(this, this->bitmap->h(), this->bitmap->w());
+      int start_x, end_x;
+      if (twin_tile->start_x() == (tile->end_x() - 1)) {
+        start_x = tile->start_x();
+        end_x = twin_tile->end_x();
+      } else {
+        start_x = twin_tile->start_x();
+        end_x = tile->end_x();
+      }
+
+      Cluster *cluster = new Cluster(this, this->bitmap->h(), start_x, end_x);
+
       if (twin_tile->start_x() == (tile->end_x() - 1)) {
         cluster->add(tile);
         cluster->add(twin_tile);
@@ -124,6 +134,7 @@ Finder::~Finder() {
 ProcessResult* Finder::process_info() {
   ProcessResult *pr = new ProcessResult();
   pr->polygons = std::move(this->whole_tile->to_raw_polygons());
+  pr->groups = pr->polygons.size();
   FakeCluster fake_cluster(pr->polygons, this->options);
   cpu_timer.start();
   fake_cluster.compress_coords(pr->polygons, this->options);

data/ext/cpp_polygon_finder/PolygonFinder/src/polygon/finder/concurrent/Hub.cpp

@@ -9,15 +9,15 @@
 #include "Hub.h"
 #include <cstring>
 
-Hub::Hub(int height, int width)
-    : width_(width), height_(height)
-{ size_t total_pixels = static_cast<size_t>(width) * height;
+Hub::Hub(int height, int start_x, int end_x)
+    : width_(end_x - start_x), height_(height), start_x_(start_x)
+{ size_t total_pixels = static_cast<size_t>(width_) * height;
   payloads_.resize(total_pixels);
   bitset_.resize((total_pixels + 63) / 64, 0);
   std::memset(bitset_.data(), 0, bitset_.size() * sizeof(uint64_t));
 }
 
-EndPoint* Hub::spawn_end_point() {
+int Hub::spawn_end_point() {
   endpoint_pool_.emplace_back();
-  return &endpoint_pool_.back();
+  return static_cast<int>(endpoint_pool_.size() - 1);
 }

data/ext/cpp_polygon_finder/PolygonFinder/src/polygon/finder/concurrent/Hub.h

@@ -14,16 +14,19 @@
 class EndPoint;
 class Hub {
  public:
-  explicit Hub(int height, int width);
-  EndPoint* spawn_end_point();
+  explicit Hub(int height, int start_x, int end_x);
+  int spawn_end_point();
   const int width() const { return width_; }
-  inline EndPoint* get(int key) const {
+  const int start_x() const { return start_x_; }
+  inline EndPoint* get(int key) {
     if (!is_set(key)) return nullptr;
-    return payloads_[key];
+    int index = payloads_[key];
+    return &endpoint_pool_[index];
   }
-  inline void put(int key, EndPoint* payload) {
-    payloads_[key] = payload;
+  inline EndPoint* put(int key, int index) {
+    payloads_[key] = index;
     bitset_[key >> 6] |= (1ULL << (key & 63));
+    return &endpoint_pool_[index];
   }
   inline bool is_set(int key) const {
     return (bitset_[key >> 6] & (1ULL << (key & 63)));
@@ -32,7 +35,8 @@ class Hub {
  private:
   int width_;
   int height_;
-  std::vector<EndPoint*> payloads_;
+  int start_x_;
+  std::vector<int> payloads_;
   std::deque<EndPoint> endpoint_pool_;
   Hub(const Hub&) = delete;
   Hub& operator=(const Hub&) = delete;

data/ext/cpp_polygon_finder/PolygonFinder/src/polygon/finder/concurrent/Part.h

@@ -27,6 +27,7 @@ class Part : public Queueable<Point> {
   bool is(Types type);
   bool inverts = false;
   bool trasmuted = false;
+  bool delayed = false;
   Part* next = nullptr;
   Part* prev = nullptr;
   Part* circular_next = nullptr;
@@ -40,6 +41,7 @@ class Part : public Queueable<Point> {
   const bool touched() const { return touched_; }
   void touch();
   bool intersect_part(Part* other_part);
+  void set_polyline(Polyline* polyline) { this->polyline_ = polyline; }
 
  private:
   bool touched_ = false;