xgb 0.2.0 → 0.4.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ea879f0494a877766dd500203b7f4ec091607a0a0cae609bc3be50fb134b87eb
-  data.tar.gz: 1ff52f49405628aa837f4f6ff1558b4b1a624a311b4c359c0a0703499613d35d
+  metadata.gz: bb82540880bbecdc88d82eb71b0d3fd3e7cf276f05205ddc0f1900684c5602a2
+  data.tar.gz: 1463e06dce0ae99fdee5ccc1887d1a24d537fcdac7d89fb685701566083d5600
 SHA512:
-  metadata.gz: cd2dcbbc2300bbb5900ae1cdf8abe47fd2c0f0c8dc1b825f53c5988450405298e5b8693be243ed351f7f14715e9f5963eb7c00cd39c1b25983b449398804e6f5
-  data.tar.gz: 969c4472a3926f8cd6569269817086ed047e4e64132fd92ab1bb78ab5e8f294c3a4342cfc4e121b8a2866b8f534d1ffaa35b806d9968f198f87bb4de06a64348
+  metadata.gz: 4869e2465af9d824e56dfd180eff0a7b8aacc0dd9788cfb9f5d429e83c51b5bceb93b8ad93fcdb0b7b2e0ec81231204d32f738988af97304b6d2eee33fa2f709
+  data.tar.gz: 4206063987a450cbb82bb9d2ea0cc7cfa8bf8bb9df483d4f088943cc41ba9d396afef81136f9090e65249ed2a3d78806c3fca70de5e198a15f6199ed8b3a01e8

data/CHANGELOG.md

@@ -1,3 +1,29 @@
+## 0.4.1 (2020-08-26)
+
+- Updated XGBoost to 1.2.0
+
+## 0.4.0 (2020-05-17)
+
+- Updated XGBoost to 1.1.0
+- Changed default `learning_rate` and `max_depth` for Scikit-Learn API to match Python
+- Added support for Rover
+- Improved performance of Numo datasets
+- Improved error message when OpenMP not found on Mac
+
+## 0.3.1 (2020-04-16)
+
+- Added `feature_names` and `feature_types` to `DMatrix`
+- Added feature names to `dump`
+
+## 0.3.0 (2020-02-19)
+
+- Updated XGBoost to 1.0.0
+
+## 0.2.1 (2020-02-11)
+
+- Fixed `Could not find XGBoost` error on some Linux platforms
+- Fixed `SignalException` on Windows
+
 ## 0.2.0 (2020-01-26)
 
 - Prefer `XGBoost` over `Xgb`

data/NOTICE.txt

@@ -1,3 +1,4 @@
+Copyright XGBoost contributors
 Copyright 2019-2020 Andrew Kane
 
 Licensed under the Apache License, Version 2.0 (the "License");

data/README.md

@@ -2,7 +2,7 @@
 
 [XGBoost](https://github.com/dmlc/xgboost) - high performance gradient boosting - for Ruby
 
-[![Build Status](https://travis-ci.org/ankane/xgboost.svg?branch=master)](https://travis-ci.org/ankane/xgboost)
+[![Build Status](https://travis-ci.org/ankane/xgboost.svg?branch=master)](https://travis-ci.org/ankane/xgboost) [![Build status](https://ci.appveyor.com/api/projects/status/s8umwyuahvj68m6p/branch/master?svg=true)](https://ci.appveyor.com/project/ankane/xgboost/branch/master)
 
 ## Installation
 
@@ -12,9 +12,11 @@ Add this line to your application’s Gemfile:
 gem 'xgb'
 ```
 
-## Getting Started
+On Mac, also install OpenMP:
 
-This library follows the [Python API](https://xgboost.readthedocs.io/en/latest/python/python_api.html), with the `get_` and `set_` prefixes removed from methods to make it more Ruby-like.
+```sh
+brew install libomp
+```
 
 ## Learning API
 
@@ -70,7 +72,7 @@ CV
 XGBoost.cv(params, dtrain, nfold: 3, verbose_eval: true)
 ```
 
-Set metadata about a model [master]
+Set metadata about a model
 
 ```ruby
 booster["key"] = "value"
@@ -135,16 +137,22 @@ Data can be an array of arrays
 [[1, 2, 3], [4, 5, 6]]
 ```
 
-Or a Daru data frame
+Or a Numo array
 
 ```ruby
-Daru::DataFrame.from_csv("houses.csv")
+Numo::NArray.cast([[1, 2, 3], [4, 5, 6]])
 ```
 
-Or a Numo NArray
+Or a Rover data frame
 
 ```ruby
-Numo::DFloat.new(3, 2).seq
+Rover.read_csv("houses.csv")
+```
+
+Or a Daru data frame
+
+```ruby
+Daru::DataFrame.from_csv("houses.csv")
 ```
 
 ## Helpful Resources
@@ -155,11 +163,13 @@ Numo::DFloat.new(3, 2).seq
 ## Related Projects
 
 - [LightGBM](https://github.com/ankane/lightgbm) - LightGBM for Ruby
-- [Eps](https://github.com/ankane/eps) - Machine Learning for Ruby
+- [Eps](https://github.com/ankane/eps) - Machine learning for Ruby
 
 ## Credits
 
-Thanks to the [xgboost](https://github.com/PairOnAir/xgboost-ruby) gem for serving as an initial reference.
+This library follows the [Python API](https://xgboost.readthedocs.io/en/latest/python/python_api.html), with the `get_` and `set_` prefixes removed from methods to make it more Ruby-like.
+
+Thanks to the [xgboost](https://github.com/PairOnAir/xgboost-ruby) gem for showing how to use FFI.
 
 ## History
 
@@ -174,11 +184,12 @@ Everyone is encouraged to help improve this project. Here are a few ways you can
 - Write, clarify, or fix documentation
 - Suggest or add new features
 
-To get started with development and testing:
+To get started with development:
 
 ```sh
 git clone https://github.com/ankane/xgboost.git
 cd xgboost
 bundle install
+bundle exec rake vendor:all
 bundle exec rake test
 ```

data/lib/xgboost.rb

@@ -31,7 +31,8 @@ module XGBoost
       booster = Booster.new(params: params)
       num_feature = dtrain.num_col
       booster.set_param("num_feature", num_feature)
-      booster.feature_names = num_feature.times.map { |i| "f#{i}" }
+      booster.feature_names = dtrain.feature_names
+      booster.feature_types = dtrain.feature_types
       evals ||= []
 
       if early_stopping_rounds
@@ -156,6 +157,14 @@ module XGBoost
       eval_hist
     end
 
+    def lib_version
+      major = ::FFI::MemoryPointer.new(:int)
+      minor = ::FFI::MemoryPointer.new(:int)
+      patch = ::FFI::MemoryPointer.new(:int)
+      FFI.XGBoostVersion(major, minor, patch)
+      "#{major.read_int}.#{minor.read_int}.#{patch.read_int}"
+    end
+
     private
 
     def mean(arr)

data/lib/xgboost/booster.rb

@@ -1,6 +1,6 @@
 module XGBoost
   class Booster
-    attr_accessor :best_iteration, :feature_names
+    attr_accessor :best_iteration, :feature_names, :feature_types
 
     def initialize(params: nil, model_file: nil)
       @handle = ::FFI::MemoryPointer.new(:pointer)
@@ -25,11 +25,8 @@ module XGBoost
     end
 
     def eval_set(evals, iteration)
-      dmats = ::FFI::MemoryPointer.new(:pointer, evals.size)
-      dmats.write_array_of_pointer(evals.map { |v| v[0].handle_pointer })
-
-      evnames = ::FFI::MemoryPointer.new(:pointer, evals.size)
-      evnames.write_array_of_pointer(evals.map { |v| ::FFI::MemoryPointer.from_string(v[1]) })
+      dmats = array_of_pointers(evals.map { |v| v[0].handle_pointer })
+      evnames = array_of_pointers(evals.map { |v| string_pointer(v[1]) })
 
       out_result = ::FFI::MemoryPointer.new(:pointer)
 
@@ -52,7 +49,7 @@ module XGBoost
       ntree_limit ||= 0
       out_len = ::FFI::MemoryPointer.new(:uint64)
       out_result = ::FFI::MemoryPointer.new(:pointer)
-      check_result FFI.XGBoosterPredict(handle_pointer, data.handle_pointer, 0, ntree_limit, out_len, out_result)
+      check_result FFI.XGBoosterPredict(handle_pointer, data.handle_pointer, 0, ntree_limit, 0, out_len, out_result)
       out = out_result.read_pointer.read_array_of_float(read_uint64(out_len))
       num_class = out.size / data.num_row
       out = out.each_slice(num_class).to_a if num_class > 1
@@ -67,7 +64,13 @@ module XGBoost
     def dump(fmap: "", with_stats: false, dump_format: "text")
       out_len = ::FFI::MemoryPointer.new(:uint64)
       out_result = ::FFI::MemoryPointer.new(:pointer)
-      check_result FFI.XGBoosterDumpModelEx(handle_pointer, fmap, with_stats ? 1 : 0, dump_format, out_len, out_result)
+
+      names = feature_names || []
+      fnames = array_of_pointers(names.map { |fname| string_pointer(fname) })
+      ftypes = array_of_pointers(feature_types || Array.new(names.size, string_pointer("float")))
+
+      check_result FFI.XGBoosterDumpModelExWithFeatures(handle_pointer, names.size, fnames, ftypes, with_stats ? 1 : 0, dump_format, out_len, out_result)
+
       out_result.read_pointer.get_array_of_string(0, read_uint64(out_len))
     end
 
@@ -155,7 +158,7 @@ module XGBoost
     end
 
     def [](key_name)
-      key = ::FFI::MemoryPointer.from_string(key_name)
+      key = string_pointer(key_name)
       success = ::FFI::MemoryPointer.new(:int)
       out_result = ::FFI::MemoryPointer.new(:pointer)
 
@@ -165,8 +168,8 @@ module XGBoost
     end
 
     def []=(key_name, raw_value)
-      key = ::FFI::MemoryPointer.from_string(key_name)
-      value = raw_value.nil? ? nil : ::FFI::MemoryPointer.from_string(raw_value)
+      key = string_pointer(key_name)
+      value = raw_value.nil? ? nil : string_pointer(raw_value)
 
       check_result FFI.XGBoosterSetAttr(handle_pointer, key, value)
     end
@@ -188,6 +191,14 @@ module XGBoost
       @handle.read_pointer
     end
 
+    def array_of_pointers(values)
+      ::FFI::MemoryPointer.new(:pointer, values.size).write_array_of_pointer(values)
+    end
+
+    def string_pointer(value)
+      ::FFI::MemoryPointer.from_string(value.to_s)
+    end
+
     include Utils
   end
 end

data/lib/xgboost/classifier.rb

@@ -1,6 +1,6 @@
 module XGBoost
   class Classifier < Model
-    def initialize(max_depth: 3, learning_rate: 0.1, n_estimators: 100, objective: "binary:logistic", importance_type: "gain", **options)
+    def initialize(n_estimators: 100, objective: "binary:logistic", importance_type: "gain", **options)
       super
     end
 

data/lib/xgboost/dmatrix.rb

@@ -1,6 +1,6 @@
 module XGBoost
   class DMatrix
-    attr_reader :data
+    attr_reader :data, :feature_names, :feature_types
 
     def initialize(data, label: nil, weight: nil, missing: Float::NAN)
       @data = data
@@ -15,21 +15,42 @@ module XGBoost
         elsif daru?(data)
           nrow, ncol = data.shape
           flat_data = data.map_rows(&:to_a).flatten
-        elsif narray?(data)
+          @feature_names = data.each_vector.map(&:name)
+          @feature_types =
+            data.each_vector.map(&:db_type).map do |v|
+              case v
+              when "INTEGER"
+                "int"
+              when "DOUBLE"
+                "float"
+              else
+                raise Error, "Unknown feature type: #{v}"
+              end
+            end
+        elsif numo?(data)
           nrow, ncol = data.shape
-          flat_data = data.flatten.to_a
+        elsif rover?(data)
+          nrow, ncol = data.shape
+          @feature_names = data.keys
+          data = data.to_numo
         else
           nrow = data.count
           ncol = data.first.count
           flat_data = data.flatten
         end
 
-        handle_missing(flat_data, missing)
         c_data = ::FFI::MemoryPointer.new(:float, nrow * ncol)
-        c_data.write_array_of_float(flat_data)
+        if numo?(data)
+          c_data.write_bytes(data.cast_to(Numo::SFloat).to_string)
+        else
+          handle_missing(flat_data, missing)
+          c_data.write_array_of_float(flat_data)
+        end
         check_result FFI.XGDMatrixCreateFromMat(c_data, nrow, ncol, missing, @handle)
 
         ObjectSpace.define_finalizer(self, self.class.finalize(handle_pointer))
+
+        @feature_names ||= ncol.times.map { |i| "f#{i}" }
       end
 
       self.label = label if label
@@ -60,7 +81,7 @@ module XGBoost
     def group=(group)
       c_data = ::FFI::MemoryPointer.new(:int, group.size)
       c_data.write_array_of_int(group)
-      check_result FFI.XGDMatrixSetGroup(handle_pointer, c_data, group.size)
+      check_result FFI.XGDMatrixSetUIntInfo(handle_pointer, "group", c_data, group.size)
     end
 
     def num_row
@@ -120,10 +141,14 @@ module XGBoost
       defined?(Daru::DataFrame) && data.is_a?(Daru::DataFrame)
     end
 
-    def narray?(data)
+    def numo?(data)
       defined?(Numo::NArray) && data.is_a?(Numo::NArray)
     end
 
+    def rover?(data)
+      defined?(Rover::DataFrame) && data.is_a?(Rover::DataFrame)
+    end
+
     def handle_missing(data, missing)
       data.map! { |v| v.nil? ? missing : v }
     end

data/lib/xgboost/ffi.rb

@@ -5,19 +5,23 @@ module XGBoost
     begin
       ffi_lib XGBoost.ffi_lib
     rescue LoadError => e
-      raise e if ENV["XGB_DEBUG"]
-      raise LoadError, "Could not find XGBoost"
+      if e.message.include?("Library not loaded: /usr/local/opt/libomp/lib/libomp.dylib") && e.message.include?("Reason: image not found")
+        raise LoadError, "OpenMP not found. Run `brew install libomp`"
+      else
+        raise e
+      end
     end
 
     # https://github.com/dmlc/xgboost/blob/master/include/xgboost/c_api.h
     # keep same order
 
-    # error
+    # general
+    attach_function :XGBoostVersion, %i[pointer pointer pointer], :void
     attach_function :XGBGetLastError, %i[], :string
 
     # dmatrix
     attach_function :XGDMatrixCreateFromMat, %i[pointer uint64 uint64 float pointer], :int
-    attach_function :XGDMatrixSetGroup, %i[pointer pointer uint64], :int
+    attach_function :XGDMatrixSetUIntInfo, %i[pointer string pointer uint64], :int
     attach_function :XGDMatrixNumRow, %i[pointer pointer], :int
     attach_function :XGDMatrixNumCol, %i[pointer pointer], :int
     attach_function :XGDMatrixSliceDMatrix, %i[pointer pointer uint64 pointer], :int
@@ -32,10 +36,10 @@ module XGBoost
     attach_function :XGBoosterEvalOneIter, %i[pointer int pointer pointer uint64 pointer], :int
     attach_function :XGBoosterFree, %i[pointer], :int
     attach_function :XGBoosterSetParam, %i[pointer string string], :int
-    attach_function :XGBoosterPredict, %i[pointer pointer int int pointer pointer], :int
+    attach_function :XGBoosterPredict, %i[pointer pointer int int int pointer pointer], :int
     attach_function :XGBoosterLoadModel, %i[pointer string], :int
     attach_function :XGBoosterSaveModel, %i[pointer string], :int
-    attach_function :XGBoosterDumpModelEx, %i[pointer string int string pointer pointer], :int
+    attach_function :XGBoosterDumpModelExWithFeatures, %i[pointer int pointer pointer int string pointer pointer], :int
     attach_function :XGBoosterGetAttr, %i[pointer pointer pointer pointer], :int
     attach_function :XGBoosterSetAttr, %i[pointer pointer pointer], :int
     attach_function :XGBoosterGetAttrNames, %i[pointer pointer pointer], :int

data/lib/xgboost/model.rb

@@ -2,12 +2,8 @@ module XGBoost
   class Model
     attr_reader :booster
 
-    def initialize(max_depth: 3, learning_rate: 0.1, n_estimators: 100, objective: nil, importance_type: "gain", **options)
-      @params = {
-        max_depth: max_depth,
-        objective: objective,
-        learning_rate: learning_rate
-      }.merge(options)
+    def initialize(n_estimators: 100, importance_type: "gain", **options)
+      @params = options
       @n_estimators = n_estimators
       @importance_type = importance_type
     end

data/lib/xgboost/ranker.rb

@@ -1,6 +1,6 @@
 module XGBoost
   class Ranker < Model
-    def initialize(max_depth: 3, learning_rate: 0.1, n_estimators: 100, objective: "rank:pairwise", importance_type: "gain", **options)
+    def initialize(n_estimators: 100, objective: "rank:pairwise", importance_type: "gain", **options)
       super
     end
 

data/lib/xgboost/regressor.rb

@@ -1,6 +1,6 @@
 module XGBoost
   class Regressor < Model
-    def initialize(max_depth: 3, learning_rate: 0.1, n_estimators: 100, objective: "reg:squarederror", importance_type: "gain", **options)
+    def initialize(n_estimators: 100, objective: "reg:squarederror", importance_type: "gain", **options)
       super
     end
 

data/lib/xgboost/version.rb

@@ -1,3 +1,3 @@
 module XGBoost
-  VERSION = "0.2.0"
+  VERSION = "0.4.1"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: xgb
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.4.1
 platform: ruby
 authors:
 - Andrew Kane
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-01-27 00:00:00.000000000 Z
+date: 2020-08-27 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ffi
@@ -94,6 +94,20 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
+- !ruby/object:Gem::Dependency
+  name: rover-df
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
 description: 
 email: andrew@chartkick.com
 executables: []