xgb 0.3.1 → 0.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 295787749986ba153354f61845509079f470a5d11015c20cb89a5c58f04b133a
-  data.tar.gz: e111099514e9a9a48d51a5033214a3e897eeff209302e47e1f30d79c818833c5
+  metadata.gz: 0d128d5d386edca28045cf55d4cd246f68deee2bb8e4a26254b9549955d2d8d4
+  data.tar.gz: 9e43a00b4b5f5f451ba7744effae622cfb404e67038de3881a6edfc5a776878a
 SHA512:
-  metadata.gz: 803003072dafa5d2de97da7721042b94315b125f62f4f524f818923c5e5f8d3277624b146cedd7b88d9ca0769bec074e99717fb27ce56f5d2bf079ba627d52f7
-  data.tar.gz: b9b14da27b6f8c03415d770c31dd16fdd112922daa2af3022df8554301eb5e43d0e873f21c65da13799e7a62d832cadbabff50188783aad5b5e7a577a9334005
+  metadata.gz: c7aaa6dd25ca9987621314b7ccee0b4fb029e0e6b70cce50a7fdaf9517c8896c6ad961b36b45af0ed2e6e2698bfd6d523a0fc623202cc058aa671821d9cec893
+  data.tar.gz: 603488b7920eed380a8eb5ca3576e3d5a1f4b7ba9cdc8afa75c0385558606b7e2865c8ddb9c21db43725b3cf379f75beb8cba35cf92d7d4717c85fedc7bf8ccb

data/CHANGELOG.md

@@ -1,3 +1,11 @@
+## 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`

data/README.md

@@ -137,16 +137,22 @@ Data can be an array of arrays
 [[1, 2, 3], [4, 5, 6]]
 ```
 
-Or a Daru data frame
+Or a Numo NArray
 
 ```ruby
-Daru::DataFrame.from_csv("houses.csv")
+Numo::DFloat.new(3, 2).seq
 ```
 
-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

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

@@ -27,18 +27,25 @@ module XGBoost
                 raise Error, "Unknown feature type: #{v}"
               end
             end
-        elsif narray?(data)
+        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))
@@ -134,10 +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

@@ -2,7 +2,15 @@ module XGBoost
   module FFI
     extend ::FFI::Library
 
-    ffi_lib XGBoost.ffi_lib
+    begin
+      ffi_lib XGBoost.ffi_lib
+    rescue LoadError => e
+      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

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 = "0.3.1"
+  VERSION = "0.4.0"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: xgb
 version: !ruby/object:Gem::Version
-  version: 0.3.1
+  version: 0.4.0
 platform: ruby
 authors:
 - Andrew Kane
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-04-16 00:00:00.000000000 Z
+date: 2020-05-17 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: []