easy_ml 0.2.0.pre.rc103 → 0.2.0.pre.rc105

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ef3f840cce99d7205957fbb39a6b319a45035624dce2e4e10f681383cb088abf
-  data.tar.gz: e25100f792ad48cfa4feab7eb652a2d6c49bfc6e28f3bcb97c8150f9bdd1bfc5
+  metadata.gz: c0f70db7d7b59423ef34b5f3cd15d783ec22112977b44c1a5c7df390cabbc1c9
+  data.tar.gz: c985391e9231db8870c3a0dde14d41c9a0b75f0643b9ebb4ad30af7b57354230
 SHA512:
-  metadata.gz: 5f58395d392158d149db34ad5019a0e011164ca8d331846553e44e6564a291d88323ad0090c1c5ded60f696940b30949cba4e1a614fa9cd502e94372ef949707
-  data.tar.gz: 9497391351ad054308a985cc6b9e608f8dfef61be7417d66502cb11c26ca4f7825456b31aab010c016ac10feff791a8f7a01893743ecf11a26fabb9de7405b82
+  metadata.gz: d06176713f32434dc24b7250090a7c005b79a0c515d96e7ef35b67a28e9c5c6feb58ed2bb6a5136f98897c9ded66b8c9c5ee78018cb8056c50ef1f2491b26e0a
+  data.tar.gz: 8a049b2858d28e0f4eb7776a606a1e3375535c7ab0a13d41ed32a5c253b9fc979d9dcf705c0c7e99322340b8c0d3b21c38513885e79947347efd1b9bafb66347

data/app/models/easy_ml/column.rb

@@ -528,6 +528,18 @@ module EasyML
 
       return Polars.col(name).cast(expected_dtype).alias(name) if expected_dtype == actual_type
 
+      if encoding.present?
+        encoding_cast = case encoding.to_sym
+        when :one_hot
+          Polars.col(series.name).cast(Polars::Boolean).alias(series.name)
+        when :ordinal
+          Polars.col(series.name).cast(Polars::Int64).alias(series.name)
+        when :embedding
+          Polars.col(series.name).alias(series.name)
+        end
+        return encoding_cast
+      end
+
       cast_statement = case expected_dtype.to_s
                         when /Polars::List/
                           # we should start tracking polars args so we can know what type of list it is

data/app/models/easy_ml/dataset/learner/lazy/datetime.rb

@@ -11,7 +11,7 @@ module EasyML
 
           def unique_count
             Polars.col(column.name)
-                  .cast(column.polars_datatype)
+                  .cast(datatype)
                   .n_unique.alias("#{column.name}__unique_count")
           end
         end

data/app/models/easy_ml/dataset/learner/lazy/numeric.rb

@@ -6,27 +6,27 @@ module EasyML
           def train_query
             super.concat([
               Polars.col(column.name)
-                    .cast(column.polars_datatype)
+                    .cast(datatype)
                     .mean
                     .alias("#{column.name}__mean"),
 
               Polars.col(column.name)
-                      .cast(column.polars_datatype)
+                      .cast(datatype)
                       .median
                       .alias("#{column.name}__median"),
 
               Polars.col(column.name)
-                    .cast(column.polars_datatype)
+                    .cast(datatype)
                     .min
                     .alias("#{column.name}__min"),
 
               Polars.col(column.name)
-                    .cast(column.polars_datatype)
+                    .cast(datatype)
                     .max
                     .alias("#{column.name}__max"),
 
               Polars.col(column.name)
-                    .cast(column.polars_datatype)
+                    .cast(datatype)
                     .std
                     .alias("#{column.name}__std"),
             ])

data/app/models/easy_ml/dataset/learner/lazy/query.rb

@@ -33,6 +33,14 @@ module EasyML
             end
           end
 
+          def datatype
+            case column.polars_datatype.to_s
+            when /Polars::Categorical/ then Polars::String
+            else
+              column.polars_datatype
+            end
+          end
+
           private
 
           def full_dataset_query
@@ -45,21 +53,21 @@ module EasyML
 
           def null_count
             Polars.col(column.name)
-                  .cast(column.polars_datatype)
+                  .cast(datatype)
                   .null_count
                   .alias("#{column.name}__null_count")
           end
 
           def num_rows
             Polars.col(column.name)
-                  .cast(column.polars_datatype)
+                  .cast(datatype)
                   .len
                   .alias("#{column.name}__num_rows")
           end
 
           def most_frequent_value
             Polars.col(column.name)
-                  .cast(column.polars_datatype)
+                  .cast(datatype)
                   .filter(Polars.col(column.name).is_not_null)
                   .mode
                   .first
@@ -70,7 +78,7 @@ module EasyML
             return unless dataset.date_column.present?
 
             Polars.col(column.name)
-                  .cast(column.polars_datatype)
+                  .cast(datatype)
                   .sort_by(dataset.date_column.name, reverse: true, nulls_last: true)
                   .filter(Polars.col(column.name).is_not_null)
                   .first

data/app/models/easy_ml/models/xgboost.rb

@@ -498,7 +498,7 @@ module EasyML
         feature_cols -= [weights_col] if weights_col
 
         # Get features, labels and weights
-        exploded = explode_embeddings(xs)
+        exploded = explode_embeddings(xs.select(feature_cols))
         feature_cols = exploded.columns
         features = lazy ? exploded.collect.to_numo : exploded.to_numo
 

data/lib/easy_ml/version.rb

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
 
 module EasyML
-  VERSION = "0.2.0-rc103"
+  VERSION = "0.2.0-rc105"
 
   module Version
   end

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: easy_ml
 version: !ruby/object:Gem::Version
-  version: 0.2.0.pre.rc103
+  version: 0.2.0.pre.rc105
 platform: ruby
 authors:
 - Brett Shollenberger
 bindir: exe
 cert_chain: []
-date: 2025-03-18 00:00:00.000000000 Z
+date: 2025-03-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activerecord