rc105

Author: brettshollenberger

Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    easy_ml (0.2.0.pre.rc104)
+    easy_ml (0.2.0.pre.rc105)
       activerecord
       activerecord-import (~> 1.8.1)
       activesupport

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

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

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

@@ -6,27 +6,27 @@ class Numeric < Query
           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"),
             ])

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

@@ -33,6 +33,14 @@ def execute(split)
             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 @@ def train_query
 
           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 @@ def last_value
             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

easy_ml-0.2.0.pre.rc105.gem

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
 
 module EasyML
-  VERSION = "0.2.0-rc104"
+  VERSION = "0.2.0-rc105"
 
   module Version
   end

lib/easy_ml/version.rb

@@ -400,7 +400,7 @@
                                                                          ])
 
         column.update(datatype: "null")
-        expect(column.errors.map(&:message)).to include("Can't cast from float to null")
+        expect(column.raw.data["Age"].dtype).to eq(Polars::Null)
       end
     end
   end