polars-df 0.5.0-x86_64-linux → 0.7.0-x86_64-linux

data/lib/polars/list_name_space.rb

@@ -1,9 +1,9 @@
 module Polars
-  # Series.arr namespace.
+  # Series.list namespace.
   class ListNameSpace
     include ExprDispatch
 
-    self._accessor = "arr"
+    self._accessor = "list"
 
     # @private
     def initialize(series)
@@ -16,7 +16,7 @@ module Polars
     #
     # @example
     #   s = Polars::Series.new([[1, 2, 3], [5]])
-    #   s.arr.lengths
+    #   s.list.lengths
     #   # =>
     #   # shape: (2,)
     #   # Series: '' [u32]
@@ -119,13 +119,13 @@ module Polars
     #
     # @example
     #   s = Polars::Series.new([["foo", "bar"], ["hello", "world"]])
-    #   s.arr.join("-")
+    #   s.list.join("-")
     #   # =>
     #   # shape: (2,)
     #   # Series: '' [str]
     #   # [
-    #   #     "foo-bar"
-    #   #     "hello-world"
+    #   #         "foo-bar"
+    #   #         "hello-world"
     #   # ]
     def join(separator)
       super
@@ -180,7 +180,7 @@ module Polars
     #
     # @example
     #   s = Polars::Series.new("a", [[1, 2, 3, 4], [10, 2, 1]])
-    #   s.arr.diff
+    #   s.list.diff
     #   # =>
     #   # shape: (2,)
     #   # Series: 'a' [list[i64]]
@@ -201,7 +201,7 @@ module Polars
     #
     # @example
     #   s = Polars::Series.new("a", [[1, 2, 3, 4], [10, 2, 1]])
-    #   s.arr.shift
+    #   s.list.shift
     #   # =>
     #   # shape: (2,)
     #   # Series: 'a' [list[i64]]
@@ -225,7 +225,7 @@ module Polars
     #
     # @example
     #   s = Polars::Series.new("a", [[1, 2, 3, 4], [10, 2, 1]])
-    #   s.arr.slice(1, 2)
+    #   s.list.slice(1, 2)
     #   # =>
     #   # shape: (2,)
     #   # Series: 'a' [list[i64]]
@@ -246,7 +246,7 @@ module Polars
     #
     # @example
     #   s = Polars::Series.new("a", [[1, 2, 3, 4], [10, 2, 1]])
-    #   s.arr.head(2)
+    #   s.list.head(2)
     #   # =>
     #   # shape: (2,)
     #   # Series: 'a' [list[i64]]
@@ -267,7 +267,7 @@ module Polars
     #
     # @example
     #   s = Polars::Series.new("a", [[1, 2, 3, 4], [10, 2, 1]])
-    #   s.arr.tail(2)
+    #   s.list.tail(2)
     #   # =>
     #   # shape: (2,)
     #   # Series: 'a' [list[i64]]
@@ -291,7 +291,7 @@ module Polars
     #
     # @example
     #   df = Polars::DataFrame.new({"a" => [[1, 2, 3], [1, 2]]})
-    #   df.select([Polars.col("a").arr.to_struct])
+    #   df.select([Polars.col("a").list.to_struct])
     #   # =>
     #   # shape: (2, 1)
     #   # ┌────────────┐
@@ -315,7 +315,7 @@ module Polars
     #   Run all expression parallel. Don't activate this blindly.
     #   Parallelism is worth it if there is enough work to do per thread.
     #
-    #   This likely should not be use in the groupby context, because we already
+    #   This likely should not be use in the group by context, because we already
     #   parallel execution per group
     #
     # @return [Series]
@@ -323,14 +323,14 @@ module Polars
     # @example
     #   df = Polars::DataFrame.new({"a" => [1, 8, 3], "b" => [4, 5, 2]})
     #   df.with_column(
-    #     Polars.concat_list(["a", "b"]).arr.eval(Polars.element.rank).alias("rank")
+    #     Polars.concat_list(["a", "b"]).list.eval(Polars.element.rank).alias("rank")
     #   )
     #   # =>
     #   # shape: (3, 3)
     #   # ┌─────┬─────┬────────────┐
     #   # │ a   ┆ b   ┆ rank       │
     #   # │ --- ┆ --- ┆ ---        │
-    #   # │ i64 ┆ i64 ┆ list[f32]  │
+    #   # │ i64 ┆ i64 ┆ list[f64]  │
     #   # ╞═════╪═════╪════════════╡
     #   # │ 1   ┆ 4   ┆ [1.0, 2.0] │
     #   # │ 8   ┆ 5   ┆ [2.0, 1.0] │

data/lib/polars/name_expr.rb

@@ -0,0 +1,198 @@
+module Polars
+  # Namespace for expressions that operate on expression names.
+  class NameExpr
+    # @private
+    attr_accessor :_rbexpr
+
+    # @private
+    def initialize(expr)
+      self._rbexpr = expr._rbexpr
+    end
+
+    # Keep the original root name of the expression.
+    #
+    # @note
+    #   Due to implementation constraints, this method can only be called as the last
+    #   expression in a chain.
+    #
+    # @return [Expr]
+    #
+    # @example Prevent errors due to potential duplicate column names.
+    #   df = Polars::DataFrame.new(
+    #     {
+    #       "a" => [1, 2],
+    #       "b" => [3, 4]
+    #     }
+    #   )
+    #   df.select((Polars.lit(10) / Polars.all).name.keep)
+    #   # =>
+    #   # shape: (2, 2)
+    #   # ┌──────┬──────────┐
+    #   # │ a    ┆ b        │
+    #   # │ ---  ┆ ---      │
+    #   # │ f64  ┆ f64      │
+    #   # ╞══════╪══════════╡
+    #   # │ 10.0 ┆ 3.333333 │
+    #   # │ 5.0  ┆ 2.5      │
+    #   # └──────┴──────────┘
+    #
+    # @example Undo an alias operation.
+    #   df.with_columns((Polars.col("a") * 9).alias("c").name.keep)
+    #   # =>
+    #   # shape: (2, 2)
+    #   # ┌─────┬─────┐
+    #   # │ a   ┆ b   │
+    #   # │ --- ┆ --- │
+    #   # │ i64 ┆ i64 │
+    #   # ╞═════╪═════╡
+    #   # │ 9   ┆ 3   │
+    #   # │ 18  ┆ 4   │
+    #   # └─────┴─────┘
+    def keep
+      Utils.wrap_expr(_rbexpr.name_keep)
+    end
+
+    # Rename the output of an expression by mapping a function over the root name.
+    #
+    # @return [Expr]
+    #
+    # @example Remove a common suffix and convert to lower case.
+    #   df = Polars::DataFrame.new(
+    #     {
+    #       "A_reverse" => [3, 2, 1],
+    #       "B_reverse" => ["z", "y", "x"]
+    #     }
+    #   )
+    #   df.with_columns(
+    #     Polars.all.reverse.name.map { |c| c.delete_suffix("_reverse").downcase }
+    #   )
+    #   # =>
+    #   # shape: (3, 4)
+    #   # ┌───────────┬───────────┬─────┬─────┐
+    #   # │ A_reverse ┆ B_reverse ┆ a   ┆ b   │
+    #   # │ ---       ┆ ---       ┆ --- ┆ --- │
+    #   # │ i64       ┆ str       ┆ i64 ┆ str │
+    #   # ╞═══════════╪═══════════╪═════╪═════╡
+    #   # │ 3         ┆ z         ┆ 1   ┆ x   │
+    #   # │ 2         ┆ y         ┆ 2   ┆ y   │
+    #   # │ 1         ┆ x         ┆ 3   ┆ z   │
+    #   # └───────────┴───────────┴─────┴─────┘
+    def map(&f)
+      Utils.wrap_expr(_rbexpr.name_map(f))
+    end
+
+    # Add a prefix to the root column name of the expression.
+    #
+    # @param prefix [Object]
+    #   Prefix to add to the root column name.
+    #
+    # @return [Expr]
+    #
+    # @example
+    #   df = Polars::DataFrame.new(
+    #     {
+    #       "a" => [1, 2, 3],
+    #       "b" => ["x", "y", "z"]
+    #     }
+    #   )
+    #   df.with_columns(Polars.all.reverse.name.prefix("reverse_"))
+    #   # =>
+    #   # shape: (3, 4)
+    #   # ┌─────┬─────┬───────────┬───────────┐
+    #   # │ a   ┆ b   ┆ reverse_a ┆ reverse_b │
+    #   # │ --- ┆ --- ┆ ---       ┆ ---       │
+    #   # │ i64 ┆ str ┆ i64       ┆ str       │
+    #   # ╞═════╪═════╪═══════════╪═══════════╡
+    #   # │ 1   ┆ x   ┆ 3         ┆ z         │
+    #   # │ 2   ┆ y   ┆ 2         ┆ y         │
+    #   # │ 3   ┆ z   ┆ 1         ┆ x         │
+    #   # └─────┴─────┴───────────┴───────────┘
+    def prefix(prefix)
+      Utils.wrap_expr(_rbexpr.name_prefix(prefix))
+    end
+
+    # Add a suffix to the root column name of the expression.
+    #
+    # @param suffix [Object]
+    #   Suffix to add to the root column name.
+    #
+    # @return [Expr]
+    #
+    # @example
+    #   df = Polars::DataFrame.new(
+    #     {
+    #       "a" => [1, 2, 3],
+    #       "b" => ["x", "y", "z"]
+    #     }
+    #   )
+    #   df.with_columns(Polars.all.reverse.name.suffix("_reverse"))
+    #   # =>
+    #   # shape: (3, 4)
+    #   # ┌─────┬─────┬───────────┬───────────┐
+    #   # │ a   ┆ b   ┆ a_reverse ┆ b_reverse │
+    #   # │ --- ┆ --- ┆ ---       ┆ ---       │
+    #   # │ i64 ┆ str ┆ i64       ┆ str       │
+    #   # ╞═════╪═════╪═══════════╪═══════════╡
+    #   # │ 1   ┆ x   ┆ 3         ┆ z         │
+    #   # │ 2   ┆ y   ┆ 2         ┆ y         │
+    #   # │ 3   ┆ z   ┆ 1         ┆ x         │
+    #   # └─────┴─────┴───────────┴───────────┘
+    def suffix(suffix)
+      Utils.wrap_expr(_rbexpr.name_suffix(suffix))
+    end
+
+    # Make the root column name lowercase.
+    #
+    # @return [Expr]
+    #
+    # @example
+    #   df = Polars::DataFrame.new(
+    #     {
+    #       "ColX" => [1, 2, 3],
+    #       "ColY" => ["x", "y", "z"],
+    #     }
+    #   )
+    #   df.with_columns(Polars.all.name.to_lowercase)
+    #   # =>
+    #   # shape: (3, 4)
+    #   # ┌──────┬──────┬──────┬──────┐
+    #   # │ ColX ┆ ColY ┆ colx ┆ coly │
+    #   # │ ---  ┆ ---  ┆ ---  ┆ ---  │
+    #   # │ i64  ┆ str  ┆ i64  ┆ str  │
+    #   # ╞══════╪══════╪══════╪══════╡
+    #   # │ 1    ┆ x    ┆ 1    ┆ x    │
+    #   # │ 2    ┆ y    ┆ 2    ┆ y    │
+    #   # │ 3    ┆ z    ┆ 3    ┆ z    │
+    #   # └──────┴──────┴──────┴──────┘
+    def to_lowercase
+      Utils.wrap_expr(_rbexpr.name_to_lowercase)
+    end
+
+    # Make the root column name uppercase.
+    #
+    # @return [Expr]
+    #
+    # @example
+    #   df = Polars::DataFrame.new(
+    #     {
+    #       "ColX" => [1, 2, 3],
+    #       "ColY" => ["x", "y", "z"]
+    #     }
+    #   )
+    #   df.with_columns(Polars.all.name.to_uppercase)
+    #   # =>
+    #   # shape: (3, 4)
+    #   # ┌──────┬──────┬──────┬──────┐
+    #   # │ ColX ┆ ColY ┆ COLX ┆ COLY │
+    #   # │ ---  ┆ ---  ┆ ---  ┆ ---  │
+    #   # │ i64  ┆ str  ┆ i64  ┆ str  │
+    #   # ╞══════╪══════╪══════╪══════╡
+    #   # │ 1    ┆ x    ┆ 1    ┆ x    │
+    #   # │ 2    ┆ y    ┆ 2    ┆ y    │
+    #   # │ 3    ┆ z    ┆ 3    ┆ z    │
+    #   # └──────┴──────┴──────┴──────┘
+    def to_uppercase
+      Utils.wrap_expr(_rbexpr.name_to_uppercase)
+    end
+  end
+end

data/lib/polars/rolling_group_by.rb

@@ -2,7 +2,7 @@ module Polars
   # A rolling grouper.
   #
   # This has an `.agg` method which will allow you to run all polars expressions in a
-  # groupby context.
+  # group by context.
   class RollingGroupBy
     def initialize(
       df,
@@ -10,7 +10,8 @@ module Polars
       period,
       offset,
       closed,
-      by
+      by,
+      check_sorted
     )
       period = Utils._timedelta_to_pl_duration(period)
       offset = Utils._timedelta_to_pl_duration(offset)
@@ -21,12 +22,13 @@ module Polars
       @offset = offset
       @closed = closed
       @by = by
+      @check_sorted = check_sorted
     end
 
     def agg(aggs)
       @df.lazy
-        .groupby_rolling(
-          index_column: @time_column, period: @period, offset: @offset, closed: @closed, by: @by
+        .group_by_rolling(
+          index_column: @time_column, period: @period, offset: @offset, closed: @closed, by: @by, check_sorted: @check_sorted
         )
         .agg(aggs)
         .collect(no_optimization: true, string_cache: false)

data/lib/polars/series.rb

@@ -65,7 +65,7 @@ module Polars
           )
           .rename(name, in_place: true)
           ._s
-      elsif values.is_a?(Array)
+      elsif values.is_a?(::Array)
         self._s = sequence_to_rbseries(name, values, dtype: dtype, strict: strict, dtype_if_empty: dtype_if_empty)
       elsif defined?(Numo::NArray) && values.is_a?(Numo::NArray)
         self._s = numo_to_rbseries(name, values, strict: strict, nan_to_null: nan_to_null)
@@ -317,6 +317,10 @@ module Polars
       end
 
       if item.is_a?(Integer)
+        if item < 0
+          item = len + item
+        end
+
         return _s.get_idx(item)
       end
 
@@ -335,7 +339,7 @@ module Polars
     #
     # @return [Object]
     def []=(key, value)
-      if value.is_a?(Array)
+      if value.is_a?(::Array)
         if is_numeric || is_datelike
           set_at_idx(key, value)
           return
@@ -353,7 +357,7 @@ module Polars
         else
           raise Todo
         end
-      elsif key.is_a?(Array)
+      elsif key.is_a?(::Array)
         s = Utils.wrap_s(sequence_to_rbseries("", key, dtype: UInt32))
         self[s] = value
       elsif key.is_a?(Range)
@@ -428,6 +432,18 @@ module Polars
     end
     alias_method :all, :all?
 
+    # Check if all boolean values in the column are `false`.
+    #
+    # @return [Boolean]
+    def none?(&block)
+      if block_given?
+        apply(&block).none?
+      else
+        to_frame.select(Polars.col(name).is_not.all).to_series[0]
+      end
+    end
+    alias_method :none, :none?
+
     # Compute the logarithm to a given base.
     #
     # @param base [Float]
@@ -715,8 +731,8 @@ module Polars
     #   # │ 0   ┆ 1   ┆ 0   │
     #   # │ 0   ┆ 0   ┆ 1   │
     #   # └─────┴─────┴─────┘
-    def to_dummies(separator: "_")
-      Utils.wrap_df(_s.to_dummies(separator))
+    def to_dummies(separator: "_", drop_first: false)
+      Utils.wrap_df(_s.to_dummies(separator, drop_first))
     end
 
     # Count the unique values in a Series.
@@ -795,7 +811,7 @@ module Polars
     #   Number of valid values there should be in the window before the expression
     #   is evaluated. valid values = `length - null_count`
     # @param parallel [Boolean]
-    #   Run in parallel. Don't do this in a groupby or another operation that
+    #   Run in parallel. Don't do this in a group by or another operation that
     #   already has much parallelization.
     #
     # @return [Series]
@@ -1124,7 +1140,7 @@ module Polars
     #   #         3
     #   # ]
     def filter(predicate)
-      if predicate.is_a?(Array)
+      if predicate.is_a?(::Array)
         predicate = Series.new("", predicate)
       end
       Utils.wrap_s(_s.filter(predicate._s))
@@ -2813,7 +2829,8 @@ module Polars
       window_size,
       weights: nil,
       min_periods: nil,
-      center: false
+      center: false,
+      ddof: 1
     )
       to_frame
         .select(
@@ -2821,7 +2838,8 @@ module Polars
             window_size,
             weights: weights,
             min_periods: min_periods,
-            center: center
+            center: center,
+            ddof: ddof
           )
         )
         .to_series
@@ -2864,7 +2882,8 @@ module Polars
       window_size,
       weights: nil,
       min_periods: nil,
-      center: false
+      center: false,
+      ddof: 1
     )
       to_frame
         .select(
@@ -2872,7 +2891,8 @@ module Polars
             window_size,
             weights: weights,
             min_periods: min_periods,
-            center: center
+            center: center,
+            ddof: ddof
           )
         )
         .to_series
@@ -3089,7 +3109,7 @@ module Polars
     #   s.peak_max
     #   # =>
     #   # shape: (5,)
-    #   # Series: '' [bool]
+    #   # Series: 'a' [bool]
     #   # [
     #   #         false
     #   #         false
@@ -3098,7 +3118,7 @@ module Polars
     #   #         true
     #   # ]
     def peak_max
-      Utils.wrap_s(_s.peak_max)
+      super
     end
 
     # Get a boolean mask of the local minimum peaks.
@@ -3110,7 +3130,7 @@ module Polars
     #   s.peak_min
     #   # =>
     #   # shape: (5,)
-    #   # Series: '' [bool]
+    #   # Series: 'a' [bool]
     #   # [
     #   #         false
     #   #         true
@@ -3119,7 +3139,7 @@ module Polars
     #   #         false
     #   # ]
     def peak_min
-      Utils.wrap_s(_s.peak_min)
+      super
     end
 
     # Count the number of unique values in this Series.
@@ -3203,13 +3223,13 @@ module Polars
     #   s.interpolate
     #   # =>
     #   # shape: (5,)
-    #   # Series: 'a' [i64]
+    #   # Series: 'a' [f64]
     #   # [
-    #   #         1
-    #   #         2
-    #   #         3
-    #   #         4
-    #   #         5
+    #   #         1.0
+    #   #         2.0
+    #   #         3.0
+    #   #         4.0
+    #   #         5.0
     #   # ]
     def interpolate(method: "linear")
       super
@@ -3252,7 +3272,7 @@ module Polars
     #   s.rank
     #   # =>
     #   # shape: (5,)
-    #   # Series: 'a' [f32]
+    #   # Series: 'a' [f64]
     #   # [
     #   #         3.0
     #   #         4.5
@@ -3581,10 +3601,17 @@ module Polars
     # Create an object namespace of all list related methods.
     #
     # @return [ListNameSpace]
-    def arr
+    def list
       ListNameSpace.new(self)
     end
 
+    # Create an object namespace of all array related methods.
+    #
+    # @return [ArrayNameSpace]
+    def arr
+      ArrayNameSpace.new(self)
+    end
+
     # Create an object namespace of all binary related methods.
     #
     # @return [BinaryNameSpace]
@@ -3824,9 +3851,12 @@ module Polars
 
       if (values.nil? || values.empty?) && dtype.nil?
         dtype = dtype_if_empty || Float32
+      elsif dtype == List
+        ruby_dtype = ::Array
       end
 
       rb_temporal_types = [::Date, ::DateTime, ::Time]
+      rb_temporal_types << ActiveSupport::TimeWithZone if defined?(ActiveSupport::TimeWithZone)
 
       value = _get_first_non_none(values)
       if !value.nil?
@@ -3835,9 +3865,20 @@ module Polars
         end
       end
 
-      if !dtype.nil? && Utils.is_polars_dtype(dtype) && ruby_dtype.nil?
+      if !dtype.nil? && ![List, Unknown].include?(dtype) && Utils.is_polars_dtype(dtype) && ruby_dtype.nil?
+        if dtype == Array && !dtype.is_a?(Array) && value.is_a?(::Array)
+          dtype = Array.new(value.size)
+        end
+
         constructor = polars_type_to_constructor(dtype)
         rbseries = constructor.call(name, values, strict)
+
+        base_type = dtype.is_a?(DataType) ? dtype.class : dtype
+        if [Date, Datetime, Duration, Time, Categorical, Boolean].include?(base_type)
+          if rbseries.dtype != dtype
+            rbseries = rbseries.cast(dtype, true)
+          end
+        end
         return rbseries
       else
         if ruby_dtype.nil?
@@ -3868,7 +3909,17 @@ module Polars
           return s._s
         elsif defined?(Numo::NArray) && value.is_a?(Numo::NArray) && value.shape.length == 1
           raise Todo
-        elsif ruby_dtype == Array
+        elsif ruby_dtype == ::Array
+          if dtype.is_a?(Object)
+            return RbSeries.new_object(name, values, strict)
+          end
+          if dtype
+            srs = sequence_from_anyvalue_or_object(name, values)
+            if dtype != srs.dtype
+              srs = srs.cast(dtype, strict: false)
+            end
+            return srs
+          end
           return sequence_from_anyvalue_or_object(name, values)
         elsif ruby_dtype == Series
           return RbSeries.new_series_list(name, values.map(&:_s), strict)
@@ -3910,9 +3961,17 @@ module Polars
       UInt16 => RbSeries.method(:new_opt_u16),
       UInt32 => RbSeries.method(:new_opt_u32),
       UInt64 => RbSeries.method(:new_opt_u64),
+      Decimal => RbSeries.method(:new_decimal),
+      Date => RbSeries.method(:new_from_anyvalues),
+      Datetime => RbSeries.method(:new_from_anyvalues),
+      Duration => RbSeries.method(:new_from_anyvalues),
+      Time => RbSeries.method(:new_from_anyvalues),
       Boolean => RbSeries.method(:new_opt_bool),
       Utf8 => RbSeries.method(:new_str),
-      Binary => RbSeries.method(:new_binary)
+      Object => RbSeries.method(:new_object),
+      Categorical => RbSeries.method(:new_str),
+      Binary => RbSeries.method(:new_binary),
+      Null => RbSeries.method(:new_null)
     }
 
     SYM_TYPE_TO_CONSTRUCTOR = {
@@ -3931,8 +3990,14 @@ module Polars
     }
 
     def polars_type_to_constructor(dtype)
-      if dtype.is_a?(Class) && dtype < DataType
+      if dtype.is_a?(Array)
+        lambda do |name, values, strict|
+          RbSeries.new_array(dtype.width, dtype.inner, name, values, strict)
+        end
+      elsif dtype.is_a?(Class) && dtype < DataType
         POLARS_TYPE_TO_CONSTRUCTOR.fetch(dtype)
+      elsif dtype.is_a?(DataType)
+        POLARS_TYPE_TO_CONSTRUCTOR.fetch(dtype.class)
       else
         SYM_TYPE_TO_CONSTRUCTOR.fetch(dtype.to_sym)
       end
@@ -3944,7 +4009,9 @@ module Polars
       Float => RbSeries.method(:new_opt_f64),
       Integer => RbSeries.method(:new_opt_i64),
       TrueClass => RbSeries.method(:new_opt_bool),
-      FalseClass => RbSeries.method(:new_opt_bool)
+      FalseClass => RbSeries.method(:new_opt_bool),
+      BigDecimal => RbSeries.method(:new_decimal),
+      NilClass => RbSeries.method(:new_null)
     }
 
     def rb_type_to_constructor(dtype)