polars-df 0.14.0 → 0.15.0

data/lib/polars/string_cache.rb

@@ -1,11 +1,37 @@
 module Polars
-  # Context manager for enabling and disabling the global string cache.
+  # Class for enabling and disabling the global string cache.
+  #
+  # @example Construct two Series using the same global string cache.
+  #   s1 = nil
+  #   s2 = nil
+  #   Polars::StringCache.new do
+  #     s1 = Polars::Series.new("color", ["red", "green", "red"], dtype: Polars::Categorical)
+  #     s2 = Polars::Series.new("color", ["blue", "red", "green"], dtype: Polars::Categorical)
+  #   end
+  #
+  # @example As both Series are constructed under the same global string cache, they can be concatenated.
+  #   Polars.concat([s1, s2])
+  #   # =>
+  #   # shape: (6,)
+  #   # Series: 'color' [cat]
+  #   # [
+  #   #         "red"
+  #   #         "green"
+  #   #         "red"
+  #   #         "blue"
+  #   #         "red"
+  #   #         "green"
+  #   # ]
   class StringCache
     def initialize(&block)
       RbStringCacheHolder.hold(&block)
     end
   end
 
+  def self.string_cache(...)
+    StringCache.new(...)
+  end
+
   module Functions
     # Enable the global string cache.
     #

data/lib/polars/string_expr.rb

@@ -421,7 +421,6 @@ module Polars
     #   # │ THERE'S NO TURNING BACK ┆ There's No Turning Back │
     #   # └─────────────────────────┴─────────────────────────┘
     def to_titlecase
-      raise Todo
       Utils.wrap_expr(_rbexpr.str_to_titlecase)
     end
 

data/lib/polars/string_name_space.rb

@@ -658,7 +658,18 @@ module Polars
     #   An optional single character that should be trimmed
     #
     # @return [Series]
-    def strip(matches = nil)
+    #
+    # @example
+    #   s = Polars::Series.new([" hello ", "\tworld"])
+    #   s.str.strip_chars
+    #   # =>
+    #   # shape: (2,)
+    #   # Series: '' [str]
+    #   # [
+    #   #         "hello"
+    #   #         "world"
+    #   # ]
+    def strip_chars(matches = nil)
       super
     end
 
@@ -668,9 +679,21 @@ module Polars
     #   An optional single character that should be trimmed
     #
     # @return [Series]
-    def lstrip(matches = nil)
+    #
+    # @example
+    #   s = Polars::Series.new([" hello ", "\tworld"])
+    #   s.str.strip_chars_start
+    #   # =>
+    #   # shape: (2,)
+    #   # Series: '' [str]
+    #   # [
+    #   #         "hello "
+    #   #         "world"
+    #   # ]
+    def strip_chars_start(matches = nil)
       super
     end
+    alias_method :lstrip, :strip_chars_start
 
     # Remove trailing whitespace.
     #
@@ -678,9 +701,21 @@ module Polars
     #   An optional single character that should be trimmed
     #
     # @return [Series]
-    def rstrip(matches = nil)
+    #
+    # @example
+    #   s = Polars::Series.new([" hello ", "world\t"])
+    #   s.str.strip_chars_end
+    #   # =>
+    #   # shape: (2,)
+    #   # Series: '' [str]
+    #   # [
+    #   #         " hello"
+    #   #         "world"
+    #   # ]
+    def strip_chars_end(matches = nil)
       super
     end
+    alias_method :rstrip, :strip_chars_end
 
     # Fills the string with zeroes.
     #
@@ -695,6 +730,19 @@ module Polars
     #   Fill the value up to this length.
     #
     # @return [Series]
+    #
+    # @example
+    #   s = Polars::Series.new([-1, 123, 999999, nil])
+    #   s.cast(Polars::String).str.zfill(4)
+    #   # =>
+    #   # shape: (4,)
+    #   # Series: '' [str]
+    #   # [
+    #   #         "-001"
+    #   #         "0123"
+    #   #         "999999"
+    #   #         null
+    #   # ]
     def zfill(length)
       super
     end
@@ -758,6 +806,17 @@ module Polars
     # Modify the strings to their lowercase equivalent.
     #
     # @return [Series]
+    #
+    # @example
+    #   s = Polars::Series.new("foo", ["CAT", "DOG"])
+    #   s.str.to_lowercase
+    #   # =>
+    #   # shape: (2,)
+    #   # Series: 'foo' [str]
+    #   # [
+    #   #         "cat"
+    #   #         "dog"
+    #   # ]
     def to_lowercase
       super
     end
@@ -765,6 +824,17 @@ module Polars
     # Modify the strings to their uppercase equivalent.
     #
     # @return [Series]
+    #
+    # @example
+    #   s = Polars::Series.new("foo", ["cat", "dog"])
+    #   s.str.to_uppercase
+    #   # =>
+    #   # shape: (2,)
+    #   # Series: 'foo' [str]
+    #   # [
+    #   #         "CAT"
+    #   #         "DOG"
+    #   # ]
     def to_uppercase
       super
     end

data/lib/polars/struct_name_space.rb

@@ -23,16 +23,14 @@ module Polars
       end
     end
 
-    # Convert this Struct Series to a DataFrame.
-    #
-    # @return [DataFrame]
-    def to_frame
-      Utils.wrap_df(_s.struct_to_frame)
-    end
-
     # Get the names of the fields.
     #
     # @return [Array]
+    #
+    # @example
+    #   s = Polars::Series.new([{"a" => 1, "b" => 2}, {"a" => 3, "b" => 4}])
+    #   s.struct.fields
+    #   # => ["a", "b"]
     def fields
       if _s.nil?
         []
@@ -47,6 +45,17 @@ module Polars
     #   Name of the field
     #
     # @return [Series]
+    #
+    # @example
+    #   s = Polars::Series.new([{"a" => 1, "b" => 2}, {"a" => 3, "b" => 4}])
+    #   s.struct.field("a")
+    #   # =>
+    #   # shape: (2,)
+    #   # Series: 'a' [i64]
+    #   # [
+    #   #         1
+    #   #         3
+    #   # ]
     def field(name)
       super
     end
@@ -57,6 +66,16 @@ module Polars
     #   New names in the order of the struct's fields
     #
     # @return [Series]
+    #
+    # @example
+    #   s = Polars::Series.new([{"a" => 1, "b" => 2}, {"a" => 3, "b" => 4}])
+    #   s.struct.fields
+    #   # => ["a", "b"]
+    #
+    # @example
+    #   s = s.struct.rename_fields(["c", "d"])
+    #   s.struct.fields
+    #   # => ["c", "d"]
     def rename_fields(names)
       super
     end
@@ -64,6 +83,11 @@ module Polars
     # Get the struct definition as a name/dtype schema dict.
     #
     # @return [Object]
+    #
+    # @example
+    #   s = Polars::Series.new([{"a" => 1, "b" => 2}, {"a" => 3, "b" => 4}])
+    #   s.struct.schema
+    #   # => {"a"=>Polars::Int64, "b"=>Polars::Int64}
     def schema
       if _s.nil?
         {}

data/lib/polars/utils/various.rb

@@ -12,6 +12,10 @@ module Polars
       val.all? { |x| x.is_a?(eltype) }
     end
 
+    def self.is_path_or_str_sequence(val)
+      val.is_a?(::Array) && val.all? { |x| pathlike?(x) }
+    end
+
     def self.is_bool_sequence(val)
       val.is_a?(::Array) && val.all? { |x| x == true || x == false }
     end
@@ -42,7 +46,7 @@ module Polars
     end
 
     def self.normalize_filepath(path, check_not_directory: true)
-      path = File.expand_path(path)
+      path = File.expand_path(path) if !path.is_a?(::String) || path.start_with?("~")
       if check_not_directory && File.exist?(path) && Dir.exist?(path)
         raise ArgumentError, "Expected a file path; #{path} is a directory"
       end

data/lib/polars/utils.rb

@@ -3,12 +3,18 @@ module Polars
   module Utils
     DTYPE_TEMPORAL_UNITS = ["ns", "us", "ms"]
 
-    # TODO fix
-    def self.is_polars_dtype(data_type, include_unknown: false)
-      if data_type == Unknown
-        return include_unknown
+    def self.is_polars_dtype(dtype, include_unknown: false)
+      is_dtype = dtype.is_a?(Symbol) || dtype.is_a?(::String) || dtype.is_a?(DataType) || (dtype.is_a?(Class) && dtype < DataType)
+
+      if !include_unknown
+        is_dtype && dtype != Unknown
+      else
+        is_dtype
       end
-      data_type.is_a?(Symbol) || data_type.is_a?(::String) || data_type.is_a?(DataType) || (data_type.is_a?(Class) && data_type < DataType)
+    end
+
+    def self.is_column(obj)
+      obj.is_a?(Expr) && obj.meta.is_column
     end
 
     def self.map_rb_type_to_dtype(ruby_dtype)
@@ -100,6 +106,27 @@ module Polars
       end
     end
 
+    def self.is_selector(obj)
+      obj.is_a?(Selectors::SelectorProxy)
+    end
+
+    def self.expand_selector(target, selector, strict: true)
+      if target.is_a?(Hash)
+        target = DataFrame.new(schema: target)
+      end
+
+      if !is_selector(selector) && !is_polars_dtype(selector)
+        msg = "expected a selector; found #{selector.inspect} instead."
+        raise TypeError, msg
+      end
+
+      if is_selector(selector)
+        target.select(selector).columns
+      else
+        target.select(Polars.col(selector)).columns
+      end
+    end
+
     def self._expand_selectors(frame, *items)
       items_iter = _parse_inputs_as_iterable(items)
 
@@ -115,16 +142,24 @@ module Polars
       expanded
     end
 
-    # TODO
-    def self.is_selector(obj)
-      false
-    end
-
     def self.parse_interval_argument(interval)
       if interval.include?(" ")
         interval = interval.gsub(" ", "")
       end
       interval.downcase
     end
+
+    def self.parse_into_dtype(input)
+      if is_polars_dtype(input)
+        input
+      else
+        raise Todo
+      end
+    end
+
+    def self.re_escape(s)
+      # escapes _only_ those metachars with meaning to the rust regex crate
+      Plr.re_escape(s)
+    end
   end
 end

data/lib/polars/version.rb

@@ -1,4 +1,4 @@
 module Polars
   # @private
-  VERSION = "0.14.0"
+  VERSION = "0.15.0"
 end

data/lib/polars.rb

@@ -24,12 +24,12 @@ require_relative "polars/convert"
 require_relative "polars/plot"
 require_relative "polars/data_frame"
 require_relative "polars/data_types"
+require_relative "polars/data_type_group"
 require_relative "polars/date_time_expr"
 require_relative "polars/date_time_name_space"
 require_relative "polars/dynamic_group_by"
 require_relative "polars/exceptions"
 require_relative "polars/expr"
-require_relative "polars/functions"
 require_relative "polars/functions/as_datatype"
 require_relative "polars/functions/col"
 require_relative "polars/functions/eager"
@@ -60,6 +60,7 @@ require_relative "polars/list_name_space"
 require_relative "polars/meta_expr"
 require_relative "polars/name_expr"
 require_relative "polars/rolling_group_by"
+require_relative "polars/selectors"
 require_relative "polars/series"
 require_relative "polars/slice"
 require_relative "polars/sql_context"

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: polars-df
 version: !ruby/object:Gem::Version
-  version: 0.14.0
+  version: 0.15.0
 platform: ruby
 authors:
 - Andrew Kane
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-09-17 00:00:00.000000000 Z
+date: 2024-11-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bigdecimal
@@ -63,7 +63,9 @@ files:
 - ext/polars/src/dataframe/general.rs
 - ext/polars/src/dataframe/io.rs
 - ext/polars/src/dataframe/mod.rs
+- ext/polars/src/dataframe/serde.rs
 - ext/polars/src/error.rs
+- ext/polars/src/exceptions.rs
 - ext/polars/src/expr/array.rs
 - ext/polars/src/expr/binary.rs
 - ext/polars/src/expr/categorical.rs
@@ -78,6 +80,7 @@ files:
 - ext/polars/src/expr/struct.rs
 - ext/polars/src/file.rs
 - ext/polars/src/functions/aggregation.rs
+- ext/polars/src/functions/business.rs
 - ext/polars/src/functions/eager.rs
 - ext/polars/src/functions/io.rs
 - ext/polars/src/functions/lazy.rs
@@ -88,11 +91,15 @@ files:
 - ext/polars/src/functions/range.rs
 - ext/polars/src/functions/string_cache.rs
 - ext/polars/src/functions/whenthen.rs
+- ext/polars/src/interop/arrow/mod.rs
+- ext/polars/src/interop/arrow/to_ruby.rs
 - ext/polars/src/interop/mod.rs
 - ext/polars/src/interop/numo/mod.rs
 - ext/polars/src/interop/numo/to_numo_df.rs
 - ext/polars/src/interop/numo/to_numo_series.rs
+- ext/polars/src/lazyframe/general.rs
 - ext/polars/src/lazyframe/mod.rs
+- ext/polars/src/lazyframe/serde.rs
 - ext/polars/src/lazygroupby.rs
 - ext/polars/src/lib.rs
 - ext/polars/src/map/dataframe.rs
@@ -108,6 +115,8 @@ files:
 - ext/polars/src/series/comparison.rs
 - ext/polars/src/series/construction.rs
 - ext/polars/src/series/export.rs
+- ext/polars/src/series/general.rs
+- ext/polars/src/series/import.rs
 - ext/polars/src/series/mod.rs
 - ext/polars/src/series/scatter.rs
 - ext/polars/src/sql.rs
@@ -124,6 +133,7 @@ files:
 - lib/polars/config.rb
 - lib/polars/convert.rb
 - lib/polars/data_frame.rb
+- lib/polars/data_type_group.rb
 - lib/polars/data_types.rb
 - lib/polars/date_time_expr.rb
 - lib/polars/date_time_name_space.rb
@@ -131,7 +141,6 @@ files:
 - lib/polars/exceptions.rb
 - lib/polars/expr.rb
 - lib/polars/expr_dispatch.rb
-- lib/polars/functions.rb
 - lib/polars/functions/aggregation/horizontal.rb
 - lib/polars/functions/aggregation/vertical.rb
 - lib/polars/functions/as_datatype.rb
@@ -163,6 +172,7 @@ files:
 - lib/polars/name_expr.rb
 - lib/polars/plot.rb
 - lib/polars/rolling_group_by.rb
+- lib/polars/selectors.rb
 - lib/polars/series.rb
 - lib/polars/slice.rb
 - lib/polars/sql_context.rb
@@ -199,7 +209,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.5.16
+rubygems_version: 3.5.22
 signing_key:
 specification_version: 4
 summary: Blazingly fast DataFrames for Ruby

data/lib/polars/functions.rb

@@ -1,57 +0,0 @@
-module Polars
-  module Functions
-    # Convert categorical variables into dummy/indicator variables.
-    #
-    # @param df [DataFrame]
-    #   DataFrame to convert.
-    # @param columns [Array, nil]
-    #   A subset of columns to convert to dummy variables. `nil` means
-    #   "all columns".
-    #
-    # @return [DataFrame]
-    def get_dummies(df, columns: nil)
-      df.to_dummies(columns: columns)
-    end
-
-    # Aggregate to list.
-    #
-    # @return [Expr]
-    def to_list(name)
-      col(name).list
-    end
-
-    # Compute the spearman rank correlation between two columns.
-    #
-    # Missing data will be excluded from the computation.
-    #
-    # @param a [Object]
-    #   Column name or Expression.
-    # @param b [Object]
-    #   Column name or Expression.
-    # @param ddof [Integer]
-    #   Delta degrees of freedom
-    # @param propagate_nans [Boolean]
-    #   If `True` any `NaN` encountered will lead to `NaN` in the output.
-    #   Defaults to `False` where `NaN` are regarded as larger than any finite number
-    #   and thus lead to the highest rank.
-    #
-    # @return [Expr]
-    def spearman_rank_corr(a, b, ddof: 1, propagate_nans: false)
-      corr(a, b, method: "spearman", ddof: ddof, propagate_nans: propagate_nans)
-    end
-
-    # Compute the pearson's correlation between two columns.
-    #
-    # @param a [Object]
-    #   Column name or Expression.
-    # @param b [Object]
-    #   Column name or Expression.
-    # @param ddof [Integer]
-    #   Delta degrees of freedom
-    #
-    # @return [Expr]
-    def pearson_corr(a, b, ddof: 1)
-      corr(a, b, method: "pearson", ddof: ddof)
-    end
-  end
-end