polars-df 0.2.0 → 0.2.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ca8491538991cd918a2706c572c43994ba00558937636d7d90bb98b51f27cb76
-  data.tar.gz: a8d6667309db4c17c807d0b2a398c6acdbcb2a9c4282d07a313f0de57d10b6b7
+  metadata.gz: df03134e7edf09e86b5a4f4f9ae9a926bac4c9c0804a29c3422c32675f478825
+  data.tar.gz: e0338be1aa96d0ad082ebf8fe27e608b2906b243dd49fa837aceb7f8186947d8
 SHA512:
-  metadata.gz: b26a1722981f2616f979d8a46590cba04d3fd4f420f95923693fd3796787153bc5197b3308ee93b83a06a14cc4bcbba52f9c3046f9f366e6599c987dc8d0fd4c
-  data.tar.gz: 0da65a6f3543d500a4d62e2d5816043060156d6b94f02750d50036ecc406adbaa037e3ecefe5d23e667ce814db6580fc8336e8a4f52beceda85551c6dd4a0c96
+  metadata.gz: 75a139d30f9fdebaa84a21fa45cec8a199da76eb295e7099ceb849646a93fbc7ed80ffed18aaa8eb7bbfc53a32792b2e47101485ad31d727a47ed67d8d7e8110
+  data.tar.gz: 589f7fbc1300aadc05568308700f6a94b934e63c40bd1be0a3e7b6f564c0d55f256e2e45e926c128d80453d0e7d200b057f640b02cd6fb9aaddf5bf55dd89754

data/CHANGELOG.md

@@ -1,3 +1,15 @@
+## 0.2.2 (2023-01-20)
+
+- Added support for strings to `read_sql` method
+- Improved indexing
+- Fixed error with precompiled gem on Mac ARM
+
+## 0.2.1 (2023-01-18)
+
+- Added `read_sql` method
+- Added `to_csv` method
+- Added support for symbol keys
+
 ## 0.2.0 (2023-01-14)
 
 - Updated Polars to 0.26.1

data/Cargo.lock

@@ -1367,7 +1367,7 @@ dependencies = [
 
 [[package]]
 name = "polars"
-version = "0.2.0"
+version = "0.2.2"
 dependencies = [
  "ahash",
  "jemallocator",

data/README.md

@@ -25,7 +25,13 @@ Polars.read_csv("iris.csv")
   .collect
 ```
 
-You can follow [Polars tutorials](https://pola-rs.github.io/polars-book/user-guide/introduction.html) and convert the code to Ruby in many cases. Feel free to open an issue if you run into problems. Some methods are missing at the moment.
+You can follow [Polars tutorials](https://pola-rs.github.io/polars-book/user-guide/introduction.html) and convert the code to Ruby in many cases. Feel free to open an issue if you run into problems.
+
+## Reference
+
+- [Series](https://www.rubydoc.info/gems/polars-df/Polars/Series)
+- [DataFrame](https://www.rubydoc.info/gems/polars-df/Polars/DataFrame)
+- [LazyFrame](https://www.rubydoc.info/gems/polars-df/Polars/LazyFrame)
 
 ## Examples
 
@@ -35,6 +41,9 @@ From a CSV
 
 ```ruby
 Polars.read_csv("file.csv")
+
+# or lazily with
+Polars.scan_csv("file.csv")
 ```
 
 From Parquet
@@ -46,7 +55,7 @@ Polars.read_parquet("file.parquet")
 From Active Record
 
 ```ruby
-Polars::DataFrame.new(User.all)
+Polars.read_sql(User.all)
 ```
 
 From a hash
@@ -67,6 +76,261 @@ Polars::DataFrame.new([
 ])
 ```
 
+## Attributes
+
+Get number of rows
+
+```ruby
+df.height
+```
+
+Get column names
+
+```ruby
+df.columns
+```
+
+Check if a column exists
+
+```ruby
+df.include?(name)
+```
+
+## Selecting Data
+
+Select a column
+
+```ruby
+df["a"]
+```
+
+Select multiple columns
+
+```ruby
+df[["a", "b"]]
+```
+
+Select first rows
+
+```ruby
+df.head
+```
+
+Select last rows
+
+```ruby
+df.tail
+```
+
+## Filtering
+
+Filter on a condition
+
+```ruby
+df[Polars.col("a") == 2]
+df[Polars.col("a") != 2]
+df[Polars.col("a") > 2]
+df[Polars.col("a") >= 2]
+df[Polars.col("a") < 2]
+df[Polars.col("a") <= 2]
+```
+
+And, or, and exclusive or
+
+```ruby
+df[(Polars.col("a") > 1) & (Polars.col("b") == "two")] # and
+df[(Polars.col("a") > 1) | (Polars.col("b") == "two")] # or
+df[(Polars.col("a") > 1) ^ (Polars.col("b") == "two")] # xor
+```
+
+## Operations
+
+Basic operations
+
+```ruby
+df["a"] + 5
+df["a"] - 5
+df["a"] * 5
+df["a"] / 5
+df["a"] % 5
+df["a"] ** 2
+df["a"].sqrt
+df["a"].abs
+```
+
+Rounding
+
+```ruby
+df["a"].round(2)
+df["a"].ceil
+df["a"].floor
+```
+
+Logarithm
+
+```ruby
+df["a"].log # natural log
+df["a"].log(10)
+```
+
+Exponentiation
+
+```ruby
+df["a"].exp
+```
+
+Trigonometric functions
+
+```ruby
+df["a"].sin
+df["a"].cos
+df["a"].tan
+df["a"].asin
+df["a"].acos
+df["a"].atan
+```
+
+Hyperbolic functions
+
+```ruby
+df["a"].sinh
+df["a"].cosh
+df["a"].tanh
+df["a"].asinh
+df["a"].acosh
+df["a"].atanh
+```
+
+Summary statistics
+
+```ruby
+df["a"].sum
+df["a"].mean
+df["a"].median
+df["a"].quantile(0.90)
+df["a"].min
+df["a"].max
+df["a"].std
+df["a"].var
+```
+
+## Grouping
+
+Group
+
+```ruby
+df.groupby("a").count
+```
+
+Works with all summary statistics
+
+```ruby
+df.groupby("a").max
+```
+
+Multiple groups
+
+```ruby
+df.groupby(["a", "b"]).count
+```
+
+## Combining Data Frames
+
+Add rows
+
+```ruby
+df.vstack(other_df)
+```
+
+Add columns
+
+```ruby
+df.hstack(other_df)
+```
+
+Inner join
+
+```ruby
+df.join(other_df, on: "a")
+```
+
+Left join
+
+```ruby
+df.join(other_df, on: "a", how: "left")
+```
+
+## Encoding
+
+One-hot encoding
+
+```ruby
+df.to_dummies
+```
+
+## Conversion
+
+Array of rows
+
+```ruby
+df.rows
+```
+
+Hash of series
+
+```ruby
+df.to_h
+```
+
+CSV
+
+```ruby
+df.to_csv
+# or
+df.write_csv("data.csv")
+```
+
+Parquet
+
+```ruby
+df.write_parquet("data.parquet")
+```
+
+## Types
+
+You can specify column types when creating a data frame
+
+```ruby
+Polars::DataFrame.new(data, columns: {"a" => Polars::Int32, "b" => Polars::Float32})
+```
+
+Supported types are:
+
+- boolean - `Boolean`
+- float - `Float64`, `Float32`
+- integer - `Int64`, `Int32`, `Int16`, `Int8`
+- unsigned integer - `UInt64`, `UInt32`, `UInt16`, `UInt8`
+- string - `Utf8`, `Categorical`
+- temporal - `Date`, `Datetime`, `Time`, `Duration`
+
+Get column types
+
+```ruby
+df.schema
+```
+
+For a specific column
+
+```ruby
+df["a"].dtype
+```
+
+Cast a column
+
+```ruby
+df["a"].cast(Polars::Int32)
+```
+
 ## History
 
 View the [changelog](CHANGELOG.md)

data/ext/polars/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "polars"
-version = "0.2.0"
+version = "0.2.2"
 license = "MIT"
 authors = ["Andrew Kane <andrew@ankane.org>"]
 edition = "2021"

data/ext/polars/src/lib.rs

@@ -22,7 +22,7 @@ use magnus::{
     define_module, function, memoize, method, prelude::*, Error, RArray, RClass, RHash, RModule,
     Value,
 };
-use polars::datatypes::{DataType, TimeUnit};
+use polars::datatypes::{DataType, TimeUnit, IDX_DTYPE};
 use polars::error::PolarsResult;
 use polars::frame::DataFrame;
 use polars::functions::{diag_concat_df, hor_concat_df};
@@ -71,6 +71,7 @@ fn init() -> RbResult<()> {
     module.define_singleton_method("_sum_exprs", function!(sum_exprs, 1))?;
     module.define_singleton_method("_as_struct", function!(as_struct, 1))?;
     module.define_singleton_method("_arg_where", function!(arg_where, 1))?;
+    module.define_singleton_method("_get_idx_type", function!(get_idx_type, 0))?;
 
     let class = module.define_class("RbBatchedCsv", Default::default())?;
     class.define_singleton_method("new", function!(RbBatchedCsv::new, -1))?;
@@ -988,3 +989,7 @@ fn as_struct(exprs: RArray) -> RbResult<RbExpr> {
 fn arg_where(condition: &RbExpr) -> RbExpr {
     polars::lazy::dsl::arg_where(condition.inner.clone()).into()
 }
+
+fn get_idx_type() -> Value {
+    Wrap(IDX_DTYPE).into()
+}

data/lib/polars/data_frame.rb

@@ -17,6 +17,7 @@ module Polars
     #   the orientation is inferred by matching the columns and data dimensions. If
     #   this does not yield conclusive results, column orientation is used.
     def initialize(data = nil, columns: nil, orient: nil)
+      # TODO deprecate in favor of read_sql
       if defined?(ActiveRecord) && (data.is_a?(ActiveRecord::Relation) || data.is_a?(ActiveRecord::Result))
         result = data.is_a?(ActiveRecord::Result) ? data : data.connection.select_all(data.to_sql)
         data = {}
@@ -275,6 +276,8 @@ module Polars
     def height
       _df.height
     end
+    alias_method :count, :height
+    alias_method :length, :height
 
     # Get the width of the DataFrame.
     #
@@ -521,13 +524,13 @@ module Polars
             return df.slice(row_selection, 1)
           end
           # df[2, "a"]
-          if col_selection.is_a?(String)
+          if col_selection.is_a?(String) || col_selection.is_a?(Symbol)
             return self[col_selection][row_selection]
           end
         end
 
         # column selection can be "a" and ["a", "b"]
-        if col_selection.is_a?(String)
+        if col_selection.is_a?(String) || col_selection.is_a?(Symbol)
           col_selection = [col_selection]
         end
 
@@ -539,7 +542,7 @@ module Polars
 
         if col_selection.is_a?(Array)
           # df[.., [1, 2]]
-          if is_int_sequence(col_selection)
+          if Utils.is_int_sequence(col_selection)
             series_list = col_selection.map { |i| to_series(i) }
             df = self.class.new(series_list)
             return df[row_selection]
@@ -553,8 +556,8 @@ module Polars
 
         # select single column
         # df["foo"]
-        if item.is_a?(String)
-          return Utils.wrap_s(_df.column(item))
+        if item.is_a?(String) || item.is_a?(Symbol)
+          return Utils.wrap_s(_df.column(item.to_s))
         end
 
         # df[idx]
@@ -572,6 +575,28 @@ module Polars
           # df[["foo", "bar"]]
           return _from_rbdf(_df.select(item))
         end
+
+        if Utils.is_int_sequence(item)
+          item = Series.new("", item)
+        end
+
+        if item.is_a?(Series)
+          dtype = item.dtype
+          if dtype == Utf8
+            return _from_rbdf(_df.select(item))
+          elsif dtype == UInt32
+            return _from_rbdf(_df.take_with_series(item._s))
+          elsif [UInt8, UInt16, UInt64, Int8, Int16, Int32, Int64].include?(dtype)
+            return _from_rbdf(
+              _df.take_with_series(_pos_idxs(item, 0)._s)
+            )
+          end
+        end
+      end
+
+      # Ruby-specific
+      if item.is_a?(Expr)
+        return filter(item)
       end
 
       raise ArgumentError, "Cannot get item of type: #{item.class.name}"
@@ -797,6 +822,13 @@ module Polars
       nil
     end
 
+    # Write to comma-separated values (CSV) string.
+    #
+    # @return [String]
+    def to_csv(**options)
+      write_csv(**options)
+    end
+
     # Write to Apache Avro file.
     #
     # @param file [String]
@@ -4648,8 +4680,53 @@ module Polars
       end
     end
 
-    # def _pos_idxs
-    # end
+    def _pos_idxs(idxs, dim)
+      idx_type = Polars._get_idx_type
+
+      if idxs.is_a?(Series)
+        if idxs.dtype == idx_type
+          return idxs
+        end
+        if [UInt8, UInt16, idx_type == UInt32 ? UInt64 : UInt32, Int8, Int16, Int32, Int64].include?(idxs.dtype)
+          if idx_type == UInt32
+            if [Int64, UInt64].include?(idxs.dtype)
+              if idxs.max >= 2**32
+                raise ArgumentError, "Index positions should be smaller than 2^32."
+              end
+            end
+            if idxs.dtype == Int64
+              if idxs.min < -(2**32)
+                raise ArgumentError, "Index positions should be bigger than -2^32 + 1."
+              end
+            end
+          end
+          if [Int8, Int16, Int32, Int64].include?(idxs.dtype)
+            if idxs.min < 0
+              if idx_type == UInt32
+                if [Int8, Int16].include?(idxs.dtype)
+                  idxs = idxs.cast(Int32)
+                end
+              else
+                if [Int8, Int16, Int32].include?(idxs.dtype)
+                  idxs = idxs.cast(Int64)
+                end
+              end
+
+              idxs =
+                Polars.select(
+                  Polars.when(Polars.lit(idxs) < 0)
+                    .then(shape[dim] + Polars.lit(idxs))
+                    .otherwise(Polars.lit(idxs))
+                ).to_series
+            end
+          end
+
+          return idxs.cast(idx_type)
+        end
+      end
+
+      raise ArgumentError, "Unsupported idxs datatype."
+    end
 
     # @private
     def self.hash_to_rbdf(data, columns: nil)

data/lib/polars/data_types.rb

@@ -93,7 +93,7 @@ module Polars
   class Time < DataType
   end
 
-  # Type for wrapping arbitrary Python objects.
+  # Type for wrapping arbitrary Ruby objects.
   class Object < DataType
   end
 

data/lib/polars/io.rb

@@ -590,8 +590,33 @@ module Polars
       DataFrame._read_ndjson(file)
     end
 
-    # def read_sql
-    # end
+    # Read a SQL query into a DataFrame.
+    #
+    # @param sql [Object]
+    #   ActiveRecord::Relation or ActiveRecord::Result.
+    #
+    # @return [DataFrame]
+    def read_sql(sql)
+      if !defined?(ActiveRecord)
+        raise Error, "Active Record not available"
+      end
+
+      result =
+        if sql.is_a?(ActiveRecord::Result)
+          sql
+        elsif sql.is_a?(ActiveRecord::Relation)
+          sql.connection.select_all(sql.to_sql)
+        elsif sql.is_a?(String)
+          ActiveRecord::Base.connection.select_all(sql)
+        else
+          raise ArgumentError, "Expected ActiveRecord::Relation, ActiveRecord::Result, or String"
+        end
+      data = {}
+      result.columns.each_with_index do |k, i|
+        data[k] = result.rows.map { |r| r[i] }
+      end
+      DataFrame.new(data)
+    end
 
     # def read_excel
     # end

data/lib/polars/series.rb

@@ -263,6 +263,10 @@ module Polars
     #
     # @return [Object]
     def [](item)
+      if item.is_a?(Series) && [UInt8, UInt16, UInt32, UInt64, Int8, Int16, Int32, Int64].include?(item.dtype)
+        return Utils.wrap_s(_s.take_with_series(_pos_idxs(item)._s))
+      end
+
       if item.is_a?(Integer)
         return _s.get_idx(item)
       end
@@ -271,6 +275,10 @@ module Polars
         return Slice.new(self).apply(item)
       end
 
+      if Utils.is_int_sequence(item)
+        return Utils.wrap_s(_s.take_with_series(_pos_idxs(Series.new("", item))._s))
+      end
+
       raise ArgumentError, "Cannot get item of type: #{item.class.name}"
     end
 
@@ -287,24 +295,23 @@ module Polars
       end
 
       if key.is_a?(Series)
-        if key.dtype == :bool
+        if key.dtype == Boolean
           self._s = set(key, value)._s
-        elsif key.dtype == :u64
-          self._s = set_at_idx(key.cast(:u32), value)._s
-        elsif key.dtype == :u32
+        elsif key.dtype == UInt64
+          self._s = set_at_idx(key.cast(UInt32), value)._s
+        elsif key.dtype == UInt32
           self._s = set_at_idx(key, value)._s
         else
           raise Todo
         end
-      end
-
-      if key.is_a?(Array)
-        s = Utils.wrap_s(sequence_to_rbseries("", key, dtype: :u32))
+      elsif key.is_a?(Array)
+        s = Utils.wrap_s(sequence_to_rbseries("", key, dtype: UInt32))
+        self[s] = value
+      elsif key.is_a?(Range)
+        s = Series.new("", key, dtype: UInt32)
         self[s] = value
       elsif key.is_a?(Integer)
-        # TODO fix
-        # self[[key]] = value
-        set_at_idx(key, value)
+        self[[key]] = value
       else
         raise ArgumentError, "cannot use #{key} for indexing"
       end
@@ -1647,6 +1654,7 @@ module Polars
     def len
       _s.len
     end
+    alias_method :count, :len
     alias_method :length, :len
 
     # Cast between data types.
@@ -2183,6 +2191,7 @@ module Polars
     def arcsin
       super
     end
+    alias_method :asin, :arcsin
 
     # Compute the element-wise value for the inverse cosine.
     #
@@ -2202,6 +2211,7 @@ module Polars
     def arccos
       super
     end
+    alias_method :acos, :arccos
 
     # Compute the element-wise value for the inverse tangent.
     #
@@ -2221,6 +2231,7 @@ module Polars
     def arctan
       super
     end
+    alias_method :atan, :arctan
 
     # Compute the element-wise value for the inverse hyperbolic sine.
     #
@@ -2240,6 +2251,7 @@ module Polars
     def arcsinh
       super
     end
+    alias_method :asinh, :arcsinh
 
     # Compute the element-wise value for the inverse hyperbolic cosine.
     #
@@ -2260,6 +2272,7 @@ module Polars
     def arccosh
       super
     end
+    alias_method :acosh, :arccosh
 
     # Compute the element-wise value for the inverse hyperbolic tangent.
     #
@@ -2283,6 +2296,7 @@ module Polars
     def arctanh
       super
     end
+    alias_method :atanh, :arctanh
 
     # Compute the element-wise value for the hyperbolic sine.
     #
@@ -3520,6 +3534,59 @@ module Polars
       end
     end
 
+    def _pos_idxs(idxs)
+      idx_type = Polars._get_idx_type
+
+      if idxs.is_a?(Series)
+        if idxs.dtype == idx_type
+          return idxs
+        end
+        if [UInt8, UInt16, idx_type == UInt32 ? UInt64 : UInt32, Int8, Int16, Int32, Int64].include?(idxs.dtype)
+          if idx_type == UInt32
+            if [Int64, UInt64].include?(idxs.dtype)
+              if idxs.max >= 2**32
+                raise ArgumentError, "Index positions should be smaller than 2^32."
+              end
+            end
+            if idxs.dtype == Int64
+              if idxs.min < -(2**32)
+                raise ArgumentError, "Index positions should be bigger than -2^32 + 1."
+              end
+            end
+          end
+          if [Int8, Int16, Int32, Int64].include?(idxs.dtype)
+            if idxs.min < 0
+              if idx_type == UInt32
+                if [Int8, Int16].include?(idxs.dtype)
+                  idxs = idxs.cast(Int32)
+                end
+              else
+                if [Int8, Int16, Int32].include?(idxs.dtype)
+                  idxs = idxs.cast(Int64)
+                end
+              end
+
+              # Update negative indexes to absolute indexes.
+              return (
+                idxs.to_frame
+                .select(
+                  Polars.when(Polars.col(idxs.name) < 0)
+                    .then(len + Polars.col(idxs.name))
+                    .otherwise(Polars.col(idxs.name))
+                    .cast(idx_type)
+                )
+                .to_series(0)
+              )
+            end
+          end
+
+          return idxs.cast(idx_type)
+        end
+      end
+
+      raise ArgumentError, "Unsupported idxs datatype."
+    end
+
     def _comp(other, op)
       if other.is_a?(Series)
         return Utils.wrap_s(_s.send(op, other._s))

data/lib/polars/slice.rb

@@ -56,7 +56,7 @@ module Polars
     # Normalize slice bounds, identify unbounded and/or zero-length slices.
     def _slice_setup(s)
       # can normalize slice indices as we know object size
-      obj_len = @obj.len
+      obj_len = @obj.length
       start = if s.begin
         if s.begin < 0
           [s.begin + obj_len, 0].max

data/lib/polars/utils.rb

@@ -70,7 +70,7 @@ module Polars
     end
 
     def self.selection_to_rbexpr_list(exprs)
-      if exprs.is_a?(String) || exprs.is_a?(Expr) || exprs.is_a?(Series)
+      if exprs.is_a?(String) || exprs.is_a?(Symbol) || exprs.is_a?(Expr) || exprs.is_a?(Series)
         exprs = [exprs]
       end
 
@@ -78,9 +78,9 @@ module Polars
     end
 
     def self.expr_to_lit_or_expr(expr, str_to_lit: true)
-      if expr.is_a?(String) && !str_to_lit
+      if (expr.is_a?(String) || expr.is_a?(Symbol)) && !str_to_lit
         col(expr)
-      elsif expr.is_a?(Integer) || expr.is_a?(Float) || expr.is_a?(String) || expr.is_a?(Series) || expr.nil?
+      elsif expr.is_a?(Integer) || expr.is_a?(Float) || expr.is_a?(String) || expr.is_a?(Symbol) || expr.is_a?(Series) || expr.nil?
         lit(expr)
       elsif expr.is_a?(Expr)
         expr
@@ -181,6 +181,26 @@ module Polars
       val.all? { |x| x.is_a?(eltype) }
     end
 
+    def self.is_bool_sequence(val)
+      val.is_a?(Array) && val.all? { |x| x == true || x == false }
+    end
+
+    def self.is_dtype_sequence(val)
+      val.is_a?(Array) && val.all? { |x| is_polars_dtype(x) }
+    end
+
+    def self.is_int_sequence(val)
+      val.is_a?(Array) && _is_iterable_of(val, Integer)
+    end
+
+    def self.is_expr_sequence(val)
+      val.is_a?(Array) && _is_iterable_of(val, Expr)
+    end
+
+    def self.is_rbexpr_sequence(val)
+      val.is_a?(Array) && _is_iterable_of(val, RbExpr)
+    end
+
     def self.is_str_sequence(val, allow_str: false)
       if allow_str == false && val.is_a?(String)
         false

data/lib/polars/version.rb

@@ -1,4 +1,4 @@
 module Polars
   # @private
-  VERSION = "0.2.0"
+  VERSION = "0.2.2"
 end

data/lib/polars.rb

@@ -7,6 +7,7 @@ end
 
 # stdlib
 require "date"
+require "stringio"
 
 # modules
 require_relative "polars/expr_dispatch"

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: polars-df
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.2.2
 platform: ruby
 authors:
 - Andrew Kane
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-01-15 00:00:00.000000000 Z
+date: 2023-01-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys