polars-df 0.2.1 → 0.2.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4a06a5c0bdca450318f76aaf5e810795e663d07ccf58d94e903307abef87a964
-  data.tar.gz: 2589b519a564b46acc2bffb8c15c8f0abadd24ef374c1e2746b6711faf3e0c48
+  metadata.gz: df03134e7edf09e86b5a4f4f9ae9a926bac4c9c0804a29c3422c32675f478825
+  data.tar.gz: e0338be1aa96d0ad082ebf8fe27e608b2906b243dd49fa837aceb7f8186947d8
 SHA512:
-  metadata.gz: bdf428c391a3d31b98021a080194b7e4dfb34266a0e7634d3676e903e732f3328123b3711c9b36a037a97b1ed18df2b129cf2d980703ec2d39cc6dfae2279eac
-  data.tar.gz: '0749f2c7fe1f5fb7b10954ae16db47b15e54a1d042c91aae8dcdf3515f5db00ed68a1e82fa9a46546af2f4051356dac1dfb886dc7ac0852e58e92b68d09983ef'
+  metadata.gz: 75a139d30f9fdebaa84a21fa45cec8a199da76eb295e7099ceb849646a93fbc7ed80ffed18aaa8eb7bbfc53a32792b2e47101485ad31d727a47ed67d8d7e8110
+  data.tar.gz: 589f7fbc1300aadc05568308700f6a94b934e63c40bd1be0a3e7b6f564c0d55f256e2e45e926c128d80453d0e7d200b057f640b02cd6fb9aaddf5bf55dd89754

data/CHANGELOG.md

@@ -1,3 +1,9 @@
+## 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

data/Cargo.lock

@@ -1367,7 +1367,7 @@ dependencies = [
 
 [[package]]
 name = "polars"
-version = "0.2.1"
+version = "0.2.2"
 dependencies = [
  "ahash",
  "jemallocator",

data/README.md

@@ -41,6 +41,9 @@ From a CSV
 
 ```ruby
 Polars.read_csv("file.csv")
+
+# or lazily with
+Polars.scan_csv("file.csv")
 ```
 
 From Parquet
@@ -135,9 +138,9 @@ df[Polars.col("a") <= 2]
 And, or, and exclusive or
 
 ```ruby
-df[(Polars.col("a") > 100) & (Polars.col("b") == "one")] # and
-df[(Polars.col("a") > 100) | (Polars.col("b") == "one")] # or
-df[(Polars.col("a") > 100) ^ (Polars.col("b") == "one")] # xor
+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

data/ext/polars/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "polars"
-version = "0.2.1"
+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

@@ -277,6 +277,7 @@ module Polars
       _df.height
     end
     alias_method :count, :height
+    alias_method :length, :height
 
     # Get the width of the DataFrame.
     #
@@ -541,7 +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]
@@ -574,6 +575,23 @@ 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
@@ -4662,8 +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/io.rb

@@ -606,8 +606,10 @@ module Polars
           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 or ActiveRecord::Result"
+          raise ArgumentError, "Expected ActiveRecord::Relation, ActiveRecord::Result, or String"
         end
       data = {}
       result.columns.each_with_index do |k, i|

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
@@ -3527,6 +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

@@ -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 = "0.2.1"
+  VERSION = "0.2.2"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: polars-df
 version: !ruby/object:Gem::Version
-  version: 0.2.1
+  version: 0.2.2
 platform: ruby
 authors:
 - Andrew Kane
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-01-18 00:00:00.000000000 Z
+date: 2023-01-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys