polars-df 0.2.4 → 0.2.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 5afa804963fc59154e8adde6034e07d1a2a90c077b8e0b72d2ad3ca49af34e8e
-  data.tar.gz: b2291436080973ad43595ef25559cdc81669b07986a6779c85c4c372e220e39a
+  metadata.gz: 122f88f6a702c0252a98404b8ecccddcd0195cadd79b7062f451cb23f5d3b8e3
+  data.tar.gz: 67588b6a0aa9829af179c8e3d5339b125b50dd6941e02cea4fc29aedd2e29217
 SHA512:
-  metadata.gz: 7d7aa73d12e9a322de83db5f69b4ab0f5e9605b737776a83476f917cf32ad1ef6770c79e80a810be68a9144206fef6c3931af3285763741232b8ba87871d0a06
-  data.tar.gz: 2f9daceaba3edd0671a650ab1a2da6ddc6591f768e52a4c8a8f01d7e3fd573320c4464d0081762daeb75d41475bb4e87e298b426b507ef6c7bb9d3164149e7e6
+  metadata.gz: 6393dcab24e4001c7d1b9149ef441317d8ba76571850071ed829b569e055c9ad4d4ffcbdb256ede8225922b56378984ebb113570051d27b2761a50f1b23fa168
+  data.tar.gz: 4356270d7f15d964eedd84bf37c05054323a19128fa5840563829af88f46107e1277534ae976ea078808807b7c80433196eb351f7160118d1251d1dd4ddbbc4b

data/CHANGELOG.md

@@ -1,3 +1,8 @@
+## 0.2.5 (2023-02-01)
+
+- Added support for glob patterns to `read_csv` method
+- Added support for symbols to more methods
+
 ## 0.2.4 (2023-01-29)
 
 - Added support for more types when creating a data frame from an array of hashes

data/README.md

@@ -73,9 +73,9 @@ From an array of hashes
 
 ```ruby
 Polars::DataFrame.new([
-  {"a" => 1, "b" => "one"},
-  {"a" => 2, "b" => "two"},
-  {"a" => 3, "b" => "three"}
+  {a: 1, b: "one"},
+  {a: 2, b: "two"},
+  {a: 3, b: "three"}
 ])
 ```
 

data/lib/polars/batched_csv_reader.rb

@@ -29,7 +29,7 @@ module Polars
       eol_char: "\n",
       new_columns: nil
     )
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         path = Utils.format_path(file)
       end
 

data/lib/polars/data_frame.rb

@@ -94,7 +94,7 @@ module Polars
       sample_size: 1024,
       eol_char: "\n"
     )
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         path = Utils.format_path(file)
       else
         path = nil
@@ -124,7 +124,39 @@ module Polars
         columns = [columns]
       end
       if file.is_a?(String) && file.include?("*")
-        raise Todo
+        dtypes_dict = nil
+        if !dtype_list.nil?
+          dtypes_dict = dtype_list.to_h
+        end
+        if !dtype_slice.nil?
+          raise ArgumentError, "cannot use glob patterns and unnamed dtypes as `dtypes` argument; Use dtypes: Mapping[str, Type[DataType]"
+        end
+        scan = Polars.scan_csv(
+          file,
+          has_header: has_header,
+          sep: sep,
+          comment_char: comment_char,
+          quote_char: quote_char,
+          skip_rows: skip_rows,
+          dtypes: dtypes_dict,
+          null_values: null_values,
+          ignore_errors: ignore_errors,
+          infer_schema_length: infer_schema_length,
+          n_rows: n_rows,
+          low_memory: low_memory,
+          rechunk: rechunk,
+          skip_rows_after_header: skip_rows_after_header,
+          row_count_name: row_count_name,
+          row_count_offset: row_count_offset,
+          eol_char: eol_char
+        )
+        if columns.nil?
+          return _from_rbdf(scan.collect._df)
+        elsif is_str_sequence(columns, allow_str: false)
+          return _from_rbdf(scan.select(columns).collect._df)
+        else
+          raise ArgumentError, "cannot use glob patterns and integer based projection as `columns` argument; Use columns: List[str]"
+        end
       end
 
       projection, columns = Utils.handle_projection_columns(columns)
@@ -170,7 +202,7 @@ module Polars
       row_count_offset: 0,
       low_memory: false
     )
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -194,7 +226,7 @@ module Polars
 
     # @private
     def self._read_avro(file, columns: nil, n_rows: nil)
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
       projection, columns = Utils.handle_projection_columns(columns)
@@ -211,7 +243,7 @@ module Polars
       rechunk: true,
       memory_map: true
     )
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
       if columns.is_a?(String)
@@ -237,7 +269,7 @@ module Polars
 
     # @private
     def self._read_json(file)
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -246,7 +278,7 @@ module Polars
 
     # @private
     def self._read_ndjson(file)
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -698,7 +730,7 @@ module Polars
       pretty: false,
       row_oriented: false
     )
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -713,7 +745,7 @@ module Polars
     #
     # @return [nil]
     def write_ndjson(file)
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -803,7 +835,7 @@ module Polars
         return buffer.string.force_encoding(Encoding::UTF_8)
       end
 
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -841,7 +873,7 @@ module Polars
       if compression.nil?
         compression = "uncompressed"
       end
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -860,7 +892,7 @@ module Polars
       if compression.nil?
         compression = "uncompressed"
       end
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -902,7 +934,7 @@ module Polars
       if compression.nil?
         compression = "uncompressed"
       end
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 

data/lib/polars/io.rb

@@ -268,7 +268,7 @@ module Polars
       _check_arg_is_1byte("comment_char", comment_char, false)
       _check_arg_is_1byte("quote_char", quote_char, true)
 
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -384,7 +384,7 @@ module Polars
       storage_options: nil,
       low_memory: false
     )
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -435,7 +435,7 @@ module Polars
       row_count_name: nil,
       row_count_offset: 0
     )
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -463,7 +463,7 @@ module Polars
     #
     # @return [DataFrame]
     def read_avro(file, columns: nil, n_rows: nil)
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -786,7 +786,7 @@ module Polars
     #
     # @return [Hash]
     def read_ipc_schema(file)
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -800,7 +800,7 @@ module Polars
     #
     # @return [Hash]
     def read_parquet_schema(file)
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 

data/lib/polars/lazy_frame.rb

@@ -106,7 +106,7 @@ module Polars
       storage_options: nil,
       memory_map: true
     )
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -156,7 +156,7 @@ module Polars
     #
     # @return [LazyFrame]
     def self.read_json(file)
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
 
@@ -263,7 +263,7 @@ module Polars
     #
     # @return [nil]
     def write_json(file)
-      if file.is_a?(String) || (defined?(Pathname) && file.is_a?(Pathname))
+      if Utils.pathlike?(file)
         file = Utils.format_path(file)
       end
       _ldf.write_json(file)

data/lib/polars/lazy_functions.rb

@@ -15,7 +15,7 @@ module Polars
       if name.is_a?(DataType)
         Utils.wrap_expr(_dtype_cols([name]))
       elsif name.is_a?(Array)
-        if name.length == 0 || name[0].is_a?(String) || name[0].is_a?(Symbol)
+        if name.length == 0 || Utils.strlike?(name[0])
           name = name.map { |v| v.is_a?(Symbol) ? v.to_s : v }
           Utils.wrap_expr(RbExpr.cols(name))
         elsif Utils.is_polars_dtype(name[0])
@@ -119,7 +119,7 @@ module Polars
     def max(column)
       if column.is_a?(Series)
         column.max
-      elsif column.is_a?(String) || column.is_a?(Symbol)
+      elsif Utils.strlike?(column)
         col(column).max
       else
         exprs = Utils.selection_to_rbexpr_list(column)
@@ -141,7 +141,7 @@ module Polars
     def min(column)
       if column.is_a?(Series)
         column.min
-      elsif column.is_a?(String) || column.is_a?(Symbol)
+      elsif Utils.strlike?(column)
         col(column).min
       else
         exprs = Utils.selection_to_rbexpr_list(column)
@@ -156,7 +156,7 @@ module Polars
     def sum(column)
       if column.is_a?(Series)
         column.sum
-      elsif column.is_a?(String) || column.is_a?(Symbol)
+      elsif Utils.strlike?(column)
         col(column.to_s).sum
       elsif column.is_a?(Array)
         exprs = Utils.selection_to_rbexpr_list(column)
@@ -356,7 +356,7 @@ module Polars
     def cumsum(column)
       if column.is_a?(Series)
         column.cumsum
-      elsif column.is_a?(String)
+      elsif Utils.strlike?(column)
         col(column).cumsum
       else
         cumfold(lit(0).cast(:u32), ->(a, b) { a + b }, column).alias("cumsum")
@@ -380,10 +380,10 @@ module Polars
     #
     # @return [Expr]
     def spearman_rank_corr(a, b, ddof: 1, propagate_nans: false)
-      if a.is_a?(String)
+      if Utils.strlike?(a)
         a = col(a)
       end
-      if b.is_a?(String)
+      if Utils.strlike?(b)
         b = col(b)
       end
       Utils.wrap_expr(RbExpr.spearman_rank_corr(a._rbexpr, b._rbexpr, ddof, propagate_nans))
@@ -400,10 +400,10 @@ module Polars
     #
     # @return [Expr]
     def pearson_corr(a, b, ddof: 1)
-      if a.is_a?(String)
+      if Utils.strlike?(a)
         a = col(a)
       end
-      if b.is_a?(String)
+      if Utils.strlike?(b)
         b = col(b)
       end
       Utils.wrap_expr(RbExpr.pearson_corr(a._rbexpr, b._rbexpr, ddof))
@@ -418,10 +418,10 @@ module Polars
     #
     # @return [Expr]
     def cov(a, b)
-      if a.is_a?(String)
+      if Utils.strlike?(a)
         a = col(a)
       end
-      if b.is_a?(String)
+      if Utils.strlike?(b)
         b = col(b)
       end
       Utils.wrap_expr(RbExpr.cov(a._rbexpr, b._rbexpr))
@@ -486,7 +486,7 @@ module Polars
     #
     # @return [Expr]
     def any(name)
-      if name.is_a?(String)
+      if Utils.strlike?(name)
         col(name).any
       else
         fold(lit(false), ->(a, b) { a.cast(:bool) | b.cast(:bool) }, name).alias("any")
@@ -589,7 +589,7 @@ module Polars
     def all(name = nil)
       if name.nil?
         col("*")
-      elsif name.is_a?(String) || name.is_a?(Symbol)
+      elsif Utils.strlike?(name)
         col(name).all
       else
         raise Todo
@@ -1137,7 +1137,7 @@ module Polars
     #   # │ 2022-10-25 07:31:39 │
     #   # └─────────────────────┘
     def from_epoch(column, unit: "s", eager: false)
-      if column.is_a?(String)
+      if Utils.strlike?(column)
         column = col(column)
       elsif !column.is_a?(Series) && !column.is_a?(Expr)
         column = Series.new(column)

data/lib/polars/utils.rb

@@ -177,6 +177,14 @@ module Polars
       value == true || value == false
     end
 
+    def self.strlike?(value)
+      value.is_a?(String) || value.is_a?(Symbol)
+    end
+
+    def self.pathlike?(value)
+      value.is_a?(String) || (defined?(Pathname) && value.is_a?(Pathname))
+    end
+
     def self._is_iterable_of(val, eltype)
       val.all? { |x| x.is_a?(eltype) }
     end

data/lib/polars/version.rb

@@ -1,4 +1,4 @@
 module Polars
   # @private
-  VERSION = "0.2.4"
+  VERSION = "0.2.5"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: polars-df
 version: !ruby/object:Gem::Version
-  version: 0.2.4
+  version: 0.2.5
 platform: ruby
 authors:
 - Andrew Kane
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-01-30 00:00:00.000000000 Z
+date: 2023-02-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys