daru_lite 0.1.1 → 0.1.2

data/spec/data_frame/filterable_example.rb

@@ -0,0 +1,250 @@
+shared_examples_for 'a filterable DataFrame' do
+  describe '#uniq' do
+    let(:df) { DaruLite::DataFrame.from_csv 'spec/fixtures/duplicates.csv' }
+
+    context 'with no args' do
+      subject { df.uniq }
+
+      it 'returns the correct result' do
+        expect(subject.shape.first).to eq 30
+      end
+    end
+
+    context 'given a vector' do
+      subject { df.uniq('color') }
+
+      it 'returns the correct result' do
+        expect(subject.shape.first).to eq 2
+      end
+    end
+
+    context 'given an array of vectors' do
+      subject { df.uniq("color", "director_name") }
+
+      it 'returns the correct result' do
+        expect(subject.shape.first).to eq 29
+      end
+    end
+  end
+
+  describe "#filter" do
+    let(:df) { DaruLite::DataFrame.new({ a: [1,2,3], b: [2,3,4] }) }
+
+    context 'avis is row' do
+      subject { df.filter(:row) { |r| r[:a] % 2 == 0 } }
+
+      it { is_expected.to eq(df.filter_rows { |r| r[:a] % 2 == 0 }) }
+    end
+
+    context 'avis is vector' do
+      subject { df.filter(:vector) { |v| v[0] == 1 } }
+
+      it { is_expected.to eq(df.filter_vectors { |v| v[0] == 1 }) }
+    end
+
+    context 'avis is unknown' do
+      subject { df.filter(:kitten) {} }
+
+      it { expect { subject }.to raise_error ArgumentError, /axis/ }
+    end
+  end
+
+
+  describe '#reject_values' do
+    let(:df) do
+      DaruLite::DataFrame.new({
+        a: [1,    2,          3,   nil,        Float::NAN, nil, 1,   7],
+        b: [:a,  :b,          nil, Float::NAN, nil,        3,   5,   8],
+        c: ['a',  Float::NAN, 3,   4,          3,          5,   nil, 7]
+      }, index: 11..18)
+    end
+    before { df.to_category :b }
+
+    context 'remove nils only' do
+      subject { df.reject_values nil }
+
+      it { is_expected.to be_a DaruLite::DataFrame }
+      its(:'b.type') { is_expected.to eq :category }
+      its(:'a.to_a') { is_expected.to eq [1, 2, 7] }
+      its(:'b.to_a') { is_expected.to eq [:a, :b, 8] }
+      its(:'c.to_a') { is_expected.to eq ['a', Float::NAN, 7] }
+      its(:'index.to_a') { is_expected.to eq [11, 12, 18] }
+    end
+
+    context 'remove Float::NAN only' do
+      subject { df.reject_values Float::NAN }
+
+      it { is_expected.to be_a DaruLite::DataFrame }
+      its(:'b.type') { is_expected.to eq :category }
+      its(:'a.to_a') { is_expected.to eq [1, 3, nil, 1, 7] }
+      its(:'b.to_a') { is_expected.to eq [:a, nil, 3, 5, 8] }
+      its(:'c.to_a') { is_expected.to eq ['a', 3, 5, nil, 7] }
+      its(:'index.to_a') { is_expected.to eq [11, 13, 16, 17, 18] }
+    end
+
+    context 'remove both nil and Float::NAN' do
+      subject { df.reject_values nil, Float::NAN }
+
+      it { is_expected.to be_a DaruLite::DataFrame }
+      its(:'b.type') { is_expected.to eq :category }
+      its(:'a.to_a') { is_expected.to eq [1, 7] }
+      its(:'b.to_a') { is_expected.to eq [:a, 8] }
+      its(:'c.to_a') { is_expected.to eq ['a', 7] }
+      its(:'index.to_a') { is_expected.to eq [11, 18] }
+    end
+
+    context 'any other values' do
+      subject { df.reject_values 1, 5 }
+
+      it { is_expected.to be_a DaruLite::DataFrame }
+      its(:'b.type') { is_expected.to eq :category }
+      its(:'a.to_a') { is_expected.to eq [2, 3, nil, Float::NAN, 7] }
+      its(:'b.to_a') { is_expected.to eq [:b, nil, Float::NAN, nil, 8] }
+      its(:'c.to_a') { is_expected.to eq [Float::NAN, 3, 4, 3, 7] }
+      its(:'index.to_a') { is_expected.to eq [12, 13, 14, 15, 18] }
+    end
+
+    context 'when resultant dataframe has one row' do
+      subject { df.reject_values 1, 2, 3, 4, 5, nil, Float::NAN }
+
+      it { is_expected.to be_a DaruLite::DataFrame }
+      its(:'b.type') { is_expected.to eq :category }
+      its(:'a.to_a') { is_expected.to eq [7] }
+      its(:'b.to_a') { is_expected.to eq [8] }
+      its(:'c.to_a') { is_expected.to eq [7] }
+      its(:'index.to_a') { is_expected.to eq [18] }
+    end
+
+    context 'when resultant dataframe is empty' do
+      subject { df.reject_values 1, 2, 3, 4, 5, 6, 7, nil, Float::NAN }
+
+      it { is_expected.to be_a DaruLite::DataFrame }
+      its(:'b.type') { is_expected.to eq :category }
+      its(:'a.to_a') { is_expected.to eq [] }
+      its(:'b.to_a') { is_expected.to eq [] }
+      its(:'c.to_a') { is_expected.to eq [] }
+      its(:'index.to_a') { is_expected.to eq [] }
+    end
+  end
+
+  describe "#keep_row_if" do
+    subject { df.keep_row_if { |row| row[:a] % 10 == 0 } }
+
+    let(:index) { [:one, :two, :three, :four, :five] }
+    let(:order) { [:a, :b, :c] }
+    let(:df)  do
+      DaruLite::DataFrame.new({
+          b: [10, 12, 20, 23, 30],
+          a: [50, 30, 30, 1, 5],
+          c: [10, 20, 30, 40, 50]
+        },
+        order:,
+        index:
+      )
+    end
+
+    context DaruLite::Index do
+      it "keeps row if block evaluates to true" do
+        subject
+        expect(df).to eq(
+          DaruLite::DataFrame.new(
+            { b: [10, 12, 20], a: [50, 30, 30], c: [10, 20, 30] },
+            order:,
+            index: index[..2]
+          )
+        )
+      end
+    end
+
+    context DaruLite::CategoricalIndex do
+      let (:index) { DaruLite::CategoricalIndex.new([:a, 1, 1, :a, :c]) }
+
+      it "keeps row if block evaluates to true" do
+        subject
+        expect(df).to eq(
+          DaruLite::DataFrame.new(
+            { b: [10, 12, 20], a: [50, 30, 30], c: [10, 20, 30] },
+            order:,
+            index: DaruLite::CategoricalIndex.new([:a, 1, 1])
+          )
+        )
+      end
+    end
+  end
+
+  describe "#keep_vector_if" do
+    it "keeps vector if block evaluates to true" do
+      df.keep_vector_if do |vector|
+        vector == [1,2,3,4,5].dv(nil, [:one, :two, :three, :four, :five])
+      end
+
+      expect(df).to eq(DaruLite::DataFrame.new({a: [1,2,3,4,5]}, order: [:a],
+        index: [:one, :two, :three, :four, :five]))
+    end
+  end
+
+  describe "#filter_vectors" do
+    context DaruLite::Index do
+      subject { df.filter_vectors { |vector| vector[0] == 1 } }
+
+      let(:df) { DaruLite::DataFrame.new({ a: [1,2,3], b: [2,3,4] }) }
+
+      it "filters vectors" do
+        expect(subject).to eq(DaruLite::DataFrame.new({a: [1,2,3]}))
+      end
+    end
+  end
+
+  describe "#filter_rows" do
+    context DaruLite::Index do
+      subject { df.filter_rows { |r| r[:a] != 2 } }
+
+      let(:df) { DaruLite::DataFrame.new a: 1..3, b: 4..6 }
+
+      it "preserves names of vectors" do
+        expect(subject[:a].name).to eq(df[:a].name)
+      end
+
+      context "when specified no index" do
+        subject { df.filter_rows { |row| row[:a] % 2 == 0 } }
+
+        let(:df) { DaruLite::DataFrame.new({ a: [1,2,3], b: [2,3,4] }) }
+
+        it "filters rows" do
+          expect(subject).to eq(DaruLite::DataFrame.new({ a: [2], b: [3] }, order: [:a, :b], index: [1]))
+        end
+      end
+
+      context "when specified numerical index" do
+        subject { df.filter_rows { |row| row[:a] % 2 == 0 } }
+
+        let(:df) { DaruLite::DataFrame.new({ a: [1,2,3], b: [2,3,4] }, index: [1,2,3]) }
+
+        it "filters rows" do
+          expect(subject).to eq(DaruLite::DataFrame.new({ a: [2], b: [3] }, order: [:a, :b], index: [2]))
+        end
+      end
+    end
+  end
+
+  context "#filter_vector" do
+    subject { df.filter_vector(:id) { |c| c[:id] == 2 or c[:id] == 4 } }
+
+    let(:df) do
+      DaruLite::DataFrame.new(
+        {
+          id: DaruLite::Vector.new([1, 2, 3, 4, 5]),
+          name: DaruLite::Vector.new(%w(Alex Claude Peter Franz George)),
+          age: DaruLite::Vector.new([20, 23, 25, 27, 5]),
+          city: DaruLite::Vector.new(['New York', 'London', 'London', 'Paris', 'Tome']),
+          a1: DaruLite::Vector.new(['a,b', 'b,c', 'a', nil, 'a,b,c'])
+        },
+        order: [:id, :name, :age, :city, :a1]
+      )
+    end
+
+    it "creates new vector with the data of a given field for which block returns true" do
+      expect(subject).to eq(DaruLite::Vector.new([2,4]))
+    end
+  end
+end

data/spec/data_frame/indexable_example.rb

@@ -0,0 +1,221 @@
+shared_examples_for 'an indexable DataFrame' do
+  describe "#set_index" do
+    let(:df) do
+      DaruLite::DataFrame.new(
+        {
+          a: [1,2,3,4,5],
+          b: ['a','b','c','d','e'],
+          c: [11,22,33,44,55]
+        }
+      )
+    end
+
+    it "sets a particular column as the index and deletes that column" do
+      df.set_index(:b)
+      expect(df).to eq(
+        DaruLite::DataFrame.new({
+          a: [1,2,3,4,5],
+          c: [11,22,33,44,55]
+          }, index: ['a','b','c','d','e'])
+        )
+    end
+
+    it "sets a particular column as index but keeps that column" do
+      expect(df.set_index(:c, keep: true)).to eq(
+        DaruLite::DataFrame.new({
+          a: [1,2,3,4,5],
+          b: ['a','b','c','d','e'],
+          c: [11,22,33,44,55]
+          }, index: [11,22,33,44,55]))
+      expect(df[:c]).to eq(df[:c])
+    end
+
+    it "sets categorical index if categorical is true" do
+      data = {
+        a: [1, 2, 3, 4, 5],
+        b: [:a, 1, :a, 1, 'c'],
+        c: %w[a b c d e]
+      }
+      df = DaruLite::DataFrame.new(data)
+      df.set_index(:b, categorical: true)
+      expected = DaruLite::DataFrame.new(
+        data.slice(:a, :c),
+        index: DaruLite::CategoricalIndex.new(data[:b])
+      )
+      expect(df).to eq(expected)
+    end
+
+    it "raises error if all elements in the column aren't unique" do
+      jholu = DaruLite::DataFrame.new({
+        a: ['a','b','a'],
+        b: [1,2,4]
+        })
+
+      expect {
+        jholu.set_index(:a)
+      }.to raise_error(ArgumentError)
+    end
+
+    it "sets multiindex if array is given" do
+      df = DaruLite::DataFrame.new({
+        a: %w[a a b b],
+        b: [1, 2, 1, 2],
+        c: %w[a b c d]
+      })
+      df.set_index(%i[a b])
+      expected =
+        DaruLite::DataFrame.new(
+          { c: %w[a b c d] },
+          index: DaruLite::MultiIndex.from_tuples(
+            [['a', 1], ['a', 2], ['b', 1], ['b', 2]]
+          )
+        ).tap do |df|
+          df.index.name = %i[a b]
+          df
+        end
+      expect(df).to eq(expected)
+    end
+  end
+
+  describe "#reindex" do
+    subject { df.reindex(DaruLite::Index.new([1,3,0,8,2])) }
+
+    let(:df) do
+      DaruLite::DataFrame.new({
+        a: [1,2,3,4,5],
+        b: [11,22,33,44,55],
+        c: %w(a b c d e)
+      })
+    end
+
+    it "re indexes and aligns accordingly" do
+      expect(subject).to eq(
+        DaruLite::DataFrame.new(
+          {
+            a: [2,4,1,nil,3],
+            b: [22,44,11,nil,33],
+            c: ['b','d','a',nil,'c']
+          },
+          index: DaruLite::Index.new([1,3,0,8,2])
+        )
+      )
+    end
+
+    it { is_expected.to_not eq(df) }
+  end
+
+  describe '#reset_index' do
+    context 'when Index' do
+      subject do
+        DaruLite::DataFrame.new(
+          {'vals' => [1,2,3,4,5]},
+          index: DaruLite::Index.new(%w[a b c d e], name: 'indices')
+        ).reset_index
+      end
+
+      it { is_expected.to eq DaruLite::DataFrame.new(
+        'indices' => %w[a b c d e],
+        'vals' => [1,2,3,4,5]
+      )}
+    end
+
+    context 'when MultiIndex' do
+      subject do
+        mi = DaruLite::MultiIndex.from_tuples([
+          [0, 'a'], [0, 'b'], [1, 'a'], [1, 'b']
+        ])
+        mi.name = %w[nums alphas]
+        DaruLite::DataFrame.new(
+          {'vals' => [1,2,3,4]},
+          index: mi
+        ).reset_index
+      end
+
+      it { is_expected.to eq DaruLite::DataFrame.new(
+        'nums' => [0,0,1,1],
+        'alphas' => %w[a b a b],
+        'vals' => [1,2,3,4]
+      )}
+    end
+  end
+
+  describe "#index=" do
+    let(:df) do
+      DaruLite::DataFrame.new({
+        a: [1,2,3,4,5],
+        b: [11,22,33,44,55],
+        c: %w(a b c d e)
+      })
+    end
+
+    it "simply reassigns the index" do
+      df.index = DaruLite::Index.new(['4','foo', :bar, 0, 23])
+      expect(df.row['foo']).to eq(DaruLite::Vector.new([2,22,'b'], index: [:a,:b,:c]))
+    end
+
+    it "raises error for improper length index" do
+      expect {
+        df.index = DaruLite::Index.new([1,2])
+      }.to raise_error(ArgumentError)
+    end
+
+    it "is able to accept array" do
+      df.index = (1..5).to_a
+      expect(df.index).to eq DaruLite::Index.new (1..5).to_a
+    end
+  end
+
+  describe "#reindex_vectors" do
+    it "re indexes vectors and aligns accordingly" do
+      df = DaruLite::DataFrame.new({
+        a: [1,2,3,4,5],
+        b: [11,22,33,44,55],
+        c: %w(a b c d e)
+      })
+
+      ans = df.reindex_vectors(DaruLite::Index.new([:b, 'a', :a]))
+      expect(ans).to eq(DaruLite::DataFrame.new({
+        :b  => [11,22,33,44,55],
+        'a' => [nil, nil, nil, nil, nil],
+        :a  => [1,2,3,4,5]
+      }, order: [:b, 'a', :a]))
+    end
+
+    it 'raises ArgumentError if argument was not an index' do
+      df = DaruLite::DataFrame.new([])
+      expect { df.reindex_vectors([]) }.to raise_error(ArgumentError)
+    end
+  end
+
+  describe "#vectors=" do
+    let(:df) do
+      DaruLite::DataFrame.new({
+        a: [1,2,3,4,5],
+        b: [11,22,33,44,55],
+        c: %w(a b c d e)
+      })
+    end
+
+    it "simply reassigns vectors" do
+      df.vectors = DaruLite::Index.new(['b',0,'m'])
+
+      expect(df.vectors).to eq(DaruLite::Index.new(['b',0,'m']))
+      expect(df['b']).to eq(DaruLite::Vector.new([1,2,3,4,5]))
+      expect(df[0]).to eq(DaruLite::Vector.new([11,22,33,44,55]))
+      expect(df['m']).to eq(DaruLite::Vector.new(%w(a b c d e)))
+    end
+
+    it "raises error for improper length index" do
+      expect {
+        df.vectors = DaruLite::Index.new([1,2,'3',4,'5'])
+      }.to raise_error(ArgumentError)
+    end
+
+    it "change name of vectors in @data" do
+      new_index_array = [:k, :l, :m]
+      df.vectors = DaruLite::Index.new(new_index_array)
+
+      expect(df.data.map { |vector| vector.name }).to eq(new_index_array)
+    end
+  end
+end