bmg 0.18.2 → 0.18.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2c9b490328160c9816102f7dd34b3ee46207ab4362cd2af6d8623c9fc5d4b87a
-  data.tar.gz: d113dc14ae1811f2b5565af9e84a0ffbc82b7842c46284af7fcad3bc9f3b7470
+  metadata.gz: 340efbb84309b24fd818bd8ff0c0f40b1f2ba3b3f087e3af0ea574fbd1f7c4ef
+  data.tar.gz: d7c6af70390c893f6e198f89a019141b4d5727326183f1663aa7d2e15a7ef694
 SHA512:
-  metadata.gz: fd42b8787cc093c6452386760744d7480c64c8be186ab73b87a548a1b853643b46a6ad48cf019da9d77fddbb38c824eb0e0bef3ef5fa2f333f1d1240f5a0bd84
-  data.tar.gz: 7bd0cd32a28963123b03b59152fa03d6470bbdf1a3c6cfe0b17a6667a80d97e65d5309def4e31ec9cbb47353d409ad9ad2a6d1fbc9c2ec10797460852836ce35
+  metadata.gz: 69562558e784e35f6c8a81bf8b32f8f1a24bbabc176ca8fc109e8151ad324d15dece34d58cf44196a78387b645ae21ec2cb361cc71e11f93b14fd0c916bd1573
+  data.tar.gz: c25d6966c201c265d2b42dd976e29faf075a8ea285d609883f1f4a9d3c52c941e4c690741e5e277bd4ea8a7d8cfefa4d9adea3a575e2c6006b6993ad767351fb

data/lib/bmg/operator/allbut.rb

@@ -30,6 +30,7 @@ module Bmg
     public
 
       def each
+        return to_enum unless block_given?
         seen = {}
         @operand.each do |tuple|
           allbuted = tuple_allbut(tuple)

data/lib/bmg/operator/autosummarize.rb

@@ -41,6 +41,7 @@ module Bmg
       end
 
       def each(&bl)
+        return to_enum unless block_given?
         h = {}
         @operand.each do |tuple|
           key = key(tuple)

data/lib/bmg/operator/autowrap.rb

@@ -43,6 +43,7 @@ module Bmg
       end
 
       def each
+        return to_enum unless block_given?
         @operand.each do |tuple|
           yield autowrap_tuple(tuple)
         end

data/lib/bmg/operator/constants.rb

@@ -23,6 +23,7 @@ module Bmg
     public
 
       def each
+        return to_enum unless block_given?
         @operand.each do |tuple|
           yield extend_it(tuple)
         end

data/lib/bmg/operator/extend.rb

@@ -26,6 +26,7 @@ module Bmg
     public
 
       def each
+        return to_enum unless block_given?
         @operand.each do |tuple|
           yield extend_it(tuple)
         end

data/lib/bmg/operator/group.rb

@@ -32,6 +32,7 @@ module Bmg
     public
 
       def each(&bl)
+        return to_enum unless block_given?
         index = Hash.new{|h,k| h[k] = k.merge(as => empty_group) }
         operand.each do |tuple|
           key = TupleAlgebra.allbut(tuple, attrs)

data/lib/bmg/operator/image.rb

@@ -46,6 +46,7 @@ module Bmg
     public
 
       def each(*args, &bl)
+        return to_enum unless block_given?
         (options[:jit_optimized] ? self : jit_optimize)._each(*args, &bl)
       end
 

data/lib/bmg/operator/join.rb

@@ -25,6 +25,7 @@ module Bmg
     public
 
       def each
+        return to_enum unless block_given?
         index = Hash.new
         right.each_with_object(index) do |t, index|
           key = tuple_project(t, on)

data/lib/bmg/operator/matching.rb

@@ -23,6 +23,7 @@ module Bmg
     public
 
       def each
+        return to_enum unless block_given?
         index = Hash.new
         right.each_with_object(index) do |t, index|
           key = tuple_project(t, on)

data/lib/bmg/operator/not_matching.rb

@@ -23,6 +23,7 @@ module Bmg
     public
 
       def each
+        return to_enum unless block_given?
         index = Hash.new
         right.each_with_object(index) do |t, index|
           key = tuple_project(t, on)

data/lib/bmg/operator/page.rb

@@ -30,6 +30,7 @@ module Bmg
     public
 
       def each(&bl)
+        return to_enum unless block_given?
         page_size = options[:page_size]
         @operand.to_a
           .sort(&comparator)

data/lib/bmg/operator/project.rb

@@ -29,6 +29,7 @@ module Bmg
     public
 
       def each
+        return to_enum unless block_given?
         seen = {}
         @operand.each do |tuple|
           projected = tuple_project(tuple)

data/lib/bmg/operator/rename.rb

@@ -29,6 +29,7 @@ module Bmg
     public
 
       def each
+        return to_enum unless block_given?
         @operand.each do |tuple|
           yield rename(tuple, renaming)
         end

data/lib/bmg/operator/restrict.rb

@@ -26,6 +26,7 @@ module Bmg
       end
 
       def each
+        return to_enum unless block_given?
         @operand.each do |tuple|
           yield(tuple) if @predicate.evaluate(tuple)
         end

data/lib/bmg/operator/rxmatch.rb

@@ -32,6 +32,7 @@ module Bmg
     public
 
       def each
+        return to_enum unless block_given?
         @operand.each do |tuple|
           against = attrs.map{|a| tuple[a] }.join(" ")
           matcher = self.matcher

data/lib/bmg/operator/summarize.rb

@@ -13,7 +13,7 @@ module Bmg
         @type = type
         @operand = operand
         @by = by
-        @summarization = Summarize.compile(summarization)
+        @summarization = Summarizer.summarization(summarization)
       end
 
     protected
@@ -23,6 +23,7 @@ module Bmg
     public
 
       def each
+        return to_enum unless block_given?
         # summary key => summarization memo, starting with least
         result = Hash.new{|h,k|
           h[k] = Hash[@summarization.map{|k,v|
@@ -56,22 +57,6 @@ module Bmg
         [ by, summarization ]
       end
 
-    private
-
-      # Compile a summarization hash so that every value is a Summarizer
-      # instance
-      def self.compile(summarization)
-        Hash[summarization.map{|k,v|
-          summarizer = case v
-          when Summarizer then v
-          when Symbol     then Summarizer.send(v, k)
-          else
-            raise ArgumentError, "Unexpected summarizer #{k} => #{v}"
-          end
-          [ k, summarizer ]
-        }]
-      end
-
     end # class Summarize
   end # module Operator
 end # module Bmg

data/lib/bmg/operator/transform.rb

@@ -28,6 +28,7 @@ module Bmg
     public
 
       def each
+        return to_enum unless block_given?
         t = transformer
         @operand.each do |tuple|
           yield t.call(tuple)

data/lib/bmg/operator/union.rb

@@ -37,6 +37,7 @@ module Bmg
       end
 
       def each(&bl)
+        return to_enum unless block_given?
         if all?
           operands.each do |op|
             op.each(&bl)

data/lib/bmg/reader/csv.rb

@@ -5,30 +5,36 @@ module Bmg
 
       DEFAULT_OPTIONS = {
         :headers => true,
-        :return_headers => false
+        :return_headers => false,
+        :smart => true
       }
 
-      def initialize(type, path, options = {})
+      def initialize(type, path_or_io, options = {})
         @type = type
-        @path = path
+        @path_or_io = path_or_io
         @options = DEFAULT_OPTIONS.merge(options)
-        @options[:col_sep] ||= infer_col_sep
-        @options[:quote_char] ||= infer_quote_char
+        if @options[:smart] && !@path_or_io.is_a?(IO)
+          @options[:col_sep] ||= infer_col_sep
+          @options[:quote_char] ||= infer_quote_char
+        end
       end
 
       def each
+        return to_enum unless block_given?
         require 'csv'
-        ::CSV.foreach(@path, @options) do |row|
-          yield tuple(row)
+        with_io do |io|
+          ::CSV.new(io, csv_options).each do |row|
+            yield tuple(row)
+          end
         end
       end
 
       def to_ast
-        [ :csv, @path, @options ]
+        [ :csv, @path_or_io, @options ]
       end
 
       def to_s
-        "(csv #{path})"
+        "(csv #{@path_or_io})"
       end
       alias :inspect :to_s
 
@@ -47,7 +53,16 @@ module Bmg
       end
 
       def text_portion
-        @text_portion ||= File.foreach(@path).first(10).join("\n")
+        @text_portion ||= with_io{|io| io.readlines(10).join("\n") }
+      end
+
+      def with_io(&bl)
+        case @path_or_io
+        when IO, StringIO
+          bl.call(@path_or_io)
+        else
+          File.open(@path_or_io, "r", &bl)
+        end
       end
 
       # Finds the best candidate among `candidates` for a separator
@@ -61,6 +76,10 @@ module Bmg
         snif.size > 0 ? snif[0][0] : default
       end
 
+      def csv_options
+        @csv_options ||= @options.dup.tap{|opts| opts.delete(:smart) }
+      end
+
     end # class Csv
   end # module Reader
 end # module Bmg

data/lib/bmg/reader/excel.rb

@@ -4,7 +4,8 @@ module Bmg
       include Reader
 
       DEFAULT_OPTIONS = {
-        skip: 0
+        skip: 0,
+        row_num: true
       }
 
       def initialize(type, path, options = {})
@@ -14,6 +15,7 @@ module Bmg
       end
 
       def each
+        return to_enum unless block_given?
         require 'roo'
         xlsx = Roo::Spreadsheet.open(@path, @options)
         headers = nil
@@ -25,7 +27,11 @@ module Bmg
             if i==0
               headers = row.map(&:to_sym)
             else
-              tuple = (0...headers.size).each_with_object({}){|i,t| t[headers[i]] = row[i] }
+              init = init_tuple(i)
+              tuple = (0...headers.size)
+                .each_with_object(init){|i,t|
+                  t[headers[i]] = row[i]
+                }
               yield(tuple)
             end
           end
@@ -36,10 +42,23 @@ module Bmg
       end
 
       def to_s
-        "(excel #{path})"
+        "(excel #{@path})"
       end
       alias :inspect :to_s
 
+    private
+
+      def init_tuple(i)
+        case as = @options[:row_num]
+        when TrueClass
+          { :row_num => i }
+        when FalseClass
+          {}
+        when Symbol
+          { :"#{as}" => i }
+        end
+      end
+
     end # class Excel
   end # module Reader
 end # module Bmg

data/lib/bmg/relation.rb

@@ -27,6 +27,12 @@ module Bmg
       }
     end
 
+    def with_type_attrlist
+      return self if type.knows_attrlist?
+      attrs = self.first.keys
+      with_type(type.with_attrlist(attrs))
+    end
+
     def with_typecheck
       dup.tap{|r|
         r.type = r.type.with_typecheck

data/lib/bmg/relation/in_memory.rb

@@ -8,7 +8,6 @@ module Bmg
         @type = type
       end
       attr_accessor :type
-      protected :type=
       attr_reader :operand
 
     public

data/lib/bmg/sequel/relation.rb

@@ -9,6 +9,7 @@ module Bmg
       attr_reader :sequel_db
 
       def each(&bl)
+        return to_enum unless block_given?
         dataset.each(&bl)
       end
 

data/lib/bmg/sql/relation.rb

@@ -133,8 +133,8 @@ module Bmg
         _instance(type, builder, expr)
       end
 
-      def _summarize(type, by, summarization)
-        summarization = Operator::Summarize.compile(summarization)
+      def _summarize(type, by, defs)
+        summarization = ::Bmg::Summarizer.summarization(defs)
         if can_compile_summarization?(summarization)
           expr = before_use(self.expr)
           expr = Processor::Summarize.new(by, summarization, builder).call(self.expr)

data/lib/bmg/summarizer.rb

@@ -50,6 +50,21 @@ module Bmg
       end
     end
 
+    # Converts some summarization definitions to a Hash of
+    # summarizers.
+    def self.summarization(defs)
+      Hash[defs.map{|k,v|
+        summarizer = case v
+        when Summarizer then v
+        when Symbol     then Summarizer.send(v, k)
+        when Proc       then Summarizer.by_proc(&v)
+        else
+          raise ArgumentError, "Unexpected summarizer #{k} => #{v}"
+        end
+        [ k, summarizer ]
+      }]
+    end
+
     # Returns the default options to use
     #
     # @return the default aggregation options
@@ -80,7 +95,16 @@ module Bmg
     # @param the current iterated tuple
     # @return updated memo value
     def happens(memo, tuple)
-      value = @functor.is_a?(Proc) ? @functor.call(tuple) : tuple[@functor]
+      value = case @functor
+      when Proc
+        @functor.call(tuple)
+      when NilClass
+        tuple
+      when Symbol
+        tuple[@functor]
+      else
+        tuple[@functor]
+      end
       _happens(memo, value)
     end
 
@@ -128,5 +152,9 @@ require_relative 'summarizer/max'
 require_relative 'summarizer/avg'
 require_relative 'summarizer/variance'
 require_relative 'summarizer/stddev'
+require_relative 'summarizer/percentile'
 require_relative 'summarizer/collect'
+require_relative 'summarizer/distinct'
 require_relative 'summarizer/concat'
+require_relative 'summarizer/by_proc'
+require_relative 'summarizer/multiple'

data/lib/bmg/summarizer/avg.rb

@@ -16,13 +16,13 @@ module Bmg
       end
 
       # Collects one more value + the sum of all
-      def _happens(memo, val) 
+      def _happens(memo, val)
         [memo.first + val, memo.last + 1]
       end
 
       # Finalizes the computation.
-      def finalize(memo) 
-        memo.first / memo.last 
+      def finalize(memo)
+        memo.first / memo.last
       end
 
     end # class Avg

data/lib/bmg/summarizer/by_proc.rb

@@ -0,0 +1,41 @@
+module Bmg
+  class Summarizer
+    #
+    # Generic summarizer that takes a Proc àla each_with_object.
+    #
+    # Example:
+    #
+    #   # direct ruby usage
+    #   Bmg::Summarizer.by_proc{|t,memo| ... }.summarize(...)
+    #
+    class ByProc < Summarizer
+
+      def initialize(least, by_proc)
+        @least = least
+        @by_proc = by_proc
+      end
+
+      # Returns [] as least value.
+      def least()
+        @least
+      end
+
+      # Adds val to the memo array
+      def happens(memo, val)
+        @by_proc.call(val, memo)
+      end
+
+      def finalize(memo)
+        memo
+      end
+
+    end # class ByProc
+
+    # Factors a distinct summarizer
+    def self.by_proc(least = nil, proc = nil, &bl)
+      least, proc = nil, least if least.is_a?(Proc)
+      ByProc.new(least, proc || bl)
+    end
+
+  end # class Summarizer
+end # module Bmg

data/lib/bmg/summarizer/distinct.rb

@@ -0,0 +1,36 @@
+module Bmg
+  class Summarizer
+    #
+    # Collect the distinct values as an array.
+    #
+    # Example:
+    #
+    #   # direct ruby usage
+    #   Bmg::Summarizer.distinct(:qty).summarize(...)
+    #
+    class Distinct < Summarizer
+
+      # Returns [] as least value.
+      def least()
+        {}
+      end
+
+      # Adds val to the memo array
+      def _happens(memo, val)
+        memo[val] = true
+        memo
+      end
+
+      def finalize(memo)
+        memo.keys
+      end
+
+    end # class Distinct
+
+    # Factors a distinct summarizer
+    def self.distinct(*args, &bl)
+      Distinct.new(*args, &bl)
+    end
+
+  end # class Summarizer
+end # module Bmg

data/lib/bmg/summarizer/multiple.rb

@@ -0,0 +1,46 @@
+module Bmg
+  class Summarizer
+    #
+    # A summarizer that collects multiple summarization as a wrapped
+    # tuple.
+    #
+    # Example:
+    #
+    #   # direct ruby usage
+    #   Bmg::Summarizer.multiple(x: ..., y: ...).summarize(...)
+    #
+    class Multiple < Summarizer
+
+      def initialize(defs)
+        @summarization = Summarizer.summarization(defs)
+      end
+
+      # Returns [] as least value.
+      def least()
+        @summarization.each_pair.each_with_object({}){|(k,v),memo|
+          memo[k] = v.least
+        }
+      end
+
+      # Adds val to the memo array
+      def happens(memo, val)
+        @summarization.each_pair.each_with_object({}){|(k,v),memo2|
+          memo2[k] = v.happens(memo[k], val)
+        }
+      end
+
+      def finalize(memo)
+        @summarization.each_pair.each_with_object({}){|(k,v),memo2|
+          memo2[k] = v.finalize(memo[k])
+        }
+      end
+
+    end # class Multiple
+
+    # Factors a distinct summarizer
+    def self.multiple(defs)
+      Multiple.new(defs)
+    end
+
+  end # class Summarizer
+end # module Bmg

data/lib/bmg/summarizer/percentile.rb

@@ -0,0 +1,47 @@
+module Bmg
+  class Summarizer
+    #
+    # Percentile summarizer.
+    #
+    # Example:
+    #
+    #   # direct ruby usage
+    #   Bmg::Summarizer.percentile(:qty, 50).summarize(...)
+    #
+    class Percentile < Summarizer
+
+      def initialize(attribute = nil, nth = 50, &bl)
+        attribute, nth = nil, attribute if attribute.is_a?(Integer)
+        super(*[attribute].compact, &bl)
+        @nth = nth
+      end
+
+      # Returns [] as least value.
+      def least()
+        []
+      end
+
+      # Collects the value
+      def _happens(memo, val)
+        memo << val
+      end
+
+      # Finalizes the computation.
+      def finalize(memo)
+        return nil if memo.empty?
+        index = memo.size * (@nth / 100.0)
+        above = [[index.ceil - 1, memo.size - 1].min, 0].max
+        below = [index.floor - 1, 0].max
+        sorted = memo.sort
+        (sorted[above] + sorted[below]) / 2
+      end
+
+    end # class Avg
+
+    # Factors an average summarizer
+    def self.percentile(*args, &bl)
+      Percentile.new(*args, &bl)
+    end
+
+  end # class Summarizer
+end # module Bmg

data/lib/bmg/version.rb

@@ -2,7 +2,7 @@ module Bmg
   module Version
     MAJOR = 0
     MINOR = 18
-    TINY  = 2
+    TINY  = 3
   end
   VERSION = "#{Version::MAJOR}.#{Version::MINOR}.#{Version::TINY}"
 end

data/lib/bmg/writer.rb

@@ -1 +1,17 @@
+module Bmg
+  module Writer
+
+  protected
+
+    def infer_headers(from)
+      attrlist = if from.is_a?(Type) && from.knows_attrlist?
+        from.to_attrlist
+      elsif from.is_a?(Hash)
+        from.keys
+      end
+      attrlist ? output_preferences.order_attrlist(attrlist) : nil
+    end
+
+  end # module Writer
+end # module Bmg
 require_relative 'writer/csv'

data/lib/bmg/writer/csv.rb

@@ -26,17 +26,6 @@ module Bmg
         to_s ? string_or_io.string : string_or_io
       end
 
-    private
-
-      def infer_headers(from)
-        attrlist = if from.is_a?(Type) && from.knows_attrlist?
-          from.to_attrlist
-        elsif from.is_a?(Hash)
-          from.keys
-        end
-        attrlist ? output_preferences.order_attrlist(attrlist) : nil
-      end
-
     end # class Csv
   end # module Writer
 end # module Bmg

data/lib/bmg/writer/xlsx.rb

@@ -0,0 +1,68 @@
+module Bmg
+  module Writer
+    class Xlsx
+      include Writer
+
+      DEFAULT_OPTIONS = {
+      }
+
+      def initialize(csv_options, output_preferences = nil)
+        @csv_options = DEFAULT_OPTIONS.merge(csv_options)
+        @output_preferences = OutputPreferences.dress(output_preferences)
+      end
+      attr_reader :csv_options, :output_preferences
+
+      def call(relation, path)
+        require 'write_xlsx'
+        dup._call(relation, path)
+      end
+
+    protected
+      attr_reader :workbook, :worksheet
+
+      def _call(relation, path)
+        @workbook = WriteXLSX.new(path)
+        @worksheet = workbook.add_worksheet
+
+        headers = infer_headers(relation.type)
+        relation.each_with_index do |tuple,i|
+          headers = infer_headers(tuple) if headers.nil?
+          headers.each_with_index do |h,i|
+            worksheet.write_string(0, i, h)
+          end if i == 0
+          headers.each_with_index do |h,j|
+            meth, *args = write_pair(tuple[h])
+            worksheet.send(meth, 1+i, j, *args)
+          end
+        end
+
+        workbook.close
+        path
+      end
+
+      def write_pair(value)
+        case value
+        when Numeric
+          [:write_number, value]
+        when Date
+          [:write_date_time, value, date_format]
+        else
+          [:write_string, value.to_s]
+        end
+      end
+
+      def date_format
+        @date_format ||= workbook.add_format(:num_format => 'yyyy-mm-dd')
+      end
+
+    end # class Xlsx
+  end # module Writer
+  module Relation
+
+    def to_xlsx(options = {}, path = nil, preferences = nil)
+      options, path = {}, options unless options.is_a?(Hash)
+      Writer::Xlsx.new(options, preferences).call(self, path)
+    end
+
+  end # module Relation
+end # module Bmg

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bmg
 version: !ruby/object:Gem::Version
-  version: 0.18.2
+  version: 0.18.3
 platform: ruby
 authors:
 - Bernard Lambeau
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-04-16 00:00:00.000000000 Z
+date: 2021-05-06 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: predicate
@@ -86,6 +86,20 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '2.8'
+- !ruby/object:Gem::Dependency
+  name: write_xlsx
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.0'
 - !ruby/object:Gem::Dependency
   name: sequel
   requirement: !ruby/object:Gem::Requirement
@@ -251,11 +265,15 @@ files:
 - lib/bmg/sql/version.rb
 - lib/bmg/summarizer.rb
 - lib/bmg/summarizer/avg.rb
+- lib/bmg/summarizer/by_proc.rb
 - lib/bmg/summarizer/collect.rb
 - lib/bmg/summarizer/concat.rb
 - lib/bmg/summarizer/count.rb
+- lib/bmg/summarizer/distinct.rb
 - lib/bmg/summarizer/max.rb
 - lib/bmg/summarizer/min.rb
+- lib/bmg/summarizer/multiple.rb
+- lib/bmg/summarizer/percentile.rb
 - lib/bmg/summarizer/stddev.rb
 - lib/bmg/summarizer/sum.rb
 - lib/bmg/summarizer/variance.rb
@@ -269,6 +287,7 @@ files:
 - lib/bmg/version.rb
 - lib/bmg/writer.rb
 - lib/bmg/writer/csv.rb
+- lib/bmg/writer/xlsx.rb
 - tasks/gem.rake
 - tasks/test.rake
 homepage: http://github.com/enspirit/bmg