bmg 0.18.13 → 0.19.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 0ceadf8947aa041841941225e88396cc37ff7319
-  data.tar.gz: 9c695d3305101ff4a73bcdef5e3788648e9b2d27
+  metadata.gz: 5d3871be740059340140a52f8b47892ef64effee
+  data.tar.gz: 721d3073d321a0d5b6bc3a1a8426522e35cc7138
 SHA512:
-  metadata.gz: 70db4550decd609934e1d8d643983766b5b50b9f0354b2405dfdc710b9580a885e4754c907ed5011856e905308af9647c029f0dd612657e72b45af0465eaced2
-  data.tar.gz: 5ed5e7f6d1d9baa8070112a36229e8961e4b5a531acbdac808fcf1f5c008fc91ae3ceadc5c3b4c1d0ace22ca6f861074e79b6d957498740a8fba3d5e99875068
+  metadata.gz: 28e1cd068d476e15f0577f70d9e84ca74f2db6e4dc849945c6c94bcf280ad76cc160276d00705cb98a2a5c61db785352a805c3c76034251b1da41c5f511669d5
+  data.tar.gz: b36c7e59b259c5072a37b64f33b422c376794bca64ed02623f7124acd786c1e5f3792a9264ce6cc1f672c2dd281c34e806965a7f0a67bb585afd81a67b830caa

data/lib/bmg/reader/csv.rb

@@ -10,20 +10,18 @@ module Bmg
       }
 
       def initialize(type, path_or_io, options = {})
-        @type = type
+        require 'csv'
+
         @path_or_io = path_or_io
-        @options = DEFAULT_OPTIONS.merge(options)
-        if @options[:smart] && !@path_or_io.is_a?(IO)
-          @options[:col_sep] ||= infer_col_sep
-          @options[:quote_char] ||= infer_quote_char
-        end
+        @options = handle_options(options)
+        @type = handle_type(type)
       end
 
       def each
         return to_enum unless block_given?
-        require 'csv'
-        with_io do |io|
-          ::CSV.new(io, **csv_options).each do |row|
+
+        with_csv do |csv|
+          csv.each do |row|
             yield tuple(row)
           end
         end
@@ -44,33 +42,70 @@ module Bmg
         row.to_hash.each_with_object({}){|(k,v),h| h[k.to_sym] = v }
       end
 
+      def handle_type(type)
+        return type if type.knows_attrlist?
+
+        type.with_attrlist(infer_attrlist)
+      end
+
+      def infer_attrlist
+        with_csv do |csv|
+          csv.each do |row|
+            return tuple(row).keys
+          end
+        end
+      end
+
+      def handle_options(options)
+        options = DEFAULT_OPTIONS.merge(options)
+        if options[:smart] && !@path_or_io.is_a?(IO)
+          options[:col_sep] ||= infer_col_sep
+          options[:quote_char] ||= infer_quote_char(options[:col_sep])
+        end
+        options
+      end
+
       def infer_col_sep
         sniff(text_portion, [",","\t",";"], ",")
       end
 
-      def infer_quote_char
-        sniff(text_portion, ["'","\""], "\"")
+      def infer_quote_char(col_sep)
+        sniff(text_portion, ['"',"'"], '"'){|quote|
+          /#{quote}#{col_sep}#{quote}|^#{quote}|#{quote}$/
+        }
       end
 
       def text_portion
-        @text_portion ||= with_io{|io| io.readlines(10).join("\n") }
+        @text_portion ||= with_io{|io| io.readlines(50).join("\n") }
       end
 
       def with_io(&bl)
         case @path_or_io
         when IO, StringIO
+          @path_or_io.rewind if @path_or_io.respond_to?(:rewind)
           bl.call(@path_or_io)
         else
           File.open(@path_or_io, "r", &bl)
         end
       end
 
+      def with_csv(&bl)
+        with_io do |io|
+          yield ::CSV.new(io, **csv_options)
+        end
+      end
+
       # Finds the best candidate among `candidates` for a separator
       # found in `str`. If none is found, returns `default`.
-      def sniff(str, candidates, default)
+      def sniff(str, candidates, default, &bl)
         snif = {}
         candidates.each {|delim|
-          snif[delim] = str.count(delim)
+          counter = bl ? bl.call(delim) : delim
+          snif[delim] = if counter.is_a?(Regexp)
+            str.scan(counter).length
+          else
+            str.count(counter)
+          end
         }
         snif = snif.sort {|a,b| b[1] <=> a[1] }
         snif.size > 0 ? snif[0][0] : default

data/lib/bmg/reader/excel.rb

@@ -10,32 +10,27 @@ module Bmg
       }
 
       def initialize(type, path, options = {})
-        @type = type
+        require 'roo'
         @path = path
         @options = DEFAULT_OPTIONS.merge(options)
+        @type = type.knows_attrlist? ? type : type.with_attrlist(infer_attrlist)
       end
 
       def each
         return to_enum unless block_given?
-        require 'roo'
-        xlsx = Roo::Spreadsheet.open(@path, @options)
-        headers = nil
-        xlsx.sheet(@options[:sheet])
-          .each
-          .drop(@options[:skip])
-          .each_with_index
-          .each do |row, i|
-            if i==0
-              headers = row.map{|c| c.to_s.strip.to_sym }
-            else
-              init = init_tuple(i)
-              tuple = (0...headers.size)
-                .each_with_object(init){|i,t|
-                  t[headers[i]] = row[i]
-                }
-              yield(tuple)
-            end
-          end
+
+        headers = type.attrlist
+        headers = headers[1..-1] if generate_row_num?
+        start_at = @options[:skip] + 2
+        end_at = spreadsheet.last_row
+        (start_at..end_at).each do |i|
+          row = spreadsheet.row(i)
+          init = init_tuple(i - start_at + 1)
+          tuple = (0...headers.size).each_with_object(init){|i,t|
+            t[headers[i]] = row[i]
+          }
+          yield(tuple)
+        end
       end
 
       def to_ast
@@ -49,17 +44,37 @@ module Bmg
 
     private
 
-      def init_tuple(i)
+      def spreadsheet
+        @spreadsheet ||= Roo::Spreadsheet
+          .open(@path, @options)
+          .sheet(@options[:sheet])
+      end
+
+      def infer_attrlist
+        row = spreadsheet.row(1+@options[:skip])
+        attrlist = row.map{|c| c.to_s.strip.to_sym }
+        attrlist.unshift(row_num_name) if generate_row_num?
+        attrlist
+      end
+
+      def generate_row_num?
+        !!@options[:row_num]
+      end
+
+      def row_num_name
         case as = @options[:row_num]
-        when TrueClass
-          { :row_num => i }
-        when FalseClass
-          {}
-        when Symbol
-          { :"#{as}" => i }
+        when TrueClass  then :row_num
+        when Symbol     then as
+        else nil
         end
       end
 
+      def init_tuple(i)
+        return {} unless generate_row_num?
+
+        { row_num_name => i }
+      end
+
     end # class Excel
   end # module Reader
-end # module Bmg
+end # module Bmg

data/lib/bmg/relation/materialized.rb

@@ -5,6 +5,7 @@ module Bmg
 
       def initialize(operand)
         @operand = operand
+        @materialized = nil
       end
 
       def type
@@ -40,9 +41,9 @@ module Bmg
     private
 
       def _materialize
-        return @operand if @operand.is_a?(Relation::InMemory)
+        return @materialized if @materialized
 
-        @operand = Relation::InMemory.new(operand.type, operand.to_a)
+        @materialized = Relation::InMemory.new(operand.type, operand.to_a)
       end
 
     end # class Materialized

data/lib/bmg/sql/processor/semi_join.rb

@@ -39,9 +39,6 @@ module Bmg
             subquery = All.new(builder).call(subquery)
             if commons.size == 0
               builder.exists(subquery)
-            elsif commons.size == 1
-              identifier = left.desaliaser(true)[commons.to_a.first]
-              Predicate::Factory.in(identifier, Predicate::Factory.opaque(subquery))
             else
               join_pre  = join_predicate(left, subquery, commons)
               subquery  = expand_where_clause(subquery, join_pre)

data/lib/bmg/version.rb

@@ -1,8 +1,8 @@
 module Bmg
   module Version
     MAJOR = 0
-    MINOR = 18
-    TINY  = 13
+    MINOR = 19
+    TINY  = 0
   end
   VERSION = "#{Version::MAJOR}.#{Version::MINOR}.#{Version::TINY}"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bmg
 version: !ruby/object:Gem::Version
-  version: 0.18.13
+  version: 0.19.0
 platform: ruby
 authors:
 - Bernard Lambeau
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2022-01-25 00:00:00.000000000 Z
+date: 2022-04-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: predicate