arc-furnace 0.1.12 → 0.1.13

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: d72d07c3b8825ef0dafa1cea574ae5fdd9fb8bfd
-  data.tar.gz: d7c6dd81528cce65a1bd3919d78369c63587cd09
+  metadata.gz: 401438a69d2165a093c17faf0e8c3fc10b960155
+  data.tar.gz: 870456329b0d027b6f627b97355bc91c42854c8d
 SHA512:
-  metadata.gz: 95e6a1c890f47069d918d706e2f8f466c0854b356f4ef0d56a54d74be7881b9a31f9e81e323e75d8c1d732dd4a16bd9907517e0439d3423334ee70a5e5d93fbe
-  data.tar.gz: a0d046f0cf4f882be6507c24e5e067b32d95bd1317e636c0a0bd3f71f04e560b4848486abec9248f83ef59ab6f237bcc139d639ce5f6b8c759430e868793326d
+  metadata.gz: 2a5e9147e64a189b1874d9a06f5fff23d0f9c89b06208058dc6232ab520b342fa83d104fc0e0d5f7d8fd21e8cfb44b577fa8fa54668e117ad2f3d9086c01322f
+  data.tar.gz: 0e14d178fc361a6ea7cb7e8a96be7537064f2cb95ea2a487fe00d0044ea14d824b89b5d9cfe5871361c5f2d778b3273a6c8d1094b0432940395c89b2920e3f7c

data/lib/arc-furnace/csv_source.rb

@@ -1,33 +1,25 @@
-require 'arc-furnace/source'
 require 'arc-furnace/csv_to_hash_with_duplicate_headers'
+require 'arc-furnace/enumerator_source'
 require 'csv'
 
 module ArcFurnace
-  class CSVSource < Source
+  class CSVSource < EnumeratorSource
     include CSVToHashWithDuplicateHeaders
-    private_attr_reader :csv
     attr_reader :value, :file
 
     def initialize(filename: , encoding: 'UTF-8')
       @file = File.open(filename, encoding: encoding)
-      @csv = CSV.new(file, encoding: encoding, headers: true).each
-      advance
+      super()
     end
 
-    # Is this source empty?
-    def empty?
-      !value
+    def finalize
+      file.close
     end
 
-    def advance
-      @value =
-        begin
-          csv_to_hash_with_duplicates(csv.next) if csv
-        rescue StopIteration
-          file.close
-          @csv = nil
-          nil
-        end
+    def build_enumerator
+      Enumerator.new do |yielder|
+        CSV.new(file, headers: true).each { |row| yielder << csv_to_hash_with_duplicates(row) }
+      end
     end
   end
 end

data/lib/arc-furnace/enumerator_source.rb

@@ -8,9 +8,16 @@ module ArcFurnace
 
     def initialize
       @enumerator = build_enumerator
+      preprocess
       advance
     end
 
+    # Called after setting up the enumerator but before advancing it
+    # Use this to extract header rows for instance.
+    def preprocess
+      # nothing
+    end
+
     # Is this source empty?
     def empty?
       !value

data/lib/arc-furnace/excel_source.rb

@@ -4,7 +4,7 @@ require 'roo'
 module ArcFurnace
   class ExcelSource < EnumeratorSource
 
-    private_attr_reader :excel, :enumerator
+    private_attr_reader :excel, :header_row
     attr_reader :value
 
     def initialize(filename: , sheet: nil)
@@ -19,21 +19,30 @@ module ArcFurnace
       @excel.close if @excel
     end
 
-    def build_enumerator
-      header_row = excel.row(1)
+    def preprocess
+      enumerator.next
+    end
 
-      last_row_index = excel.last_row
-      current_row_index = 2
+    def extract_cell_value(cell)
+      if cell
+        coerced_value = cell.type == :string ? cell.value : cell.excelx_value.try(:to_s).try(:strip)
+        coerced_value unless coerced_value.blank?
+      end
+    end
 
+    def build_enumerator
       Enumerator.new do |yielder|
-        until current_row_index > last_row_index
-          row = header_row.each_with_object(::Hash.new).each_with_index do |(header, result), index|
-            value = excel.cell(current_row_index, index + 1)
-            coerced_value = (value.is_a?(String) ? value : excel.excelx_value(current_row_index, index + 1)).try(:to_s).try(:strip)
-            result[header] = coerced_value unless coerced_value.blank?
-          end
-          current_row_index += 1
-          yielder << row
+        excel.each_row_streaming do |row|
+          yielder <<
+              if header_row
+                row.each_with_object({}) do |cell, result|
+                  value = extract_cell_value(cell)
+                  result[header_row[cell.coordinate.column - 1]] = value if value
+                end
+              else
+                # First time, return the header row so we can save it.
+                @header_row = row.map { |value| extract_cell_value(value) }
+              end
         end
       end
     end

data/lib/arc-furnace/version.rb

@@ -1,3 +1,3 @@
 module ArcFurnace
-  VERSION = "0.1.12"
+  VERSION = "0.1.13"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: arc-furnace
 version: !ruby/object:Gem::Version
-  version: 0.1.12
+  version: 0.1.13
 platform: ruby
 authors:
 - Daniel Spangenberger
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2016-01-05 00:00:00.000000000 Z
+date: 2016-01-15 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: msgpack