arc-furnace 0.1.14 → 0.1.15

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: c014a227e52695c9f5a91ac292f21ff1850785f6
-  data.tar.gz: b273460b4eafdf5be12d57d3bc721c98eb4b324f
+  metadata.gz: e5dfedc8371c24fad4931ca94741d943a0e696d9
+  data.tar.gz: ef7e1108e5f1f1695f235670877f417b51455cbe
 SHA512:
-  metadata.gz: 661e36f39e6389337bd70f0b12a618a646345353c271a96087894c3ed2b616520519d10d2c4072b8a139f2174a04ffc63b4ad7d7339f6c22b8730222fda104f0
-  data.tar.gz: c47fdb6fa20296507f24ff79076816fbca19b06a6c94b9b3b151c3cda6d571049e138a46fe38e6b845ed992a412977c29e2692d697919849c53372f1bb69cec3
+  metadata.gz: 4f9303126e3d7b7e0aa0c0f682defc33c2694206eb90ee92cde2d30aa186fdbe056a9e68d7ecfd82622e3419c68f35e3da003700e008ca48c1b4f4afc3bd6000
+  data.tar.gz: 29322368c00d734c17497c564d24588eec8a86a0603337642c345a0efde48cb0c72f72118a0611dd1b61d3e7fae4831a9a928f252e9c6d7b8ec5e0c33d10a450

data/.hound.yml

@@ -0,0 +1,2 @@
+ruby:
+  config_file: .ruby-style.yml

data/.ruby-style.yml

@@ -0,0 +1,13 @@
+Style/StringLiterals:
+  EnforcedStyle: single_quotes
+  SupportedStyles:
+    - single_quotes
+    - double_quotes
+
+Metrics/LineLength:
+  Max: 100
+  AllowHeredoc: true
+  AllowURI: true
+  URISchemes:
+    - http
+    - https

data/README.md

@@ -99,7 +99,12 @@ associated sugar in the `transform` method of `Pipeline` make this very easy (se
 
 An `Unfold` acts as a source, however, takes a source as an input and produces multiple rows for that source as an output.
 A common case for this is splitting rows into multiple rows depending upon their keys. The `BlockTransform` and associated
-sugar in the `unfold` method of `Pipeline` make this fiarly easy (see `pipeline_spec.rb`).
+sugar in the `unfold` method of `Pipeline` make this fairly easy (see `pipeline_spec.rb`).
+
+#### Observers
+
+An `Observer` acts as a source and takes a source as an input and serves as a pass-through for a stream. Observers
+are used to observe the data stream--record data for use elsewhere.
 
 #### Sinks
 

data/lib/arc-furnace/multi_excel_source.rb

@@ -0,0 +1,85 @@
+require 'arc-furnace/source'
+require 'roo'
+
+module ArcFurnace
+  class MultiExcelSource < Source
+
+    private_attr_reader :enumerator, :header_row
+    attr_reader :value, :excel, :sheets_info_array
+    
+    # Sheets is in the format of:
+    # [
+    #   { filename: 'foo.xlsx', sheet: 'sheet name' },
+    #   { filename: 'foo2.xlsx', sheet: 'sheet name' }
+    # ]
+    #
+    # The value for the :sheet key points to the sheet that we want to parse.
+    # If sheets are not explicitly indicated, they will not be parsed.
+
+    def initialize(sheets_info_array: [])
+      @sheets_info_array = sheets_info_array.reverse
+      open_next_file
+    end
+
+    private
+
+    def preprocess
+      enumerator.next
+    end
+
+    def advance
+      advance_in_current_file || open_next_file
+    end
+
+    def advance_in_current_file
+      @value =
+          begin
+            enumerator.next if enumerator
+          rescue StopIteration
+            @enumerator = nil
+            nil
+          end
+      value
+    end
+
+    def open_next_file
+      excel.close if excel
+      @excel = nil
+      @header_row = nil
+      if sheets_info_array.empty?
+        nil
+      else
+        sheets_info = sheets_info_array.pop
+        @excel = Roo::Excelx.new(sheets_info[:filename])
+        @excel.default_sheet = sheets_info[:sheet]
+        @enumerator = build_enumerator
+        preprocess
+        advance
+      end
+    end
+
+    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|
+        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
+  end
+end

data/lib/arc-furnace/version.rb

@@ -1,3 +1,3 @@
 module ArcFurnace
-  VERSION = "0.1.14"
+  VERSION = "0.1.15"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: arc-furnace
 version: !ruby/object:Gem::Version
-  version: 0.1.14
+  version: 0.1.15
 platform: ruby
 authors:
 - Daniel Spangenberger
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2016-01-25 00:00:00.000000000 Z
+date: 2016-02-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: msgpack
@@ -118,7 +118,9 @@ extensions: []
 extra_rdoc_files: []
 files:
 - ".gitignore"
+- ".hound.yml"
 - ".rspec"
+- ".ruby-style.yml"
 - ".travis.yml"
 - Gemfile
 - LICENSE.txt
@@ -147,6 +149,7 @@ files:
 - lib/arc-furnace/logging_error_handler.rb
 - lib/arc-furnace/merging_hash.rb
 - lib/arc-furnace/multi_csv_source.rb
+- lib/arc-furnace/multi_excel_source.rb
 - lib/arc-furnace/node.rb
 - lib/arc-furnace/nodes.rb
 - lib/arc-furnace/null_sink.rb