libhxl-ruby 0.0.1

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: 719cfb05f414be3c3f796a5549d30d68ed4e9a03
+  data.tar.gz: 75f03298ddd344a4cca609a3e3db1897da1ad35e
+SHA512:
+  metadata.gz: 29f19279e9d8f3d60b2d1d3e536193db92d70ab8936b7b866d9cfd62853548830031dfbb37e95fdd333e5c18c4402afc5097b123e629002b13a4be4e4358820e
+  data.tar.gz: d9db4b7a208fa9bf3c26d650ea9fd250d6f1b4197682648f427069cd8f13921a9dbf9a238b762f776cd33d68fc2a5862110789e3c099e92e21f8b17e88a3c17e

data/lib/libhxl-ruby.rb

@@ -0,0 +1,193 @@
+class HXLReader
+
+  def initialize
+
+  end
+
+  def self.foreach(path, &block)
+
+    table_spec = nil
+    prev_row = nil
+
+    source_row_number = -1
+    row_number = -1
+    disaggregation_position = 0
+
+    CSV.foreach(path) do |row|
+
+      source_row_number += 1
+
+      # If we don't have a table_spec yet (row of HXL tags), scan for one
+      if table_spec.nil?
+
+        table_spec = self.parse_hashtag_row(row, prev_row)
+
+        next if table_spec
+
+      end
+      prev_row = row
+      next if table_spec.nil?
+
+
+
+      disaggregation_position = 0
+
+      loop do
+        # Next logical row
+        row_number += 1
+
+        hxl_row, disaggregation_position = parse_row(row,
+                                                     table_spec,
+                                                     disaggregation_position,
+                                                     row_number,
+                                                     source_row_number)
+        yield hxl_row
+
+        break unless disaggregation_position < table_spec.get_disaggregation_count
+      end
+
+
+    end
+
+    raise HXLFormatError.new('HXL hashtag row not found') if table_spec.nil?
+  end
+
+  def self.parse_row(row, table_spec, disaggregation_position, row_number, source_row_number)
+
+    hxl_fields = []
+    col_num = -1
+
+    seen_fixed = false
+    row.each_with_index do |value, source_col_number|
+
+      col_spec = table_spec.col_specs[source_col_number]
+
+      # Only parse HXL columns
+      next if col_spec.column.hxl_tag.nil?
+
+      if col_spec.fixed_column
+        # Looking at disaggregation
+
+        if !seen_fixed
+          col_num += 1
+          raw_position = table_spec.get_raw_position(disaggregation_position)
+
+          hxl_fields.push HXLValue.new(table_spec.col_specs[raw_position].fixed_column,
+                                table_spec.col_specs[raw_position].fixed_value,
+                                col_num,
+                                source_col_number)
+
+          col_num += 1
+          hxl_fields.push HXLValue.new(table_spec.col_specs[raw_position].column,
+                                row[raw_position],
+                                col_num,
+                                source_col_number)
+
+          seen_fixed = true
+          disaggregation_position += 1
+
+        end
+      else
+        # Regular column
+        col_num += 1
+        hxl_fields.push HXLValue.new(table_spec.col_specs[source_col_number].column,
+                              value,
+                              col_num,
+                              source_col_number)
+
+
+      end
+    end
+
+    hxl_row = HXLRow.new(
+      table_spec.hxl_headers,
+      hxl_fields,
+      false,
+      row_number,
+      source_row_number)
+
+    return hxl_row, disaggregation_position
+  end
+
+  def self.parse_hashtag_row(row, prev_row)
+
+    # Try parsing the current raw CSV data row as a HXL hashtag row.
+    # Returns a HXLTableSpec on success, or None on failure
+
+    seen_header = false
+    table_spec = HXLTableSpec.new
+
+    row.each_with_index do |value, col_num|
+      value = value.strip if value
+      col_spec = nil
+
+      if !value.nil? && !value.empty?
+        col_spec = self.parse_hashtag(col_num, value)
+        return nil if col_spec.nil?
+
+        seen_header = true
+
+        if col_spec.fixed_column
+          col_spec.fixed_value = prev_row[col_num]
+        end
+
+      else
+        col_spec = HXLColSpec.new col_num, HXLColumn.new
+      end
+
+      table_spec.push col_spec
+    end
+
+    return table_spec if seen_header
+
+    nil
+  end
+
+  def self.parse_hashtag(source_col_number, value)
+
+    # Pattern for a single tag
+    tag_regex = /(#[\w]+)(?:\/([[:alpha:]]{2}))?/
+
+    # Pattern for full tag spec (optional second tag following '+')
+    full_regex = /^\s*#{tag_regex}(?:\s*\+\s*#{tag_regex})?$/
+
+    result = full_regex.match value
+    col_spec = nil
+
+    if result
+      col1 = nil
+      col2 = nil
+
+      if result[3]
+        # There were two tags
+        col1 = HXLColumn.new result[1], result[2]
+        col2 = HXLColumn.new result[3], result[4]
+        col_spec = HXLColSpec.new source_col_number, col2, col1
+
+      else
+        col1 = HXLColumn.new result[1], result[2]
+        col_spec = HXLColSpec.new source_col_number, col1
+      end
+    end
+
+    col_spec
+  end
+
+  def self.parse_table_spec(row, prev_row)
+    # Search for the HXL hashtag row
+    # Returns a HXLTableSpec on success. Throws an exception on failure.
+
+    raw = self.parse_source_row
+  end
+
+  def self.parse_source_row
+  end
+
+end
+
+require 'libhxl-ruby/hxl_format_error'
+require 'libhxl-ruby/hxl_table_spec'
+require 'libhxl-ruby/hxl_row'
+require 'libhxl-ruby/hxl_value'
+require 'libhxl-ruby/hxl_column'
+require 'libhxl-ruby/hxl_col_spec'

data/lib/libhxl-ruby/hxl_col_spec.rb

@@ -0,0 +1,19 @@
+class HXLReader::HXLColSpec
+
+  attr_reader :fixed_column, :source_col_number, :column
+  attr_accessor :fixed_value
+
+  # Column metadata for parsing a HXL CSV file
+  #
+  # This class captures the way a column is encoded in the input CSV
+  # file, which might be different from the logical structure of the
+  # HXL data. Used only during parsing.
+
+  def initialize(source_col_number, column = nil, fixed_column = nil, fixed_value = nil)
+    @source_col_number = source_col_number
+    @column = column
+    @fixed_column = fixed_column
+    @fixed_value = fixed_value
+  end
+
+end

data/lib/libhxl-ruby/hxl_column.rb

@@ -0,0 +1,20 @@
+class HXLReader::HXLColumn
+
+  attr_reader :hxl_tag, :language_code, :header_text
+
+  def initialize(hxl_tag = nil, language_code = nil)
+    @hxl_tag = hxl_tag
+    @language_code = language_code
+    @header_text = self.pretty_tag
+  end
+
+  def pretty_tag
+    return '' unless @hxl_tag
+
+    @hxl_tag
+      .gsub(/^#/, '')
+      .gsub(/_(date|deg|id|link|num)$/, '')
+      .gsub('_', ' ')
+      .upcase
+  end
+end

data/lib/libhxl-ruby/hxl_format_error.rb

@@ -0,0 +1,2 @@
+class HXLReader::HXLFormatError < StandardError
+end

data/lib/libhxl-ruby/hxl_row.rb

@@ -0,0 +1,27 @@
+require 'csv'
+
+class HXLReader::HXLRow < CSV::Row
+  # An iterable row of HXL value objects
+
+  attr_reader :row_number, :source_row_number
+
+
+  def initialize(headers, fields, header_row = false, row_number = nil, source_row_number = nil)
+    super headers, fields, header_row
+    @row_number = row_number
+    @source_row_number = source_row_number
+
+  end
+
+  def to_s
+    s = '<HXLRow';
+    s += "\n  rowNumber: " + @row_number.to_s
+    s += "\n  sourceRowNumber: " + @sourceRowNumber.to_s
+    s += "\n"
+    s += @row.map(&:to_s).join('\n ')
+    s += "\n>"
+
+    s
+  end
+
+end

data/lib/libhxl-ruby/hxl_table_spec.rb

@@ -0,0 +1,45 @@
+class HXLReader::HXLTableSpec
+
+  attr_reader :col_specs
+
+  # Table metadata for parsing a HXL dataset
+
+  def initialize
+    @col_specs = []
+  end
+
+  def push(col_spec)
+    @col_specs.push col_spec
+  end
+
+  def hxl_headers
+    headers = []
+    seen_fixed = false
+    @col_specs.each do |spec|
+      if spec.fixed_column && !seen_fixed
+        headers.push spec.fixed_column.hxl_tag
+        headers.push spec.column.hxl_tag unless spec.column.hxl_tag.nil?
+        seen_fixed = true
+      elsif !spec.fixed_column
+        headers.push spec.column.hxl_tag unless spec.column.hxl_tag.nil?
+      end
+    end
+
+    headers
+  end
+
+  def get_disaggregation_count
+    (@col_specs.select { |col_spec| col_spec.fixed_column }).length
+  end
+
+  def get_raw_position(disaggregation_position)
+    @col_specs.each_with_index do |col_spec, i|
+      disaggregation_position -= 1 if col_spec.fixed_column
+
+      return i if disaggregation_position < 0
+    end
+
+    return -1
+  end
+
+end

data/lib/libhxl-ruby/hxl_value.rb

@@ -0,0 +1,12 @@
+class HXLReader::HXLValue
+  # A single HXL value at the intersection of a row and column
+  attr_reader :column, :value, :col_num, :source_col_num
+
+  def initialize(column, value, col_num, source_col_num)
+    @column = column
+    @value = value
+    @col_num = col_num
+    @source_col_num = source_col_num
+  end
+
+end

metadata

@@ -0,0 +1,65 @@
+--- !ruby/object:Gem::Specification
+name: libhxl-ruby
+version: !ruby/object:Gem::Version
+  version: 0.0.1
+platform: ruby
+authors:
+- Ben Rudolph
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2014-10-19 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: rspec
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '3.1'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '3.1'
+description: A simple gem to parse your HXL files
+email: rudolphben@gmail.com
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- lib/libhxl-ruby.rb
+- lib/libhxl-ruby/hxl_col_spec.rb
+- lib/libhxl-ruby/hxl_column.rb
+- lib/libhxl-ruby/hxl_format_error.rb
+- lib/libhxl-ruby/hxl_row.rb
+- lib/libhxl-ruby/hxl_table_spec.rb
+- lib/libhxl-ruby/hxl_value.rb
+homepage: https://github.com/benrudolph/libhxl-ruby
+licenses:
+- MIT
+metadata: {}
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 2.4.2
+signing_key: 
+specification_version: 4
+summary: A HXL parser for ruby
+test_files: []
+has_rdoc: