infoboxer 0.2.7 → 0.2.8

data/lib/infoboxer/parser/paragraphs.rb

@@ -1,9 +1,10 @@
 # encoding: utf-8
+
 module Infoboxer
   class Parser
     module Paragraphs
       include Tree
-      
+
       def paragraphs(until_pattern = nil)
         nodes = Nodes[]
         until @context.eof?
@@ -18,46 +19,42 @@ module Infoboxer
 
       private
 
-        def paragraph(until_pattern)
-          case @context.current
-          when /^(?<level>={2,})\s*(?<text>.+?)\s*\k<level>$/
-            heading(Regexp.last_match[:text], Regexp.last_match[:level])
-          when /^\s*{\|/
-            table
-          when /^[\*\#:;]./
-            list(until_pattern)
-          when /^-{4,}/
-            HR.new
-          when /^\s*$/
-            # will, when merged, close previous paragraph or add spaces to <pre>
-            EmptyParagraph.new(@context.current)
-          when /^ (?!\s*{{)/ # Lookahead, because spaces before template are ignored
-            pre(until_pattern)
-          else
-            Paragraph.new(short_inline(until_pattern))
-          end
+      def paragraph(until_pattern)
+        case @context.current
+        when /^(?<level>={2,})\s*(?<text>.+?)\s*\k<level>$/
+          heading(Regexp.last_match[:text], Regexp.last_match[:level])
+        when /^\s*{\|/
+          table
+        when /^[\*\#:;]./
+          list(until_pattern)
+        when /^-{4,}/
+          HR.new
+        when /^\s*$/
+          # will, when merged, close previous paragraph or add spaces to <pre>
+          EmptyParagraph.new(@context.current)
+        when /^ (?!\s*{{)/ # Lookahead, because spaces before template are ignored
+          pre(until_pattern)
+        else
+          Paragraph.new(short_inline(until_pattern))
         end
+      end
 
-        def heading(text, level)
-          Heading.new(Parser.inline(text), level.length)
-        end
+      def heading(text, level)
+        Heading.new(Parser.inline(text), level.length)
+      end
 
-        # http://en.wikipedia.org/wiki/Help:List
-        def list(until_pattern)
-          marker = @context.scan(/^([*\#:;]+)\s*/).strip
-          List.construct(marker.chars.to_a, short_inline(until_pattern))
-        end
+      # http://en.wikipedia.org/wiki/Help:List
+      def list(until_pattern)
+        marker = @context.scan(/^([*\#:;]+)\s*/).strip
+        List.construct(marker.chars.to_a, short_inline(until_pattern))
+      end
 
-        # FIXME: in fact, there's some formatting, that should work inside pre
-        def pre(until_pattern)
-          @context.skip(/^ /)
-          str = if until_pattern
-            @context.scan_until(/(#{until_pattern}|$)/)
-          else
-            @context.current
-          end
-          Pre.new(Nodes[Text.new(str)])
-        end
+      # FIXME: in fact, there's some formatting, that should work inside pre
+      def pre(until_pattern)
+        @context.skip(/^ /)
+        str = until_pattern ? @context.scan_until(/(#{until_pattern}|$)/) : @context.current
+        Pre.new(Nodes[Text.new(str)])
+      end
 
       require_relative 'table'
       include Parser::Table

data/lib/infoboxer/parser/table.rb

@@ -1,15 +1,19 @@
 # encoding: utf-8
+
 module Infoboxer
   class Parser
     # http://en.wikipedia.org/wiki/Help:Table
     module Table
       include Tree
-      
+
       def table
-         @context.current =~ /^\s*{\|/ or
+        @context.current =~ /^\s*{\|/ or
           @context.fail!('Something went wrong: trying to parse not a table')
 
+        log 'Starting to parse table'
+
         prms = table_params
+        log "Table params found #{prms}"
         table = Tree::Table.new(Nodes[], prms)
 
         @context.next!
@@ -20,7 +24,7 @@ module Infoboxer
         end
 
         # FIXME: not the most elegant way, huh?
-        table.children.reject!{|r| r.children.empty?}
+        table.children.reject! { |r| r.children.empty? }
 
         table
       end
@@ -34,55 +38,53 @@ module Infoboxer
         case @context.current
         when /^\s*\|}(.*)$/                 # table end
           @context.scan(/^\s*\|}/)
-          return false # should not continue
-
+          return false
         when /^\s*!/                        # heading (th) in a row
           table_cells(table, TableHeading)
-
         when /^\s*\|\+/                     # caption
           table_caption(table)
-          
         when /^\s*\|-(.*)$/                 # row start
-          table_row(table, $1)
+          table_row(table, Regexp.last_match(1))
 
         when /^\s*\|/                       # cell in row
           table_cells(table)
-
         when /^\s*{{/                       # template can be at row level
           table_template(table)
-
         when nil
           return false
-
         else
           return table_cell_cont(table)
         end
+
         true # should continue parsing
       end
 
       def table_row(table, param_str)
+        log 'Table row found'
         table.push_children(TableRow.new(Nodes[], parse_params(param_str)))
       end
 
       def table_caption(table)
+        log 'Table caption found'
         @context.skip(/^\s*\|\+\s*/)
 
         children = inline(/^\s*([|!]|{\|)/)
-        @context.prev! # compensate next! which will be done in table()
+        @context.prev! if @context.eol? # compensate next! which will be done in table()
         table.push_children(TableCaption.new(children.strip))
       end
 
       def table_cells(table, cell_class = TableCell)
-        table.push_children(TableRow.new()) unless table.children.last.is_a?(TableRow)
+        log 'Table cells found'
+        table.push_children(TableRow.new) unless table.children.last.is_a?(TableRow)
         row = table.children.last
 
         @context.skip(/\s*[!|]\s*/)
         guarded_loop do
-          if @context.check(/[^|{|\[]+\|([^\|]|$)/)
-            params = parse_params(@context.scan_until(/\|/))
-          else
-            params = {}
-          end
+          params = if @context.check(/[^|{|\[]+\|([^\|]|$)/)
+                     parse_params(@context.scan_until(/\|/))
+                   else
+                     {}
+                   end
           content = short_inline(/(\|\||!!)/)
           row.push_children(cell_class.new(content, params))
           break if @context.eol?
@@ -91,7 +93,7 @@ module Infoboxer
 
       def table_template(table)
         contents = paragraph(/^\s*([|!]|{\|)/).to_templates?
-        
+
         if (row = table.children.last).is_a?(TableRow)
           if (cell = row.children.last).is_a?(BaseCell)
             cell.push_children(*contents)
@@ -117,16 +119,13 @@ module Infoboxer
                       nil
                     end
 
-        if !container
+        unless container
           # return "table not continued" unless row is empty
-          if @context.current.empty?
-            return true
-          else
-            @context.prev!
-            return false
-          end
+          return true if @context.current.empty?
+          @context.prev!
+          return false
         end
-        
+
         container.push_children(paragraph(/^\s*([|!]|{\|)/))
         table.push_children(container) unless container.parent
         true

data/lib/infoboxer/parser/template.rb

@@ -1,4 +1,5 @@
 # encoding: utf-8
+
 module Infoboxer
   class Parser
     module Template
@@ -11,7 +12,9 @@ module Infoboxer
       # More about magic words: https://www.mediawiki.org/wiki/Help:Magic_words
       def template
         name = @context.scan_continued_until(/\||:|}}/) or
-          @context.fail!("Template name not found")
+          @context.fail!('Template name not found')
+
+        log "Parsing template #{name}"
 
         name.strip!
         vars = @context.eat_matched?('}}') ? Nodes[] : template_vars
@@ -19,6 +22,8 @@ module Infoboxer
       end
 
       def template_vars
+        log 'Parsing template variables'
+
         num = 1
         res = Nodes[]
 
@@ -31,11 +36,14 @@ module Infoboxer
             name = num
             num += 1
           end
+          log "Variable #{name} found"
 
           value = long_inline(/\||}}/)
-          unless value.empty? && name.is_a?(Numeric) # it was just empty line otherwise
-            res << Var.new(name.to_s, value)
-          end
+
+          # it was just empty line otherwise
+          res << Var.new(name.to_s, value) unless value.empty? && name.is_a?(Numeric)
+
+          log 'Variable value found'
 
           break if @context.eat_matched?('}}')
           @context.eof? and @context.fail!("Unexpected break of template variables: #{res}")

data/lib/infoboxer/parser/util.rb

@@ -1,19 +1,20 @@
 # encoding: utf-8
+
 module Infoboxer
   class Parser
     module Util
       attr_reader :re
 
-      FORMATTING = /(
+      FORMATTING = %r((
         '''''|'''|''  |     # bold, italic, bold italic
         \[\[          |     # link
         {{            |     # template
-        \[[a-z]+:\/\/ |     # external link
+        \[[a-z]+://   |     # external link
         <nowiki[^>]*> |     # nowiki
         <ref[^>]*>    |     # reference
         <math>        |     # math
         <                   # HTML tag
-      )/x
+      ))x
 
       INLINE_EOL = %r[(?=   # if we have ahead... (not scanned, just checked
         </ref>        |     # <ref> closed
@@ -24,7 +25,8 @@ module Infoboxer
         </ref>        |     # <ref> closed
         }}            |     # or template closed
         (?<!\])\](?!\])     # or ext.link closed,
-                            # the madness with look-ahead/behind means "match single bracket but not double"
+                            # the madness with look-ahead/behind means
+                            # "match single bracket but not double"
       )]x
 
       # FIXME: ok, NOW it's officially ridiculous
@@ -34,24 +36,22 @@ module Infoboxer
         \]\]                # or int.link closed
       )]x
 
-
       def make_regexps
         {
           file_namespace: /(#{@context.traits.file_namespace.join('|')}):/,
           formatting: FORMATTING,
-          inline_until_cache: Hash.new{|h, r|
+          inline_until_cache: Hash.new { |h, r|
             h[r] = Regexp.union(*[r, FORMATTING, /$/].compact.uniq)
           },
-          short_inline_until_cache: Hash.new{|h, r|
+          short_inline_until_cache: Hash.new { |h, r|
             h[r] = Regexp.union(*[r, INLINE_EOL, FORMATTING, /$/].compact.uniq)
           },
-          short_inline_until_cache_brackets: Hash.new{|h, r|
+          short_inline_until_cache_brackets: Hash.new { |h, r|
             h[r] = Regexp.union(*[r, INLINE_EOL_BRACK, FORMATTING, /$/].compact.uniq)
           },
-          short_inline_until_cache_brackets2: Hash.new{|h, r|
+          short_inline_until_cache_brackets2: Hash.new { |h, r|
             h[r] = Regexp.union(*[r, INLINE_EOL_BRACK2, FORMATTING, /$/].compact.uniq)
           }
-
         }
       end
 
@@ -67,11 +67,7 @@ module Infoboxer
           if scan.peek(1) == '='
             scan.skip(/=\s*/)
             q = scan.scan(/['"]/)
-            if q
-              value = scan.scan_until(/#{q}|$/).sub(q, '')
-            else
-              value = scan.scan_until(/\s|$/)
-            end
+            value = q ? scan.scan_until(/#{q}|$/).sub(q, '') : scan.scan_until(/\s|$/)
             params[name.to_sym] = value
           else
             params[name.to_sym] = name
@@ -89,7 +85,6 @@ module Infoboxer
             @context.fail!("Infinite loop on position #{pos_after.last}")
         end
       end
-
     end
   end
 end

data/lib/infoboxer/parser.rb

@@ -1,6 +1,8 @@
 # encoding: utf-8
+
 require 'ostruct'
 require 'procme'
+require 'logger'
 
 module Infoboxer
   class Parser
@@ -47,12 +49,13 @@ module Infoboxer
         end
       end
     end
-    
+
     include Tree
 
     def initialize(context)
       @context = context
       @re = OpenStruct.new(make_regexps)
+      @logger = Logger.new(STDOUT).tap { |l| l.level = Logger::FATAL }
     end
 
     require_relative 'parser/inline'
@@ -65,6 +68,10 @@ module Infoboxer
 
     require_relative 'parser/util'
     include Parser::Util
+
+    def log(msg)
+      @logger.info "#{msg} | #{@context.lineno}:#{@context.colno}: #{@context.current}"
+    end
   end
 end
 

data/lib/infoboxer/templates/base.rb

@@ -2,7 +2,7 @@ module Infoboxer
   module Templates
     class Base < Tree::Template
       include Tree
-      
+
       class << self
         attr_accessor :template_name, :template_options
 
@@ -16,7 +16,7 @@ module Infoboxer
       end
 
       def ==(other)
-        other.kind_of?(Tree::Template) && _eq(other)
+        other.is_a?(Tree::Template) && _eq(other)
       end
 
       protected
@@ -49,7 +49,7 @@ module Infoboxer
     # Used for {Set} definitions.
     class Replace < Base
       def replace
-        fail(NotImplementedError, "Descendants should define :replace")
+        fail(NotImplementedError, 'Descendants should define :replace')
       end
 
       def text

data/lib/infoboxer/templates/set.rb

@@ -1,4 +1,5 @@
 # encoding: utf-8
+
 module Infoboxer
   module Templates
     # Base class for defining set of templates, used for some site/domain.
@@ -20,7 +21,7 @@ module Infoboxer
 
       # @private
       def find(name)
-        _, template = @templates.detect{|m, t| m === name.downcase}
+        _, template = @templates.detect { |m, _t| m === name.downcase }
         template || Base
       end
 
@@ -43,15 +44,15 @@ module Infoboxer
       #   def from
       #     fetch_date('1', '2', '3')
       #   end
-      # 
+      #
       #   def to
       #     fetch_date('4', '5', '6') || Date.today
       #   end
-      # 
+      #
       #   def value
       #     (to - from).to_i / 365 # FIXME: obviously
       #   end
-      # 
+      #
       #   def text
       #     "#{value} years"
       #   end
@@ -100,7 +101,7 @@ module Infoboxer
       # Expected to be used inside Set definition block.
       def replace(*replacements)
         case
-        when replacements.count == 2 && replacements.all?{|r| r.is_a?(String)}
+        when replacements.count == 2 && replacements.all? { |r| r.is_a?(String) }
           name, what = *replacements
           setup_class(name, Replace) do
             define_method(:replace) do
@@ -108,8 +109,8 @@ module Infoboxer
             end
           end
         when replacements.count == 1 && replacements.first.is_a?(Hash)
-          replacements.first.each do |name, what|
-            replace(name, what)
+          replacements.first.each do |nm, rep|
+            replace(nm, rep)
           end
         else
           fail(ArgumentError, "Can't call :replace with #{replacements.join(', ')}")
@@ -163,13 +164,13 @@ module Infoboxer
       def setup_class(name, base_class, options = {}, &definition)
         match = options.fetch(:match, name.downcase)
         base = options.fetch(:base, base_class)
-        base = self.find(base) if base.is_a?(String)
+        base = find(base) if base.is_a?(String)
 
-        Class.new(base, &definition).tap{|cls|
+        Class.new(base, &definition).tap do |cls|
           cls.template_name = name
           cls.template_options = options
           @templates.unshift [match, cls]
-        }
+        end
       end
     end
   end

data/lib/infoboxer/tree/compound.rb

@@ -1,4 +1,5 @@
 # encoding: utf-8
+
 module Infoboxer
   module Tree
     # Base class for all nodes with children.
@@ -6,7 +7,7 @@ module Infoboxer
       def initialize(children = Nodes.new, params = {})
         super(params)
         @children = Nodes[*children]
-        @children.each{|c| c.parent = self}
+        @children.each { |c| c.parent = self }
       end
 
       # List of children
@@ -24,7 +25,7 @@ module Infoboxer
       # @private
       # Internal, used by {Parser}
       def push_children(*nodes)
-        nodes.each{|c| c.parent = self}.each do |n|
+        nodes.each { |c| c.parent = self }.each do |n|
           @children << n
         end
       end
@@ -40,15 +41,15 @@ module Infoboxer
           "#{indent(level)}#{children.first.text} <#{descr}>\n"
         else
           "#{indent(level)}<#{descr}>\n" +
-            children.map(&call(to_tree: level+1)).join
+            children.map(&call(to_tree: level + 1)).join
         end
       end
 
       # Kinda "private" methods, used by Parser only -------------------
-      
+
       # @private
       # Internal, used by {Parser}
-      def can_merge?(other)
+      def can_merge?(_other)
         false
       end
 
@@ -80,7 +81,7 @@ module Infoboxer
 
       def _eq(other)
         children == other.children
-      end      
+      end
     end
   end
 end

data/lib/infoboxer/tree/document.rb

@@ -1,4 +1,5 @@
 # encoding: utf-8
+
 module Infoboxer
   module Tree
     # Represents entire document.

data/lib/infoboxer/tree/html.rb

@@ -1,8 +1,9 @@
 # encoding: utf-8
+
 module Infoboxer
   module Tree
     module HTMLTagCommons
-      BLOCK_TAGS = %w[div p br] # FIXME: are some other used in WP?
+      BLOCK_TAGS = %w[div p br].freeze # FIXME: are some other used in WP?
 
       def text
         super + (BLOCK_TAGS.include?(tag) ? "\n" : '')
@@ -27,7 +28,7 @@ module Infoboxer
         # even empty tag, for ex., <br>, should not be dropped!
         false
       end
-      
+
       private
 
       def descr
@@ -47,12 +48,12 @@ module Infoboxer
         super(attrs)
         @tag = tag
       end
-      
+
       attr_reader :tag
       alias_method :attrs, :params
 
       include HTMLTagCommons
-      
+
       private
 
       def descr

data/lib/infoboxer/tree/image.rb

@@ -1,4 +1,5 @@
 # encoding: utf-8
+
 module Infoboxer
   module Tree
     # Represents image (or other media file).
@@ -16,16 +17,16 @@ module Infoboxer
       # @return [Nodes]
       attr_reader :caption
 
-      # @!attribute [r] path 
+      # @!attribute [r] path
       # @!attribute [r] type
-      # @!attribute [r] location 
+      # @!attribute [r] location
       # @!attribute [r] alignment
       # @!attribute [r] link
-      # @!attribute [r] alt 
+      # @!attribute [r] alt
 
       def_readers :path, :type,
-        :location, :alignment, :link,
-        :alt
+                  :location, :alignment, :link,
+                  :alt
 
       def border?
         !params[:border].to_s.empty?
@@ -42,8 +43,8 @@ module Infoboxer
       def to_tree(level = 0)
         super(level) +
           if caption && !caption.empty?
-            indent(level+1) + "caption:\n" +
-              caption.children.map(&call(to_tree: level+2)).join
+            indent(level + 1) + "caption:\n" +
+            caption.children.map(&call(to_tree: level + 2)).join
           else
             ''
           end

data/lib/infoboxer/tree/inline.rb

@@ -1,4 +1,5 @@
 # encoding: utf-8
+
 module Infoboxer
   module Tree
     # Represents italic text.
@@ -20,17 +21,15 @@ module Infoboxer
         super(label || Nodes.new([Text.new(link)]), link: link)
       end
 
-      #@!attribute [r] link
+      # @!attribute [r] link
 
       def_readers :link
     end
 
     # External link. Has other nodes as a contents, and, err, link (url).
     class ExternalLink < Link
-    
-      #@!attribute [r] url
-      #  synonym for `#link`
-      
+      # @!attribute [r] url
+      #   synonym for `#link`
       alias_method :url, :link
     end
   end

data/lib/infoboxer/tree/linkable.rb

@@ -23,20 +23,18 @@ module Infoboxer
       # @return [String]
       def url
         # FIXME: fragile as hell.
-        page.url.sub(/[^\/]+$/, link.gsub(' ', '_'))
+        page.url.sub(%r{[^/]+$}, link.tr(' ', '_'))
       end
 
       protected
 
       def page
-        page = lookup_parents(MediaWiki::Page).first or
-          fail("Not in a page from real source")
+        lookup_parents(MediaWiki::Page).first or fail('Not in a page from real source')
       end
 
       def client
-        page.client or fail("MediaWiki client not set")
+        page.client or fail('MediaWiki client not set')
       end
-
     end
   end
 end