infoboxer 0.2.1 → 0.2.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 3d9f535874913b90dc945c37405ed54aff97a53a
-  data.tar.gz: a25dc355939431517d55ffe1d0aa0cbcbef0fff7
+  metadata.gz: 69ee012f6cd8bb3e923289a4a921e81798ffd2aa
+  data.tar.gz: 42686bced509c310051ba7bc895c68d26a96bd2f
 SHA512:
-  metadata.gz: b5fc7579e52f108b56de01e9ca6bcd3b8c54eaec08a4003d15553f2b91318052c6b6f2d3bfac76ff7a913651d9e73925e7c9e87e2ca0c36df26a36107789e735
-  data.tar.gz: 7692f32da19229dbb5e03e65f8be4db6e4cd605ba5c8d9e2f95200734a288ba8ff8ccb1f518bd365a91c82ead2c1639c495f81538ad983d1e9c9f8f4434e9b2f
+  metadata.gz: cc237b46fe0f59a2ba04d3b53da4a7361cd974e50e1d18f688a077093da664a7d205dd363ddc2622a9af4c1a6f1168315d90fb6867e65c0a61bd60e5e453e81a
+  data.tar.gz: d844446421cffd8e7f6a8fe885f0f1918c772f9e2e7c1d289a1e90b544fde9b3c2b41c8956df7338aab9ad2ae5ca7f5ec9d67fb2d3f16da34367ed5b96a5aa22

data/CHANGELOG.md

@@ -1,5 +1,13 @@
 # Infoboxer's change log
 
+## 0.2.2 (2016-01-03)
+
+Fixes:
+* more sophisticated table parsing;
+* empty `<nowiki/>` is parsed properly;
+* inline unclosed markup inside wikilinks works;
+* `MediaWiki::Traits` can now be continued in several places.
+
 ## 0.2.1 (2015-12-21)
 
 * `infoboxer` binary properly registered.

data/README.md

@@ -4,7 +4,7 @@
 [![Build Status](https://travis-ci.org/molybdenum-99/infoboxer.svg?branch=master)](https://travis-ci.org/molybdenum-99/infoboxer)
 [![Coverage Status](https://coveralls.io/repos/molybdenum-99/infoboxer/badge.svg?branch=master&service=github)](https://coveralls.io/github/molybdenum-99/infoboxer?branch=master)
 [![Code Climate](https://codeclimate.com/github/molybdenum-99/infoboxer/badges/gpa.svg)](https://codeclimate.com/github/molybdenum-99/infoboxer)
-[![Molybdenum-99 Gitter](https://badges.gitter.im/molybdenum-99.png)](https://gitter.im/molybdenum-99)
+[![Infoboxer Gitter](https://badges.gitter.im/molybdenum-99/infoboxer.svg)](https://gitter.im/molybdenum-99/infoboxer)
 
 **Infoboxer** is pure-Ruby Wikipedia (and generic MediaWiki) client and
 parser, targeting information extraction (hence the name).

data/lib/infoboxer/media_wiki/traits.rb

@@ -59,7 +59,8 @@ module Infoboxer
         # [English Wikipedia traits](https://github.com/molybdenum-99/infoboxer/blob/master/lib/infoboxer/definitions/en.wikipedia.org.rb)
         # for example implementation.
         def for(domain, &block)
-          Class.new(self, &block).domain(domain)
+          Traits.domains[domain].tap{|c| c && c.instance_eval(&block)} ||
+            Class.new(self, &block).domain(domain)
         end
 
         # @private

data/lib/infoboxer/parser/inline.rb

@@ -33,8 +33,10 @@ module Infoboxer
         nodes = Nodes[]
         guarded_loop do
           # FIXME: quick and UGLY IS HELL JUST TRYING TO MAKE THE SHIT WORK
-          if @context.inline_eol_sign
+          if @context.inline_eol_sign == /^\]/
             chunk = @context.scan_until(re.short_inline_until_cache_brackets[until_pattern])
+          elsif @context.inline_eol_sign == /^\]\]/
+            chunk = @context.scan_until(re.short_inline_until_cache_brackets2[until_pattern])
           else
             chunk = @context.scan_until(re.short_inline_until_cache[until_pattern])
           end
@@ -97,7 +99,7 @@ module Infoboxer
           when '{{'
             template
           when /<nowiki([^>]*)>/
-            nowiki
+            nowiki($1)
           when /<ref([^>]*)\/>/
             reference($1, true)
           when /<ref([^>]*)>/
@@ -114,7 +116,12 @@ module Infoboxer
         # [[a|b]]
         def wikilink
           link = @context.scan_continued_until(/\||\]\]/)
-          caption = inline(/\]\]/) if @context.matched == '|'
+          if @context.matched == '|'
+            @context.push_eol_sign(/^\]\]/)
+            caption = inline(/\]\]/)
+            @context.pop_eol_sign
+          end
+
           Wikilink.new(link, caption)
         end
 
@@ -136,8 +143,12 @@ module Infoboxer
           Ref.new(children, parse_params(param_str))
         end
 
-        def nowiki
-          Text.new(@context.scan_continued_until(/<\/nowiki>/))
+        def nowiki(tag_rest)
+          if tag_rest.end_with?('/')
+            Text.new('')
+          else
+            Text.new(@context.scan_continued_until(/<\/nowiki>/))
+          end
         end
       end
 

data/lib/infoboxer/parser/table.rb

@@ -52,10 +52,10 @@ module Infoboxer
           table_template(table)
 
         when nil
-          @context.fail!("End of input before table ended!")
+          return false
 
         else
-          table_cell_cont(table)
+          return table_cell_cont(table)
         end
         true # should continue parsing
       end
@@ -103,29 +103,33 @@ module Infoboxer
         end
       end
 
-      # On-the-fly TableCaption creation handles (real life) case, when
-      # table has "HTML caption":
-      #   {|
-      #   <caption>....</caption>
-      #
-      # Solution is NOT elegant or semantically "right", yet it works.
-      # Somehow.
+      # Good news, everyone! Table can be IMPLICITLY closed when it's
+      # not "cell" context.
       #
+      # Unless it's empty row, which is just skipped.
       def table_cell_cont(table)
         container = case (last = table.children.last)
-        when TableRow
-          cell = last.children.last
-          cell.is_a?(BaseCell) ? cell : TableCaption.new
-        when TableCaption
-          last
-        when nil
-          TableCaption.new
-        else
-          @context.fail!("Multiline cell inside #{last}")
+                    when TableRow
+                      last.children.last
+                    when TableCaption
+                      last
+                    else
+                      nil
+                    end
+
+        if !container
+          # return "table not continued" unless row is empty
+          if @context.current.empty?
+            return true
+          else
+            @context.prev!
+            return false
+          end
         end
-
+        
         container.push_children(paragraph(/^\s*([|!]|{\|)/))
         table.push_children(container) unless container.parent
+        true
       end
     end
   end

data/lib/infoboxer/parser/util.rb

@@ -19,13 +19,20 @@ module Infoboxer
         }}            
       )]x
 
-      INLINE_EOL_BR = %r[(?=   # if we have ahead... (not scanned, just checked
+      INLINE_EOL_BRACK = %r[(?=   # if we have ahead... (not scanned, just checked
         </ref>        |     # <ref> closed
         }}            |     # or template closed
         (?<!\])\](?!\])     # or ext.link closed,
                             # the madness with look-ahead/behind means "match single bracket but not double"
       )]x
 
+      # FIXME: ok, NOW it's officially ridiculous
+      INLINE_EOL_BRACK2 = %r[(?=   # if we have ahead... (not scanned, just checked
+        </ref>        |     # <ref> closed
+        }}            |     # or template closed
+        \]\]                # or int.link closed
+      )]x
+
 
       def make_regexps
         {
@@ -38,7 +45,10 @@ module Infoboxer
             h[r] = Regexp.union(*[r, INLINE_EOL, FORMATTING, /$/].compact.uniq)
           },
           short_inline_until_cache_brackets: Hash.new{|h, r|
-            h[r] = Regexp.union(*[r, INLINE_EOL_BR, FORMATTING, /$/].compact.uniq)
+            h[r] = Regexp.union(*[r, INLINE_EOL_BRACK, FORMATTING, /$/].compact.uniq)
+          },
+          short_inline_until_cache_brackets2: Hash.new{|h, r|
+            h[r] = Regexp.union(*[r, INLINE_EOL_BRACK2, FORMATTING, /$/].compact.uniq)
           }
           
         }

data/lib/infoboxer/tree/table.rb

@@ -25,13 +25,13 @@ module Infoboxer
       #
       # FIXME: it can easily be several table heading rows
       def heading_row
-        rows.first.children.all?(&call(matches?: TableHeading)) ?
+        rows.first && rows.first.children.all?(&call(matches?: TableHeading)) ?
           rows.first : nil
       end
 
       # For now, returns all table rows except {#heading_row}
       def body_rows
-        rows.first.children.all?(&call(matches?: TableHeading)) ?
+        rows.first && rows.first.children.all?(&call(matches?: TableHeading)) ?
           rows[1..-1] :
           rows
       end

data/lib/infoboxer/version.rb

@@ -2,6 +2,6 @@
 module Infoboxer
   MAJOR = 0
   MINOR = 2
-  PATCH = 1
+  PATCH = 2
   VERSION = [MAJOR, MINOR, PATCH].join('.')
 end