swordfish 0.0.4 → 0.0.5

checksums.yaml

@@ -1,15 +1,15 @@
 ---
 !binary "U0hBMQ==":
   metadata.gz: !binary |-
-    ZTEyODI1ZjU2OTMyODY3ZDM0NjM3OGY3ZjcyNTU1YTIwMjk1NDQ0ZA==
+    NDJiOGRjNGUyYjg3OWRjZWYwNTJhNWI5YTQ0NTgyMWUxMTdlMWJiOQ==
   data.tar.gz: !binary |-
-    YTIxMTkzNWMzNmYwNTY3YTNmOTg2YzE1MTYyZDUxYjk3YTNkZGU1MQ==
+    ZGJkYTgxMDUzZGRiZTU3M2I1MTEwMjJhNGE1NzZmMzlhNGNhZTFlOQ==
 SHA512:
   metadata.gz: !binary |-
-    ZmQ2YTk3YTcxNzFmNjNhMTg4Yjk0OGE1ZmI0ZjVmMWM5M2U1NmM1NDA3Y2Y0
-    MzUzNjI4NGQ1Y2IwOTgyNGI3MmM3M2M2OGM3OGIwMzAyY2VmZjE3NWMyNmJl
-    N2M0MjQ1OGQzZDZhMjI1M2RmN2Y4MTE0MjYyOGZjZWQ4MzJiZWY=
+    YmY4YzdlODI5NGUwYzFjOTliZmJmZmRlZjc4Njc4YzlhMmQ3MGZiYjA5NGZj
+    NmE4YjdlNTYxNmM1MzM1MWFmMjUxOTE2MDhlMGI1NTFkZTdlMWI5ZTQwOTA1
+    MjIyZGI1NzY4YzcyNGU1NWI3MDllNTIwOGZkNmU0OWE4OTc5OGU=
   data.tar.gz: !binary |-
-    YjY5NDlhZmQ0N2I0NGYzNjA0OGY4YzQ5MjY1YmI1MDM4OTIzODYxNDBjODIy
-    OGE3MjdjZTU4YmNkNmVmZjk1NDAxYjA2MmVkMzAzNzViY2ExYzg3MjU5ZjU0
-    ZGE1NTY5ZWQ1MTI2ZWU5MjJjMTI5Y2ZmYzM1NTI5YWU3YWJmNjM=
+    MjgzM2YxMTMzNDlkMDJlN2JhNWI4NzM4M2U4Y2IyOGI0MjkyOGY1MDEzZjNk
+    MjYzNGYzNzBlYzY2YzdhZGExM2M4MDMzMDBmYjYxYTRlMzNlMWQ4MmY0YWIy
+    MTllYzM2MDQwNGZiMjEwMjdmZjZhNjMxNGFlMDNhNWMzN2UwMjg=

data/lib/swordfish/formats/docx.rb

@@ -9,7 +9,6 @@ module Swordfish
 
     attr_reader :swordfish_doc   # The Swordfish::Document corresponding to the parsed document
     attr_reader :docx_archive    # The source archive
-    attr_reader :namespaces      # A hash of XML namespaces used in this doc
     
     # Parse a document and return a Swordfish::Document object
     def self.open(filepath)
@@ -48,7 +47,6 @@ module Swordfish
     # Parse the document structure XML
     def parse(document_xml)
       @xml = Nokogiri::XML(document_xml)
-      @namespaces = @xml.collect_namespaces
 
       # Iterate over each element node and dispatch it to the appropriate parser
       @xml.xpath('//w:body').children.each do |node|
@@ -170,10 +168,10 @@ module Swordfish
               text.content = run_xml.xpath('./w:t')[0].content
               get_styles_for_node(text, run_xml.xpath('./w:rPr')[0])
               texts << text
-            elsif @namespaces['xmlns:pic'] && run_xml.xpath('.//pic:pic', :pic => @namespaces['xmlns:pic']).length > 0
+            elsif run_xml.xpath('.//*[name()="pic:pic"]').length > 0
               # An image run
               image = Swordfish::Node::Image.new
-              relationship_id = run_xml.xpath('.//pic:pic/pic:blipFill/a:blip', :pic => @namespaces['xmlns:pic'], :a => @namespaces['xmlns:a'])[0]['r:embed'] rescue nil
+              relationship_id = run_xml.xpath('.//*[name()="pic:pic"]/*[name()="pic:blipFill"]/*[name()="a:blip"]')[0]['r:embed'] rescue nil
               if relationship_id
                 image.original_name = @relationships[relationship_id].split('/').last
                 @swordfish_doc.images[image.original_name] = read_image(image.original_name)
@@ -190,6 +188,17 @@ module Swordfish
             texts << link
         end
       end
+      # Clean up runs by merging them if they have identical styles
+      to_delete = []
+      texts.each_with_index do |text, idx|
+        if idx > 0
+          if text.is_a?(Swordfish::Node::Text) && texts[idx-1].is_a?(Swordfish::Node::Text) && text.style == texts[idx-1].style
+            texts[idx-1].content += text.content
+            to_delete << text
+          end
+        end
+      end
+      texts.reject! {|t| to_delete.include?(t) }
       texts
     end
 

data/lib/swordfish/stylesheet.rb

@@ -2,6 +2,8 @@
 
 module Swordfish
   class Stylesheet
+
+    attr_reader :styles
     
     # Define all supported values here
     SUPPORTED_STYLES = [
@@ -22,6 +24,11 @@ module Swordfish
       styles = [styles] unless styles.is_a?(Array)
       @styles |= styles.select{|s| SUPPORTED_STYLES.include?(s)}
     end
+
+    # Test stylesheets for equality (same styles)
+    def ==(other)
+      @styles.sort == other.styles.sort
+    end
     
     # For each supported style, define a boolean method to check its presence
     # (i.e., :bold?, :italic?, etc.)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: swordfish
 version: !ruby/object:Gem::Version
-  version: 0.0.4
+  version: 0.0.5
 platform: ruby
 authors:
 - Martin Posthumus
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-06-02 00:00:00.000000000 Z
+date: 2014-06-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler