metanorma-utils 1.8.5 → 1.9.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: c6a75f252cf191accfebe5292830c5aa14653dcf5b4c02fda98bb1f7ec8dba85
-  data.tar.gz: 49a53a02959cb26c6998458151488a6ccab8c39a4b9f627d6d16dcd2d64f634f
+  metadata.gz: 717014b944b65d1e4f5b7c7ed0c71b3ac3314a3fb9ac740ff0027e9b5088cea8
+  data.tar.gz: c5bc40281e230a272deaf0469d4b9b6dd85e4f6b124de1e64d4c5c16347b2e8c
 SHA512:
-  metadata.gz: 29fdfcc884ba45d8a6112d7fefacf8988e1fc49d24bdae7285ebc9fbfc869de88fdb75e10e6e9db4b8e5e1287c3fa3f722eb414f18879ad15df44853d2dc9882
-  data.tar.gz: d5a7e5abe280a6b5dc09cd4661ce38e722433226964aaeb4e973ce6e77aee7c439b20b0714057185ee9c7e74cd3d6dd956a79e80f3fd5031d0cbe08f721b48c2
+  metadata.gz: ef77c94f8be1071eead01e24fdbdbf1367ea2fc19ede7f1b3bf0149127863e38d56fea39a42a4685b3c65112431c251fab1507602a54b0307dae108a7ab563c9
+  data.tar.gz: 6a86994cc199526918abbb80cb79f0c099a3aa2eb3480b4d763e381a351cea0184ccce285eef28b49cb9d441037469ea4f21df91258a3b09c64bc889c76bf80f

data/lib/utils/version.rb

@@ -1,5 +1,5 @@
 module Metanorma
   module Utils
-    VERSION = "1.8.5".freeze
+    VERSION = "1.9.0".freeze
   end
 end

data/lib/utils/xml.rb

@@ -62,6 +62,38 @@ module Metanorma
         end
       end
 
+      def noko(_script = "Latn", &block)
+        fragment = ::Nokogiri::XML.parse(NOKOHEAD).fragment("")
+        ::Nokogiri::XML::Builder.with fragment, &block
+        ret = fragment
+          .to_xml(encoding: "UTF-8", indent: 0,
+                  save_with: Nokogiri::XML::Node::SaveOptions::AS_XML)
+          .lines.map do |l|
+            l.rstrip.gsub("–", "\u0096").gsub("—", "\u0097")
+              .gsub("–", "\u0096").gsub("—", "\u0097")
+          end
+        line_sanitise(ret)
+      end
+
+      # By default, carriage return in source translates to whitespace;
+      # but in CJK, it does not.  We don't want carriage returns in the final
+      # output because of CJK complications
+      def line_sanitise(ret)
+        ret.size == 1 and return ret
+        (0...ret.size).each do |i|
+          last = firstchar_xml(ret[i].reverse)
+          nextfirst = firstchar_xml(ret[i + 1])
+          /#{CJK}/o.match?(last) && /#{CJK}/o.match?(nextfirst) or
+            ret[i] += " "
+        end
+        ret
+      end
+
+      def firstchar_xml(line)
+        m = /^(<[^>]+>)*(.)/.match(line) or return ""
+        m[2]
+      end
+
       def noko_html(&block)
         doc = ::Nokogiri::XML.parse(NOKOHEAD)
         fragment = doc.fragment("")

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: metanorma-utils
 version: !ruby/object:Gem::Version
-  version: 1.8.5
+  version: 1.9.0
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2024-07-19 00:00:00.000000000 Z
+date: 2024-07-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: asciidoctor