isodoc 2.5.6 → 2.5.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7c7354811120c945f0a478b34503c89e841a49a8cffc6655e0be43f5b565ad35
-  data.tar.gz: 0f9dae634d16cbf5521eb048cbf4ff6de3d2e9d48d2284322183f598df7fc19b
+  metadata.gz: 89b2a927f8f452dd13667b90a02e9871472784e620c85978a83d6fd52015902b
+  data.tar.gz: f76d973900f03358f4411a8b479db06544e3d7ecc52971cd0c5afe60334046bd
 SHA512:
-  metadata.gz: 98e9b5625dfe9ab222ba00b03b02216e9f3ee00affa6e655c58be6acbeedb8bb05663b7d6832fbc94a76a8da02bc732a9f1b2c309532e968089990b1f50fe6ba
-  data.tar.gz: 756d64e9edbf24c6ba0ae909421e0a7c0a7f916fff8b5f6622d5792c9952728f05c8d2854888a559d3c6b36aa4a54d394ce7cfeded705fc168c0293240ea04b9
+  metadata.gz: aecc80230f4d1cde0232702931d07290c340f98c3a5e93a81603fed8c7775a1fca6dbf29cf09aa8141aece7fa3befb6d583c70690cf2cda02fcf50ce5bf75848
+  data.tar.gz: fc2d6d6fa85a1dc0edf1735068b940084b21fe20442158bc18307e9ae77581875dbf2b66be5cc4847bd2eb1096063b959c9df9695d356963e879c9c83e6a9e3b

data/lib/isodoc/function/cleanup.rb

@@ -30,33 +30,38 @@ module IsoDoc
 
         docxml.xpath("//td | //th").each do |d|
           d.traverse do |n|
-            next unless n.text?
-
-            n.replace(HTMLEntities.new.encode(
-                        break_up_long_strings(n.text),
-                      ))
+            n.text? or next
+            ret = break_up_long_str(n.text)
+            n.content = ret
           end
         end
       end
 
-      def break_up_long_strings(text)
-        return text if /^\s*$/.match?(text)
+      LONGSTR_THRESHOLD = 10
+      LONGSTR_NOPUNCT = 2
 
+      def break_up_long_str(text)
+        /^\s*$/.match?(text) and return text
         text.split(/(?=\s)/).map do |w|
-          if /^\s*$/.match(text) || (w.size < 30) then w
+          if /^\s*$/.match(text) || (w.size < LONGSTR_THRESHOLD) then w
           else
-            w.scan(/.{,30}/).map do |w1|
-              w1.size < 30 ? w1 : break_up_long_strings1(w1)
+            w.scan(/.{,#{LONGSTR_THRESHOLD}}/o).map.with_index do |w1, i|
+              w1.size < LONGSTR_THRESHOLD ? w1 : break_up_long_str1(w1, i + 1)
             end.join
           end
         end.join
       end
 
-      def break_up_long_strings1(text)
-        s = text.split(%r{(?<=[,.?+;/=])})
-        if s.size == 1 then "#{text} "
+      # break on punct every LONGSTRING_THRESHOLD chars
+      # break regardless every LONGSTRING_THRESHOLD * LONGSTR_NOPUNCT
+      def break_up_long_str1(text, iteration)
+        s = text.split(%r{(?<=[,.?+;/=(\[])})
+        if s.size == 1
+          (iteration % LONGSTR_NOPUNCT).zero? and
+            text += "\u200b"
+          text
         else
-          s[-1] = " #{s[-1]}"
+          s[-1] = "\u200b#{s[-1]}"
           s.join
         end
       end

data/lib/isodoc/html_function/html.rb

@@ -102,8 +102,6 @@ module IsoDoc
         end
       end
 
-      def table_long_strings_cleanup(docxml); end
-
       def table_attrs(node)
         ret = super
         node.at(ns("./colgroup")) and ret[:style] += "table-layout:fixed;"

data/lib/isodoc/version.rb

@@ -1,3 +1,3 @@
 module IsoDoc
-  VERSION = "2.5.6".freeze
+  VERSION = "2.5.7".freeze
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: isodoc
 version: !ruby/object:Gem::Version
-  version: 2.5.6
+  version: 2.5.7
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2023-06-05 00:00:00.000000000 Z
+date: 2023-06-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: html2doc