Module: IsoDoc::Function::Utils
- Included in:
- Common
- Defined in:
- lib/isodoc/function/utils.rb
Constant Summary collapse
- NOKOHEAD =
add namespaces for Word fragments
"<!DOCTYPE html SYSTEM\n\"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd\">\n<html xmlns=\"http://www.w3.org/1999/xhtml\">\n<head> <title></title> <meta charset=\"UTF-8\" /> </head>\n<body> </body> </html>\n"- DOCTYPE_HDR =
'<!DOCTYPE html SYSTEM '\ '"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">'
- CLAUSE_ANCESTOR =
".//ancestor::*[local-name() = 'annex' or "\ "local-name() = 'definitions' or "\ "local-name() = 'acknowledgements' or local-name() = 'term' or "\ "local-name() = 'appendix' or local-name() = 'foreword' or "\ "local-name() = 'introduction' or local-name() = 'terms' or "\ "local-name() = 'clause' or local-name() = 'references']/@id"
- NOTE_CONTAINER_ANCESTOR =
".//ancestor::*[local-name() = 'annex' or "\ "local-name() = 'foreword' or local-name() = 'appendix' or "\ "local-name() = 'introduction' or local-name() = 'terms' or "\ "local-name() = 'acknowledgements' or local-name() = 'term' or "\ "local-name() = 'clause' or local-name() = 'references' or "\ "local-name() = 'figure' or local-name() = 'formula' or "\ "local-name() = 'table' or local-name() = 'example']/@id"
Instance Method Summary collapse
- #attr_code(attributes) ⇒ Object
- #date_range(date) ⇒ Object
- #empty2nil(v) ⇒ Object
-
#extract_delims(text) ⇒ Object
avoid ‘; avoid {{ (Liquid Templates); avoid [[ (Javascript).
- #from_xhtml(xml) ⇒ Object
- #get_clause_id(node) ⇒ Object
- #get_note_container_id(node) ⇒ Object
- #header_strip(h) ⇒ Object
- #image_localfile(i) ⇒ Object
- #insert_tab(out, n) ⇒ Object
- #labelled_ancestor(node) ⇒ Object
- #liquid(doc) ⇒ Object
-
#noko(&block) ⇒ Object
block for processing XML document fragments as XHTML, to allow for HTMLentities.
- #ns(xpath) ⇒ Object
- #populate_template(docxml, _format = nil) ⇒ Object
- #save_dataimage(uri, _relative_dir = true) ⇒ Object
- #sentence_join(array) ⇒ Object
- #to_xhtml(xml) ⇒ Object
- #to_xhtml_fragment(xml) ⇒ Object
Instance Method Details
#attr_code(attributes) ⇒ Object
38 39 40 41 42 43 |
# File 'lib/isodoc/function/utils.rb', line 38 def attr_code(attributes) attributes = attributes.reject { |_, val| val.nil? }.map attributes.map do |k, v| [k, v.is_a?(String) ? HTMLEntities.new.decode(v) : v] end.to_h end |
#date_range(date) ⇒ Object
5 6 7 |
# File 'lib/isodoc/function/utils.rb', line 5 def date_range(date) self.class.date_range(date) end |
#empty2nil(v) ⇒ Object
150 151 152 153 |
# File 'lib/isodoc/function/utils.rb', line 150 def empty2nil(v) return nil if !v.nil? && v.is_a?(String) && v.empty? v end |
#extract_delims(text) ⇒ Object
avoid ‘; avoid {{ (Liquid Templates); avoid [[ (Javascript)
112 113 114 115 116 117 118 119 120 |
# File 'lib/isodoc/function/utils.rb', line 112 def extract_delims(text) @openmathdelim = "(#(" @closemathdelim = ")#)" while text.include?(@openmathdelim) || text.include?(@closemathdelim) @openmathdelim += "(" @closemathdelim += ")" end [@openmathdelim, @closemathdelim] end |
#from_xhtml(xml) ⇒ Object
70 71 72 |
# File 'lib/isodoc/function/utils.rb', line 70 def from_xhtml(xml) xml.to_xml.sub(%r{ xmlns="http://www.w3.org/1999/xhtml"}, "") end |
#get_clause_id(node) ⇒ Object
82 83 84 85 |
# File 'lib/isodoc/function/utils.rb', line 82 def get_clause_id(node) clause = node.xpath(CLAUSE_ANCESTOR) clause&.last&.text || nil end |
#get_note_container_id(node) ⇒ Object
96 97 98 99 |
# File 'lib/isodoc/function/utils.rb', line 96 def get_note_container_id(node) container = node.xpath(NOTE_CONTAINER_ANCESTOR) container&.last&.text || nil end |
#header_strip(h) ⇒ Object
122 123 124 125 126 127 128 129 130 131 132 133 134 135 |
# File 'lib/isodoc/function/utils.rb', line 122 def header_strip(h) h = h.to_s.gsub(%r{<br\s*/>}, " ").gsub(/<\/?h[123456][^>]*>/, "") .gsub(/<\/?b[^>]*>/, "") h1 = to_xhtml_fragment(h.dup) h1.traverse do |x| x.replace(" ") if x.name == "span" && /mso-tab-count/.match(x["style"]) x.remove if x.name == "img" x.remove if x.name == "span" && x["class"] == "MsoCommentReference" x.remove if x.name == "a" && x["class"] == "FootnoteRef" x.remove if x.name == "span" && /mso-bookmark/.match(x["style"]) x.replace(x.children) if x.name == "a" end from_xhtml(h1) end |
#image_localfile(i) ⇒ Object
178 179 180 181 182 183 184 185 186 |
# File 'lib/isodoc/function/utils.rb', line 178 def image_localfile(i) if /^data:/.match? i["src"] save_dataimage(i["src"], false) elsif %r{^([A-Z]:)?/}.match? i["src"] i["src"] else File.join(@localdir, i["src"]) end end |
#insert_tab(out, n) ⇒ Object
13 14 15 16 |
# File 'lib/isodoc/function/utils.rb', line 13 def insert_tab(out, n) tab = %w(Hans Hant).include?(@script) ? " " : " " [1..n].each { out << tab } end |
#labelled_ancestor(node) ⇒ Object
188 189 190 191 |
# File 'lib/isodoc/function/utils.rb', line 188 def labelled_ancestor(node) !node.ancestors("example, requirement, recommendation, permission, "\ "note, table, figure, sourcecode").empty? end |
#liquid(doc) ⇒ Object
137 138 139 |
# File 'lib/isodoc/function/utils.rb', line 137 def liquid(doc) self.class.liquid(doc) end |
#noko(&block) ⇒ Object
block for processing XML document fragments as XHTML, to allow for HTMLentities
29 30 31 32 33 34 35 36 |
# File 'lib/isodoc/function/utils.rb', line 29 def noko(&block) doc = ::Nokogiri::XML.parse(NOKOHEAD) fragment = doc.fragment("") ::Nokogiri::XML::Builder.with fragment, &block fragment.to_xml(encoding: "US-ASCII").lines.map do |l| l.gsub(/\s*\n/, "") end end |
#ns(xpath) ⇒ Object
9 10 11 |
# File 'lib/isodoc/function/utils.rb', line 9 def ns(xpath) self.class.ns(xpath) end |
#populate_template(docxml, _format = nil) ⇒ Object
155 156 157 158 159 160 161 162 163 164 |
# File 'lib/isodoc/function/utils.rb', line 155 def populate_template(docxml, _format = nil) = .get .merge(@labels ? {labels: @labels} : {}) .merge(.labels ? {labels: .labels} : {}) .merge( || {}) template = liquid(docxml) template.render(.map { |k, v| [k.to_s, empty2nil(v)] }.to_h) .gsub("<", "<").gsub(">", ">").gsub("&", "&") end |
#save_dataimage(uri, _relative_dir = true) ⇒ Object
166 167 168 169 170 171 172 173 174 175 176 |
# File 'lib/isodoc/function/utils.rb', line 166 def save_dataimage(uri, _relative_dir = true) %r{^data:(image|application)/(?<imgtype>[^;]+);base64,(?<imgdata>.+)$} =~ uri imgtype.sub!(/\+[a-z0-9]+$/, "") # svg+xml imgtype = "png" unless /^[a-z0-9]+$/.match imgtype Tempfile.open(["image", ".#{imgtype}"]) do |f| f.binmode f.write(Base64.strict_decode64(imgdata)) @tempfile_cache << f # persist to the end f.path end end |
#sentence_join(array) ⇒ Object
101 102 103 104 105 106 107 108 109 |
# File 'lib/isodoc/function/utils.rb', line 101 def sentence_join(array) return '' if array.nil? || array.empty? if array.length == 1 then array[0] else @i18n.l10n("#{array[0..-2].join(', ')} "\ "#{@i18n.and} #{array.last}", @lang, @script) end end |
#to_xhtml(xml) ⇒ Object
48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 |
# File 'lib/isodoc/function/utils.rb', line 48 def to_xhtml(xml) xml.gsub!(/<\?xml[^>]*>/, "") /<!DOCTYPE /.match(xml) || (xml = DOCTYPE_HDR + xml) xml = xml.split(/(\&[^ \r\n\t#;]+;)/).map do |t| /^(\&[^ \t\r\n#;]+;)/.match?(t) ? HTMLEntities.new.encode(HTMLEntities.new.decode(t), :hexadecimal) : t end.join("") begin Nokogiri::XML.parse(xml, &:strict) rescue Nokogiri::XML::SyntaxError => e File.open("#{@filename}.#{@format}.err", "w:UTF-8") { |f| f.write xml } abort "Malformed Output XML for #{@format}: #{e} "\ "(see #{@filename}.#{@format}.err)" end end |
#to_xhtml_fragment(xml) ⇒ Object
64 65 66 67 68 |
# File 'lib/isodoc/function/utils.rb', line 64 def to_xhtml_fragment(xml) doc = ::Nokogiri::XML.parse(NOKOHEAD) fragment = doc.fragment(xml) fragment end |