bergamasco 0.3.2 → 0.3.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 63d34d9a6d674818e54cedec4ba4ae26bfc98d1a
-  data.tar.gz: 3d373fd955263556b5766c3e1319a720d39f62b9
+  metadata.gz: 4c9fd041fa952b8ce0d759d55f1a5ea9fbb0eeba
+  data.tar.gz: 2fabed95db18a9ab876e1d6843fe01ca9d645c88
 SHA512:
-  metadata.gz: f8b6b4045b3ffacb31f7bf3f0c019b374cf534ffafa21850dd4a288295fb516e2dffec9793611e5ef8d64919cf3705f83a7eb099f790ef88a1e147b43cc1b51e
-  data.tar.gz: a06d40fd69e0f7606e0b87fcba01d151ca85e4b1e99afb2523f90eadc121da5dedb25a2a42d3b1375e4a1bcac433ba231ff4fddf194c6cb42f5ae45fb562ade4
+  metadata.gz: c073e07cb08e34c90c5afdfd36ac11fb88fc341a26f7d892d7bd9c5aed6e90632f5e937f8bcfa015e40e2c191b076ff38707c365faa2da5e91e289fa143bb7a9
+  data.tar.gz: 1855153f47914dae465a7db87cfe2910eb53fa95bec68e7defaa8e764929bf4cb946109c15b52c0243da1c5d476bdcfc9566589be925ad8ff5784b8abd941ea5

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    bergamasco (0.3.2)
+    bergamasco (0.3.3)
       activesupport (~> 4.2, >= 4.2.5)
       addressable (~> 2.3.8)
       builder (~> 3.2, >= 3.2.2)
@@ -79,4 +79,4 @@ DEPENDENCIES
   simplecov
 
 BUNDLED WITH
-   1.12.5
+   1.13.6

data/lib/bergamasco.rb

@@ -13,5 +13,5 @@ require 'addressable/uri'
 require "bergamasco/summarize"
 require "bergamasco/sanitize"
 require "bergamasco/markdown"
-require "bergamasco/jats"
+require "bergamasco/pandoc"
 require "bergamasco/whitelist_scrubber"

data/lib/bergamasco/markdown.rb

@@ -48,7 +48,7 @@ module Bergamasco
       metadata = yaml.extract!(*keys).compact
 
       content = YAML_FRONT_MATTER_REGEXP.match(file).post_match
-      html = render_html(content, options)
+      html = Bergamasco::Pandoc.convert(content, options)
       metadata["summary"] = Bergamasco::Summarize.summary_from_html(html, options)
       metadata["references"] = extract_references(html)
       metadata["date"] = metadata["date"].iso8601
@@ -59,31 +59,6 @@ module Bergamasco
       IO.write(filepath, content.to_yaml)
     end
 
-    def self.render_html(text, options={})
-      PandocRuby.new(text, options.except(:skip_yaml_header,
-                                          :separator,
-                                          :sitepath,
-                                          :authorpath,
-                                          :referencespath,
-                                          :username,
-                                          :password,
-                                          :sandbox,
-                                          :prefix,
-                                          :number)).to_html
-    rescue Errno::ENOENT
-      # fallback to commonmarker if pandoc is not installed.
-      # Commonmarker doesn't parse or ignore yaml frontmatter
-      text = split_yaml_frontmatter(text).last if options[:skip_yaml_header]
-      CommonMarker.render_html(text, :default)
-    end
-
-    def self.write_bibliograpy_to_yaml(bib_path, yaml_path)
-      yaml = `pandoc-citeproc --bib2yaml #{bib_path} 2>&1`
-      return nil if $?.exitstatus > 0
-      IO.write(yaml_path, yaml)
-      SafeYAML.load(yaml)
-    end
-
     # expects a references list generated by pandoc
     def self.extract_references(html)
       doc = Nokogiri::HTML(html)

data/lib/bergamasco/pandoc.rb

@@ -0,0 +1,60 @@
+module Bergamasco
+  module Pandoc
+    # Options understood by pandoc, taken from http://pandoc.org/MANUAL.html.
+    # Ignore all other options passed to pandoc, unless overriden.
+    AVAILABLE_OPTIONS = Set.new %w(from read to write output data-dir strict
+      parse-raw smart old-dashes base-header-level indented-code-classes filter
+      normalize preserve-tabs tab-stop track-changes file-scope extract-media
+      standalone template metadata variable print-default-template
+      print-default-data-file no-wrap wrap columns toc table-of-contents toc-depth
+      no-highlight highlight-style include-in-header include-before-body
+      include-after-body self-contained offline html5 html-q-tags ascii
+      reference-links reference-location atx-headers chapters top-level-division
+      number-sections number-offsetS no-tex-ligatures listings incremental
+      slide-level section-divs default-image-extension email-obfuscation id-prefix
+      title-prefix css reference-odt reference-docx epub-stylesheet
+      epub-cover-image epub-metadata epub-embed-font epub-chapter-level
+      latex-engine latex-engine-opt bibliography csl citation-abbreviations natbib
+      biblatex latexmathml asciimathml mathml mimetex webtex jsmath mathjax katex
+      katex-stylesheet gladtex trace dump-args ignore-args verbose bash-completion
+      list-input-formats list-output-formats list-extensions
+      list-highlight-languages list-highlight-styles)
+    ALIAS_OPTIONS = Set.new %w(f r t w o R S F p s M V D H B A 5 N i T c m)
+    ALLOWED_OPTIONS = AVAILABLE_OPTIONS + ALIAS_OPTIONS
+
+    def self.convert(text, options={})
+      options = options.select { |k, v| ALLOWED_OPTIONS.include?(k.to_s.gsub('_', '-')) }
+
+      options[:from] ||= :markdown
+      options[:to] ||= :html
+
+      PandocRuby.convert(text, options)
+    rescue Errno::ENOENT
+      puts "Pandoc is not installed"
+    end
+
+    def self.convert_to_jats(text, options={})
+      options = options.merge(template: "templates/default.jats",
+                              to: "lib/bergamasco/jats.lua",
+                              csl: "lib/bergamasco/jats.csl")
+      options = options.merge(metadata: options[:metadata]) if options[:metadata].present?
+
+      convert(text, options)
+    end
+
+    def self.write_jats(filepath, options={})
+      file = IO.read(filepath)
+      xml_path = File.join(File.dirname(filepath), File.basename(filepath, ".html.md")) + ".xml"
+      xml = convert_to_jats(file, options)
+      IO.write(xml_path, xml)
+      xml_path
+    end
+
+    def self.write_bibliograpy_to_yaml(bib_path, yaml_path)
+      yaml = `pandoc-citeproc --bib2yaml #{bib_path} 2>&1`
+      return nil if $?.exitstatus > 0
+      IO.write(yaml_path, yaml)
+      SafeYAML.load(yaml)
+    end
+  end
+end

data/lib/bergamasco/version.rb

@@ -1,3 +1,3 @@
 module Bergamasco
-  VERSION = "0.3.2"
+  VERSION = "0.3.3"
 end

data/spec/markdown_spec.rb

@@ -64,13 +64,6 @@ describe Bergamasco::Markdown do
     expect(length).to eq(text.length)
   end
 
-  it 'should write bibliography to yaml' do
-    bib_path = fixture_path + 'references.bib'
-    yaml_path = fixture_path + 'references.yaml'
-    yaml = subject.write_bibliograpy_to_yaml(bib_path, yaml_path)
-    expect(yaml["references"].length).to eq(61)
-  end
-
   it 'should update file' do
     filepath = fixture_path + 'cool-dois.html.md'
     new_metadata = { "doi" => "10.23725/0000-03VC"}
@@ -81,14 +74,14 @@ describe Bergamasco::Markdown do
   it 'should convert markdown' do
     filepath = fixture_path + 'cool-dois.html.md'
     file = IO.read(filepath)
-    html = subject.render_html(file, skip_yaml_header: true, csl: 'spec/fixtures/apa.csl', bibliography: 'spec/fixtures/references.yaml')
+    html = Bergamasco::Pandoc.convert(file, skip_yaml_header: true, csl: 'spec/fixtures/apa.csl', bibliography: 'spec/fixtures/references.yaml')
     expect(html).to start_with("<p>In 1998 Tim Berners-Lee coined the term cool URIs <span class=\"citation\">(1998)</span>, that is URIs that don’t change.")
   end
 
   it 'should extract references' do
     filepath = fixture_path + 'cool-dois.html.md'
     file = IO.read(filepath)
-    html = subject.render_html(file, skip_yaml_header: true, csl: 'spec/fixtures/apa.csl', bibliography: 'spec/fixtures/references.yaml')
+    html = Bergamasco::Pandoc.convert(file, skip_yaml_header: true, csl: 'spec/fixtures/apa.csl', bibliography: 'spec/fixtures/references.yaml')
     refs = subject.extract_references(html)
     expect(refs).to eq(["https://www.w3.org/Provider/Style/URI", "https://doi.org/10.1371/journal.pone.0115253"])
   end

data/spec/pandoc_spec.rb

@@ -0,0 +1,38 @@
+require 'spec_helper'
+
+describe Bergamasco::Pandoc do
+  subject { Bergamasco::Pandoc }
+
+  it 'should convert markdown' do
+    filepath = fixture_path + 'cool-dois.html.md'
+    file = IO.read(filepath)
+    html = subject.convert(file, skip_yaml_header: true, csl: 'spec/fixtures/apa.csl', bibliography: 'spec/fixtures/references.yaml')
+    expect(html).to start_with("<p>In 1998 Tim Berners-Lee coined the term cool URIs <span class=\"citation\">(1998)</span>, that is URIs that don’t change.")
+  end
+
+  it 'should convert to jats' do
+    filepath = fixture_path + 'cool-dois.html.md'
+    file = IO.read(filepath)
+    xml = subject.convert_to_jats(file, skip_yaml_header: true, csl: 'spec/fixtures/apa.csl', bibliography: 'spec/fixtures/references.yaml')
+    doc = Nokogiri::XML(xml)
+    article_id = doc.at_xpath("//article-id")
+    expect(article_id.text).to eq("10.23725/0000-03VC")
+    expect(article_id.values.first).to eq("doi")
+  end
+
+  it 'should write jats xml' do
+    filepath = fixture_path + 'cool-dois.html.md'
+    xml_path = subject.write_jats(filepath, skip_yaml_header: true, csl: 'spec/fixtures/apa.csl', bibliography: 'spec/fixtures/references.yaml')
+    doc = File.open(xml_path) { |f| Nokogiri::XML(f) }
+    article_id = doc.at_xpath("//article-id")
+    expect(article_id.text).to eq("10.23725/0000-03VC")
+    expect(article_id.values.first).to eq("doi")
+  end
+
+  it 'should write bibliography to yaml' do
+    bib_path = fixture_path + 'references.bib'
+    yaml_path = fixture_path + 'references.yaml'
+    yaml = subject.write_bibliograpy_to_yaml(bib_path, yaml_path)
+    expect(yaml["references"].length).to eq(61)
+  end
+end

data/spec/summarize_spec.rb

@@ -30,7 +30,7 @@ describe Bergamasco::Summarize do
   it 'should truncate at separator and convert to html' do
     filepath = fixture_path + 'cool-dois-without-yml.md'
     file = IO.read(filepath)
-    html = Bergamasco::Markdown.render_html(file, skip_yaml_header: true, csl: 'spec/fixtures/apa.csl', bibliography: 'spec/fixtures/references.bib')
+    html = Bergamasco::Pandoc.convert(file, skip_yaml_header: true, csl: 'spec/fixtures/apa.csl', bibliography: 'spec/fixtures/references.bib')
     content = subject.summary_from_html(html, skip_yaml_header: true, csl: 'spec/fixtures/apa.csl', bibliography: 'spec/fixtures/references.bib')
     expect(content).to start_with("In 1998 Tim Berners-Lee coined the term cool URIs (1998), that is URIs that don’t change.")
     expect(content).to end_with("the referenced resource.")

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: bergamasco
 version: !ruby/object:Gem::Version
-  version: 0.3.2
+  version: 0.3.3
 platform: ruby
 authors:
 - Martin Fenner
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-12-21 00:00:00.000000000 Z
+date: 2016-12-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
@@ -269,8 +269,8 @@ files:
 - lib/bergamasco.rb
 - lib/bergamasco/jats.csl
 - lib/bergamasco/jats.lua
-- lib/bergamasco/jats.rb
 - lib/bergamasco/markdown.rb
+- lib/bergamasco/pandoc.rb
 - lib/bergamasco/sanitize.rb
 - lib/bergamasco/summarize.rb
 - lib/bergamasco/version.rb
@@ -283,8 +283,8 @@ files:
 - spec/fixtures/cool-dois.yml
 - spec/fixtures/references.bib
 - spec/fixtures/references.yaml
-- spec/jats_spec.rb
 - spec/markdown_spec.rb
+- spec/pandoc_spec.rb
 - spec/sanitize_spec.rb
 - spec/spec_helper.rb
 - spec/summarize_spec.rb

data/lib/bergamasco/jats.rb

@@ -1,33 +0,0 @@
-module Bergamasco
-  module Jats
-
-    def self.render_jats(text, options={})
-      options = options.merge(template: "templates/default.jats",
-                              to: "lib/bergamasco/jats.lua",
-                              csl: "lib/bergamasco/jats.csl")
-      options = options.merge(metadata: options[:metadata]) if options[:metadata].present?
-      converter = PandocRuby.new(text, options.except(:skip_yaml_header,
-                                                      :separator,
-                                                      :sitepath,
-                                                      :authorpath,
-                                                      :referencespath,
-                                                      :username,
-                                                      :password,
-                                                      :sandbox,
-                                                      :prefix,
-                                                      :number))
-      converter.convert
-    rescue Errno::ENOENT
-      # if pandoc is not installed.
-      puts "Pandoc is not installed"
-    end
-
-    def self.write_jats(filepath, options={})
-      file = IO.read(filepath)
-      xml_path = File.join(File.dirname(filepath), File.basename(filepath, ".html.md")) + ".xml"
-      xml = render_jats(file, options)
-      IO.write(xml_path, xml)
-      xml_path
-    end
-  end
-end

data/spec/jats_spec.rb

@@ -1,24 +0,0 @@
-require 'spec_helper'
-
-describe Bergamasco::Jats do
-  subject { Bergamasco::Jats }
-
-  it 'should convert to jats' do
-    filepath = fixture_path + 'cool-dois.html.md'
-    file = IO.read(filepath)
-    xml = subject.render_jats(file, skip_yaml_header: true, csl: 'spec/fixtures/apa.csl', bibliography: 'spec/fixtures/references.yaml')
-    doc = Nokogiri::XML(xml)
-    article_id = doc.at_xpath("//article-id")
-    expect(article_id.text).to eq("10.23725/0000-03VC")
-    expect(article_id.values.first).to eq("doi")
-  end
-
-  it 'should write jats xml' do
-    filepath = fixture_path + 'cool-dois.html.md'
-    xml_path = subject.write_jats(filepath, skip_yaml_header: true, csl: 'spec/fixtures/apa.csl', bibliography: 'spec/fixtures/references.yaml')
-    doc = File.open(xml_path) { |f| Nokogiri::XML(f) }
-    article_id = doc.at_xpath("//article-id")
-    expect(article_id.text).to eq("10.23725/0000-03VC")
-    expect(article_id.values.first).to eq("doi")
-  end
-end