Module: EpubWorm::Extractors::Base

Included in:
Content, CoverReference, File, Manifest, Metadata, NcxNavigation, Spine, Version, XhtmlNavigation
Defined in:
lib/epub_worm/extractors/base.rb

Constant Summary collapse

DEFAULT_NS = 
{
  "container" => "urn:oasis:names:tc:opendocument:xmlns:container"
}.freeze

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

.extended(base) ⇒ Object 



13
14
15
 
# File 'lib/epub_worm/extractors/base.rb', line 13

def self.extended(base)
  base.instance_variable_set(:@ns, DEFAULT_NS.dup)
end

Instance Method Details

#as_xml(file) ⇒ Object 



45
46
47
 
# File 'lib/epub_worm/extractors/base.rb', line 45

def as_xml(file)
  Nokogiri::XML(file.get_input_stream.read)
end

#element_at(doc, path) ⇒ Object 



49
50
51
 
# File 'lib/epub_worm/extractors/base.rb', line 49

def element_at(doc, path)
  doc.at_xpath(path, ns)
end

#elements_at(doc, path) ⇒ Object 



53
54
55
 
# File 'lib/epub_worm/extractors/base.rb', line 53

def elements_at(doc, path)
  doc.xpath(path, ns)
end

#ns ⇒ Object 



22
23
24
 
# File 'lib/epub_worm/extractors/base.rb', line 22

def ns
  @ns || {}
end

#ns_entry(prefix, uri) ⇒ Object 



17
18
19
20
 
# File 'lib/epub_worm/extractors/base.rb', line 17

def ns_entry(prefix, uri)
  @ns ||= {}
  @ns[prefix.to_s] = uri
end

#open_opf(path) ⇒ Object 



26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
 
# File 'lib/epub_worm/extractors/base.rb', line 26

def open_opf(path)
  Zip::File.open(path) do |zip_file|
    # Get the container.xml file to get the opf file path.
    container_file = zip_file.find_entry "META-INF/container.xml"
    raise ::EpubWorm::Error, "container.xml not found" unless container_file

    # Read the container file to get opf file path.
    container_doc = Nokogiri::XML(container_file.get_input_stream.read)
    opf_file_path = element_at(container_doc, "//container:rootfile")["full-path"]

    # Read the opf file to get metadata.
    opf_file = zip_file.find_entry opf_file_path
    raise ::EpubWorm::Error, "opf file not found" unless opf_file

    opf_doc = Nokogiri::XML(opf_file.get_input_stream.read)
    yield(opf_doc, opf_file_path, zip_file) if block_given?
  end
end

#text_at(doc, path) ⇒ Object 



57
58
59
 
# File 'lib/epub_worm/extractors/base.rb', line 57

def text_at(doc, path)
  element_at(doc, path)&.text&.strip
end