Class: ArxivSync::XMLParser

Inherits:
Ox::Sax
  • Object
show all
Defined in:
lib/arxivsync/parser.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize ⇒ XMLParser

Returns a new instance of XMLParser.



30
31
32
 
# File 'lib/arxivsync/parser.rb', line 30

def initialize
  @entities = HTMLEntities.new
end

Instance Attribute Details

#papers ⇒ Object

Returns the value of attribute papers.



28
29
30
 
# File 'lib/arxivsync/parser.rb', line 28

def papers
  @papers
end

Instance Method Details

#clean(str) ⇒ Object 



47
48
49
 
# File 'lib/arxivsync/parser.rb', line 47

def clean(str)
  str.gsub(/\s+/, ' ').strip
end

#decode(string) ⇒ Object 



51
52
53
54
55
56
57
58
 
# File 'lib/arxivsync/parser.rb', line 51

def decode(string)
  str = @entities.decode(string)
  LaTeX::Decode::Base.normalize(str)
  LaTeX::Decode::Accents.decode!(str)
  LaTeX::Decode::Diacritics.decode!(str)
  LaTeX::Decode::Symbols.decode!(str)
  str
end

#end_element(name) ⇒ Object 



118
119
120
121
122
123
124
125
126
 
# File 'lib/arxivsync/parser.rb', line 118

def end_element(name)
  case name
  when :version
    @model.versions.push(@version)
  when :metadata # End of a paper entry
    @papers.push(@model)
  end
  @el = nil
end

#start_element(name, attributes = []) ⇒ Object 



34
35
36
37
38
39
40
41
42
43
44
45
 
# File 'lib/arxivsync/parser.rb', line 34

def start_element(name, attributes=[])
  @el = name
  case name
  when :ListRecords
    @papers = []
  when :metadata
    @model = Paper.new
    @model.versions = []
  when :version
    @version = Version.new
  end
end

#text(str) ⇒ Object 



60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
 
# File 'lib/arxivsync/parser.rb', line 60

def text(str)
  case @el
  # Necessary elements
  when :id
    @model.id = clean(str)
  when :submitter
    @model.submitter = decode(clean(str))
  when :title
    @model.title = decode(clean(str))
  when :authors
    # Author strings may contain strange metadata
    # Non-regex parsing to handle nested parens
    depth = 0
    no_parens = ""

    str.chars do |ch|
      case ch
      when '('
        depth += 1
      when ')'
        depth -= 1
      else
        no_parens << ch if depth == 0
      end
    end

    @model.authors = clean(no_parens).split(/,| and /)
      .map { |s| decode(clean(s)) }
      .reject { |s| s.empty? }
  when :categories
    @model.categories = clean(str).split(/\s/)
  when :abstract
    @model.abstract = decode(clean(str))

  # Optional elements
  when :comments
    @model.comments = decode(clean(str))
  when :"msc-class"
    @model.msc_class = clean(str)
  when :"report-no"
    @model.report_no = clean(str)
  when :"journal-ref"
    @model.journal_ref = clean(str)
  when :doi
    @model.doi = clean(str)
  when :proxy
    @model.proxy = clean(str)
  when :license
    @model.license = clean(str)

  # Versions
  when :date
    @version.date = Time.parse(clean(str))
  when :size
    @version.size = clean(str)
  end
end