Class: Run

46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

# File 'app/models/run.rb', line 46

def apply_fix(xml, fix, pe = nil)
  begin
    pre_fix_xml = xml.dup
  # HAX: Swallow mysterious namespace failure, come ON Noko
  rescue Java::OrgW3cDom::DOMException => e
    pre_fix_xml = Nokogiri::XML(xml.serialize, nil, 'UTF-8') {|config| config.nonet}
  end

  begin # In case of failure, catch the XML
    fix.(xml)
  rescue Fixes::Failure, StandardError => e
    pe.update(failed: true) if pe
    pre_fix_xml
  end
end

164
165
166
167
168
169
170
171
172

# File 'app/models/run.rb', line 164

def close_zipfiles(*zouts)
  zouts.each do |zout|
    begin
      zout.close
    rescue java.io.IOException => e
      # already closed, nothing to do here
    end
  end
end

26

# File 'app/models/run.rb', line 26

has_many :concrete_issues, dependent: :destroy

25

# File 'app/models/run.rb', line 25

has_and_belongs_to_many :finding_aid_versions

31
32
33
34
35
36
37
38
39
40
41
42
43

# File 'app/models/run.rb', line 31

def perform_analysis(faids)
  @checker = Checker.new(schematron, self)
  faids.each do |faid|
    faid = faid.current if faid.is_a? FindingAid
    ActiveRecord::Base.transaction do
      @checker.check(faid).each do |h|
        ConcreteIssue.create!(h)
      end
      self.finding_aid_versions << faid
      self.increment! :eads_processed
    end
  end
end

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
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140

# File 'app/models/run.rb', line 64

def perform_processing!
  raise "This run is already processed!" if run_for_processing
  update(run_for_processing: true)
  outdir = File.join(OUTPUT_DIR, "#{id}").shellescape
  indir =  File.join(INPUT_DIR,  "#{id}").shellescape
  Dir.mkdir(outdir, 0755) unless File.directory?(outdir)
  Dir.mkdir(indir, 0755) unless File.directory?(indir)

  # Stream input files to zip
  zout_in = java.util.zip.ZipOutputStream.new(File.open(File.join(indir, 'input.zip'), 'wb', 0644).to_outputstream)
  zout_out = java.util.zip.ZipOutputStream.new(File.open(File.join(outdir, 'out.zip'), 'wb', 0644).to_outputstream)

  finding_aid_versions
    .joins(:finding_aid, :concrete_issues => :issue)
    .select('finding_aid_versions.*,
             finding_aids.eadid,
             ARRAY_AGG(DISTINCT issues.identifier) AS identifiers')
    .group('finding_aids.eadid,finding_aid_versions.id')
    .each do |fa|
      add_to_zip(zout_in, fa.eadid, fa.file)

      # Preflight XML
      fa_xml = Fixes.preflights.values.reduce(fa.xml) do |xml, fix|
        apply_fix(xml, fix)
      end

      # Apply all relevant fixes to Finding Aid
      repaired = Fixes
                 .to_h
                 .select {|identifier, _| fa.identifiers.include? identifier}
                 .reduce(fa_xml) do|xml, (identifier, fix)|
        pe = processing_events.create(issue_id: schematron.issues.find_by(identifier: identifier).id,
                                      finding_aid_version_id: fa.id)
        apply_fix(xml, fix, pe)

      end # end of .reduce

      # Any problems which have fixes that exist now should theoretically
      # be things that were shadowed by the first pass, so take additional passes
      # untill either no known issues or MAX_PASSES
      MAX_PASSES.times do
        remaining_problems = schematron.issues.where(id: @checker.check_str(repaired.serialize(encoding: 'UTF-8')).map {|el| el[:issue_id]}.uniq).pluck(:identifier) & Fixes.to_h.keys

        # Run a second round of fixing if there are remaining problems
        break if remaining_problems.blank?
        repaired = Fixes
                   .to_h
                   .select {|identifier, _| remaining_problems.include? identifier}
                   .reduce(repaired) do |xml, (identifier, fix)|
          pe = processing_events.create(issue_id: schematron.issues.find_by(identifier: identifier).id,
                                        finding_aid_version_id: fa.id)
          apply_fix(xml, fix, pe)
        end
      end


      # Add notice of processing to revisiondesc
      today = DateTime.now.in_time_zone
      rd = repaired.at_xpath('/ead/eadheader/revisiondesc') || repaired.at_xpath('/ead/eadheader').add_child('<revisiondesc />').first
      rd.prepend_child(Nokogiri::XML::DocumentFragment.new(repaired, "\n" + <<-FRAGMENT.strip_heredoc + "\n"))
        <change>
          <date calendar="gregorian" era="ce" normal="#{today.strftime('%Y%m%d')}">#{today.strftime('%m/%d/%Y')}</date>
          <item>This resource was modified by the ArchivesSpace Preprocessor developed by the Harvard Library (https://github.com/harvard-library/archivesspace-preprocessor)</item>
        </change>
      FRAGMENT

      File.open(File.join(outdir, "#{fa.eadid}.xml"), 'w', 0644) do |f|
        repaired.write_xml_to(f, encoding: 'UTF-8')
      end

      add_to_zip(zout_out, fa.eadid, File.open(File.join(outdir, "#{fa.eadid}.xml"), 'r'))
  end

  update(completed_at: DateTime.now)
ensure
  close_zipfiles(zout_in, zout_out)
end

143
144
145
146

# File 'app/models/run.rb', line 143

def perform_processing_run(faids)
  perform_analysis(faids)
  perform_processing!
end

27

# File 'app/models/run.rb', line 27

has_many :processing_events, dependent: :destroy

24

# File 'app/models/run.rb', line 24

belongs_to :schematron