Module: Daengine::TeamsiteMetadataParser

Defined in:
lib/daengine/teamsite_metadata_parser.rb

Constant Summary collapse

@@translation =
{
    'TeamSite/Metadata/web_title' => 'title',
    "TeamSite/Metadata/enterprise_last_updated_date"=> 'changed_at',
    "TeamSite/Metadata/enterprise_audience_id"=> 'audiences',
    "TeamSite/Metadata/enterprise_sami_desc"=> 'sami_code',
    "TeamSite/Metadata/enterprise_last_publication_date"=> 'published_at',
    "TeamSite/Metadata/enterprise_unpublish_date"=> 'unpublished_at',
    "TeamSite/Metadata/enterprise_expiration_date"=> 'expires_at',
    "TeamSite/Metadata/enterprise_guid" => 'guid',
    "TeamSite/Metadata/shortSynopsis" => 'summary',
    "TeamSite/Metadata/business_owner" => 'business_owner',
    "TeamSite/Metadata/enterprise_product_id" => 'product_ids',
    "TeamSite/Metadata/enterprise_content_organization_id" => 'content_organization_ids',
    "TeamSite/Metadata/enterprise_program_id" => 'program_ids',
    "TeamSite/Metadata/omnitureSiteSection_codes" => 'omniture_codes'
}
@@path_tuples =
{
    "path" => 'path',
    "TeamSite/Metadata/enterprise_last_content_update_date" => 'doc_changed_at',
    "TeamSite/Metadata/enterprise_content_type_id" => 'content_type'
}
@@validations =
{
    "TeamSite/Metadata/display_on_website" => lambda { |val| /Y|1|true/ =~ val },
    "TeamSite/Metadata/enterprise_expiration_date" => lambda {|val| !val.blank? },
    # "TeamSite/Metadata/enterprise_unpublish_date" => lambda {|val| val.blank? },
    #"path" => lambda {|val| !(/\/manifest\// =~ val) }
}
@@logger =
nil

Class Method Summary collapse

Class Method Details

.log(args) ⇒ Object



42
43
44
# File 'lib/daengine/teamsite_metadata_parser.rb', line 42

def self.log(args)
  @@logger.error(args) unless @@logger.blank?
end

.logger=(some_logger) ⇒ Object



37
38
39
40
# File 'lib/daengine/teamsite_metadata_parser.rb', line 37

def self.logger=(some_logger)
  @@logger = some_logger
  self
end

.parse_tuple_file(file, last_read = nil) ⇒ Object



46
47
48
49
50
51
52
53
54
55
56
57
58
59
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
117
118
119
120
121
122
123
124
125
126
# File 'lib/daengine/teamsite_metadata_parser.rb', line 46

def self.parse_tuple_file(file, last_read = nil)
  time do
    asset = nil
    assets = {}
    docpath = {}
    valid = true
    while (line = file.gets)
      case line
        when /<\/?data-tuple>/
          if (asset.blank?)
            asset = DigitalAsset.new
          elsif (valid)
            assets[asset.guid] ||= asset.attributes # first tuple metadata wins
            assets[asset.guid]['documents_attributes'] ||= []
            assets[asset.guid]['documents_attributes'] << docpath
            # assets[asset.guid]['_id'] = asset.guid
            asset = nil; docpath = {}; valid = true;
          else
            asset = nil; docpath = {}; valid = true;
          end
        when /<tuple-field name="([^"]+)">([^<]+)<\/tuple-field>/
          if (valid)
            if @@validations[$1]
              valid = @@validations[$1].call($2)
            end
            if @@path_tuples[$1]
              docpath[@@path_tuples[$1]] = $2 # if this is one of our keys, 'send' to the method
            elsif (@@translation[$1])
              val = asset.send("#{@@translation[$1]}").respond_to?(:[]) ? $2.split(',') : $2
              asset.send("#{@@translation[$1]}=", val)
            end
          end
      end
    end
    # loop thru each doc in the collection, either replace or delete it
    error_files = []
    update_count = 0; delete_count = 0; added_count = 0;
    assets.keys.each do |key|
      da = nil
      begin
        if (!assets[key]['unpublished_at'].nil?)
          DigitalAsset.where(guid: key).try :delete_all
          delete_count += 1
        elsif (assets[key]['expires_at'].nil? || (assets[key]['expires_at'] < 1.minute.from_now))
          DigitalAsset.where(guid: key).try :delete_all
          delete_count += 1
        else
          da = DigitalAsset.find_or_initialize_by(guid: key)
          asset_docs = trim_package(assets[key]['documents_attributes'], last_read, da)
          if (!asset_docs.empty?)
            assets[key]['documents_attributes'] = asset_docs
            assets[key]['orderable'] = da.orderable unless da.new?

            creating = da.new?
            da.documents = []
            begin
              da.update_attributes!(assets[key])
            rescue Exception => e
              error_files << "#{e} ----  #{da.try(:guid)}, #{da.try(:errors).try(:full_messages)}"
            end
            creating ? added_count += 1 : update_count += 1
          else
            DigitalAsset.where(guid: key).try :delete_all
            delete_count += 1
          end
        end
      rescue Exception => e
       Daengine.log(e.message, "error") 
       Daengine.log(e.backtrace.join('\n'), "error")
      end
    end
    log_txt = "TeamsiteMetadataParser: Failed to save/update following DigitalAssets in database:\n"
    error_files.each do |asset_file|
      log_txt << "> #{asset_file}\n"
    end
    Daengine.log(log_txt, "warn") unless error_files.empty?
    Daengine.log("checking purge condition, stale count is #{DigitalAsset.stale.count}, bulk_processed is #{DigitalAsset.bulk_processed?}", 'warn')
    DigitalAsset.purge! # if the purge criteria is met, purge anything not updated
    Daengine.log("TeamsiteMetadataParser: #{added_count} records added, #{update_count} updated, #{delete_count} removed", "info")
  end
end

.timeObject



180
181
182
183
184
# File 'lib/daengine/teamsite_metadata_parser.rb', line 180

def self.time
  start = Time.now
  yield
  self.log "elapsed time was #{Time.now - start}"
end

.trim_package(asset_docs, last_read = nil, existing_da = nil) ⇒ Object



128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
# File 'lib/daengine/teamsite_metadata_parser.rb', line 128

def self.trim_package(asset_docs, last_read = nil, existing_da = nil)
  docs = []
  path = Daengine.config[:digital_assets_file_directory]
  if Dir.exist?(path)  # dont throw away assets if we can't locate the dir
    asset_docs.each do |doc|
      #exclude manifest_file
      unless doc['path'].match('\/manifest\/')
        file = File.join(path, doc['path'])
        #exclude digital_asset_files that are not in *Teamsite Staging*
        if File.exist?(file)
          docs << doc
          if(File.mtime(file) > last_read)
            begin
              exifdata = ::MiniExiftool.new(file) # spaces in filename
              pages = exifdata.pagecount
              pages = exifdata.pages if pages.blank?
              pages = exifdata.slides if pages.blank?
              if(pages.blank?)
                Daengine.log "Unable to get pages metadata for #{file}, exifdata.pagecount = #{exifdata.pagecount}, exifdata.pages = #{exifdata.pages}", "info"
              end
              doc['pages'] = pages
              doc['size'] = exifdata.filesize
              doc['mime_type'] = exifdata.mimetype
              Daengine.log "Exif data for #{file} was pages:#{doc['pages']}, size:#{doc['size']}", "info"
            rescue Exception => e
              Daengine.log "Error reading metadata from #{file} #{e.message}", "error"
            end
          elsif(existing_da && !existing_da.new?)  # copy over the existing asset metadata
           existing_doc = existing_da.documents.where(path: doc['path'])[0]
           if(existing_doc)
             doc['pages'] = existing_doc.pages
             doc['size'] = existing_doc.size
             doc['mime_type'] = existing_doc.mime_type
           else
             Daengine.log "Unable to find document with path #{file} on existing asset #{existing_da.guid} title: #{existing_da.title}", "info"
           end
           Daengine.log "Saving existing pages #{existing_doc.pages} and size #{existing_doc.size} on #{file}", "info"
          end
        else
          # the file was missing on disk, show a warning!
          Daengine.log("TeamsiteMetadataParser: Unable to locate file #{file} on disk!  Removing from metadata", "warn")
        end
      end
    end
  else
    Daengine.log("******************", "error")
    Daengine.log("ERROR!!!  Unable to locate physical teamsite files at #{path}", "error")
    Daengine.log("******************", "error")
  end
  docs
end