Sha256: 33012c45d5fc7a00b664a8c187d021434f88f0105360cdc6bc1d522da5e0f6d0
Contents?: true
Size: 1.28 KB
Versions: 11
Compression:
Stored size: 1.28 KB
Contents
module Sufia module GenericFile module FullTextIndexing extend ActiveSupport::Concern included do contains 'full_text' end def append_metadata super extract_content end private def extract_content url = Blacklight.solr_config[:url] ? Blacklight.solr_config[:url] : Blacklight.solr_config["url"] ? Blacklight.solr_config["url"] : Blacklight.solr_config[:fulltext] ? Blacklight.solr_config[:fulltext]["url"] : Blacklight.solr_config[:default]["url"] uri = URI("#{url}/update/extract?extractOnly=true&wt=json&extractFormat=text") req = Net::HTTP.new(uri.host, uri.port) resp = req.post(uri.to_s, self.content.content, { 'Content-type' => "#{self.mime_type};charset=utf-8", 'Content-Length' => self.content.content.size.to_s }) raise "URL '#{uri}' returned code #{resp.code}" unless resp.code == "200" self.content.content.rewind if self.content.content.respond_to?(:rewind) extracted_text = JSON.parse(resp.body)[''].rstrip full_text.content = extracted_text if extracted_text.present? rescue => e logger.error("Error extracting content from #{self.id}: #{e.inspect}") end end end end
Version data entries
11 entries across 11 versions & 2 rubygems