Class: BentoSearch::ScopusEngine

Inherits:
Object
  • Object
show all
Extended by:
HTTPClientPatch::IncludeClient
Includes:
SearchEngine
Defined in:
app/search_engines/bento_search/scopus_engine.rb

Overview

Supports fielded searching, sorting, pagination.

Required configuration:

  • api_key

Defaults to ‘relevance’ sort, rather than scopus’s default of date desc.

Uses the Scopus SciVerse REST API. You need to be a Scopus customer to access. api.elsevier.com www.developers.elsevier.com/action/devprojects

ToS: www.developers.elsevier.com/devcms/content-policies “Federated Search” use case. Also: www.developers.elsevier.com/cms/apiserviceagreement

Note that ToS applying to you probably means you must restrict access to search functionality to authenticated affiliated users only.

Register for an API key at “Register New Site” at developers.elsevier.com/action/devnewsite You will then need to get server IP addresses registered with Scopus too, apparently by emailing directly to dave.santucci at elsevier dot com.

Scopus API Docs:

Query syntax and search fields:

Some more docs on response elements and query elements:

Other API’s in the suite not being used by this code at present:

Support: [email protected]

TODO: Mention to Scopus: Only one author? Paging of 50 gets an error, but docs say I should be able to request 200. q

Scopus response does not seem to include language of hit, even though api allows you to restrict by language. ask scopus if we’re missing something?

Constant Summary

Constants included from SearchEngine

BentoSearch::SearchEngine::DefaultPerPage

Class Method Summary collapse

Instance Method Summary collapse

Methods included from HTTPClientPatch::IncludeClient

include_http_client

Methods included from SearchEngine

#display_configuration, #engine_id, #fill_in_search_metadata_for, #initialize, #normalized_search_arguments, #public_settable_search_args, #search

Methods included from BentoSearch::SearchEngine::Capabilities

#search_keys, #semantic_search_keys, #semantic_search_map, #sort_keys

Class Method Details

.default_configurationObject



190
191
192
193
194
195
# File 'app/search_engines/bento_search/scopus_engine.rb', line 190

def self.default_configuration
  {
    :base_url => "http://api.elsevier.com/",
    :cluster => "SCOPUS"
  }
end

.required_configurationObject



186
187
188
# File 'app/search_engines/bento_search/scopus_engine.rb', line 186

def self.required_configuration
  ["api_key"]
end

Instance Method Details

#escape_query(query) ⇒ Object

The escaping rules are not entirely clear for the API. We know colons and parens are special chars. It’s unclear how or if we can escape them, we’ll just remove them.



177
178
179
180
181
182
183
# File 'app/search_engines/bento_search/scopus_engine.rb', line 177

def escape_query(query)
  # backslash escape doesn't seem to work
  #query.gsub(/([\\\(\)\:])/) do |match|
  #  "\\#{$1}"
  #end
  query.gsub(/([\\\(\)\:])/, ' ')
end

#max_per_pageObject

Max per-page is 200, as per www.developers.elsevier.com/devcms/content-apis, bottom of page.



198
199
200
# File 'app/search_engines/bento_search/scopus_engine.rb', line 198

def max_per_page
  200
end

#multi_field_search?Boolean

Returns:

  • (Boolean)


235
236
237
# File 'app/search_engines/bento_search/scopus_engine.rb', line 235

def multi_field_search?
  true
end

#search_field_definitionsObject



202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
# File 'app/search_engines/bento_search/scopus_engine.rb', line 202

def search_field_definitions
  {
    nil           => {:semantic => :general},
    "AUTH"        => {:semantic => :author},
    "TITLE"       => {:semantic => :title},
    # controlled and author-assigned keywords
    "KEY"         => {:semantic => :subject},
    "ISBN"        => {:semantic => :isbn},
    "ISSN"        => {:semantic => :issn},
    "VOLUME"      => {:semantic => :volume},
    "ISSUE"       => {:semantic => :issue},
    "PAGEFIRST"   => {:semantic => :start_page},
    # Should we use SRCTITLE instead? I think exact match might be better?
    "EXACTSRCTITLE" => {:semantic => :source_title},
    "DOI"         => {:semantic => :doi},
    "PUBYEAR"     => {:semantic => :year}
  }
end

#search_implementation(args) ⇒ Object



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
127
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
# File 'app/search_engines/bento_search/scopus_engine.rb', line 56

def search_implementation(args)
  results = Results.new

  xml, response, exception = nil, nil, nil

  url = scopus_url(args)

  begin
    response = http_client.get( url , nil,
      # HTTP headers.
      {"X-ELS-APIKey" => configuration.api_key,
      "X-ELS-ResourceVersion" => "XOCS",
      "Accept" => "application/atom+xml"}
    )

    xml = Nokogiri::XML(response.body)
  rescue BentoSearch::RubyTimeoutClass, HTTPClient::ConfigurationError, HTTPClient::BadResponseError, Nokogiri::SyntaxError  => e
    exception = e
  end

  # handle errors
  if (response.nil? || xml.nil? || exception ||
      (! HTTP::Status.successful? response.status) ||
      xml.at_xpath("service-error") ||
      xml.at_xpath("./atom:feed/atom:entry/atom:error", xml_ns)
      )

    # UGH. Scopus reports 0 hits as an error, not entirely distinguishable
    # from an actual error. Oh well, we have to go with it.
    if (
        (response.status == 400) &&
        xml &&
        (error_xml = xml.at_xpath("./service-error/status")) &&
        (node_text(error_xml.at_xpath("./statusCode")) == "INVALID_INPUT") &&
        (node_text(error_xml.at_xpath("./statusText")).starts_with? "Result set was empty")
      )
      # PROBABLY 0 hit count, although could be something else I'm afraid.
      results.total_items = 0
      return results
    elsif (
        (response.status == 200) &&
        xml &&
        (error_xml = xml.at_xpath("./atom:feed/atom:entry/atom:error", xml_ns)) &&
        (error_xml.text == "Result set was empty")
      )
      # NEW way of Scopus reporting an error that makes no sense either
      results.total_items = 0
      return results
    else
      # real error
      results.error ||= {}
      results.error[:exception] = e
      results.error[:status] = response.status if response
      results.error[:error_info] = xml.at_xpath("service-error").text.gsub!(/[\n\t ]+/, " ") if xml
      results.error[:error_info] ||= xml.at_xpath("./atom:feed/atom:entry/atom:error", xml_ns).text if xml
      return results
    end
  end


  results.total_items = (node_text xml.at_xpath("//opensearch:totalResults", xml_ns)).to_i

  xml.xpath("//atom:entry", xml_ns).each do | entry |

    results << (item = ResultItem.new)
    if scopus_link = entry.at_xpath("atom:link[@ref='scopus']", xml_ns)
      item.link = scopus_link["href"]
    end

    item.unique_id             = node_text entry.at_xpath("dc:identifier", xml_ns)

    item.title          = node_text entry.at_xpath("dc:title", xml_ns)
    item.journal_title  = node_text entry.at_xpath("prism:publicationName", xml_ns)
    item.issn           = node_text entry.at_xpath("prism:issn", xml_ns)
    item.volume         = node_text entry.at_xpath("prism:volume", xml_ns)
    item.issue          = node_text entry.at_xpath("prism:issueIdentifier", xml_ns)
    item.doi            = node_text entry.at_xpath("prism:doi", xml_ns)

    # pages might be in startingPage/endingPage OR in pageRange
    if (start = entry.at_xpath("prism:startingPage", xml_ns))
      item.start_page = start.text.to_i
      if ( epage = entry.at_xpath("prism:endingPage", xml_ns))
        item.end_page = epage.text.to_i
      end
    elsif (range = entry.at_xpath("prism:pageRange", xml_ns))
      (spage, epage) = *range.text().split("-")
      item.start_page = spage
      item.end_page = epage
    end

    # get the year out of the date
    if date = entry.at_xpath("prism:coverDate", xml_ns)
      date.text =~ /^(\d\d\d\d)/
      item.year = $1.to_i if $1
    end

    # Authors might be in atom:authors seperated by |, or just
    # a single one in dc:creator
    if (authors = entry.at_xpath("atom:authors", xml_ns))
      authors.text.split("|").each do |author|
        item.authors << Author.new(:display => author.strip)
      end
    elsif (author = entry.at_xpath("dc:creator", xml_ns))
      item.authors << Author.new(:display => author.text.strip)
    end

    # Format we're still trying to figure out how Scopus API
    # delivers it. Here is at at least one way.
    if (doctype = entry.at_xpath("atom:subtype", xml_ns))
      item.format     = doctype_to_format(doctype.text)
      item.format_str = doctype_to_string(doctype.text)
    end

  end

  return results
end

#sort_definitionsObject



221
222
223
224
225
226
227
228
229
230
231
232
233
# File 'app/search_engines/bento_search/scopus_engine.rb', line 221

def sort_definitions
  # scopus &sort= values, not yet URI-escaped, later code will do that.
  #
  # 'refeid' key is currently undocumented on Scopus site, but
  # was given to me in email by scopus.
  {
    "title_asc"     => {:implementation => "+itemtitle"},
    "date_desc"     => {:implementation => "-datesort,+auth"},
    "relevance"     => {:implementation => "refeid" },
    "author_asc"    => {:implementation => "+auth"},
    "num_cite_desc" => {:implementation => "-numcitedby"}
  }
end