unpaywall.rb

require 'csv'
require 'faraday'
require 'faraday_middleware'
require 'pry'

# Process DOIs from Altmetrics CSV file
altmetric_csv_file = 'Altmetric - Mentions - University of York - 2019-04-16.csv'

# Store all processed unpaywall responses in new csv file
unpaywall_csv_file = 'Unpaywall_York.csv'
csv = CSV.open(unpaywall_csv_file, 'wb')
csv << ['doi', 'is_oa', 'best_oa_evidence', 'best_oa_host_type', 'best_oai_license',
            'best_oa_url_for_pdf', 'best_oa_version', 'journal_is_oa', 'genre',
            'journal_name', 'oa_status', 'published_date', 'publisher']

# Unpaywall connection
email = 'sebastian.palucha@gmail.com'
unpaywall = Faraday.new(url: 'https://api.unpaywall.org/v2') do |conn|
  conn.response :json, content_type: /\bjson$/
  conn.adapter Faraday.default_adapter
end

# For each DOI, call Unpaywall API and save returned data
CSV.foreach(altmetric_csv_file,
            encoding: 'windows-1251:utf-8',
            col_sep: ',',
            row_sep: :auto,
            headers: true) do |row|
  # Call Unpaywall API with DOI from Altmetric CSV file
  begin
    response = unpaywall.get row['DOI'], email: email
  rescue URI::InvalidURIError => e
    puts "Error #{e} while processing #{row['DOI']}"
    next
  end
  # binding.pry
  # Process Unpaywall response
  hash = {}
  hash['doi'] = response.body['doi'] # response.bidy returns hash data
  hash['is_oa'] = response.body['is_oa']
  if hash['is_oa'] == true
    # Use first set of data e.g. Unpaywall mark it as best_oa
    hash['best_oa_evidence'] = response.body['oa_locations'][0]['evidence']
    hash['best_oa_host_type'] = response.body['oa_locations'][0]['host_type']
    hash['best_oai_license'] = response.body['oa_locations'][0]['license']
    hash['best_oa_url_for_pdf'] = response.body['oa_locations'][0]['url_for_pdf']
    hash['best_oa_version'] = response.body['oa_locations'][0]['version']
  end
  # Other information about an original paper
  hash['journal_is_oa'] = response.body['journal_is_oa']
  hash['genre'] = response.body['genre']
  hash['journal_name'] = response.body['journal_name']
  hash['oa_status'] = response.body['oa_status']
  hash['published_date'] = response.body['published_date']
  hash['publisher'] = response.body['publisher']
  csv << hash.values  # Save processed data
end

csv.close