Class: Relaton::Plateau::Fetcher

Inherits:
Object
  • Object
show all
Defined in:
lib/relaton/plateau/fetcher.rb

Overview

Fetcher class to fetch data from the Plateau website

Constant Summary collapse

HANDBOOKS_URL = 
"https://www.mlit.go.jp/plateau/_next/data/1.3.0/libraries/handbooks.json".freeze
TECHNICAL_REPORTS_URL = 
"https://www.mlit.go.jp/plateau/_next/data/1.3.0/libraries/technical-reports.json".freeze

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(output, format) ⇒ Fetcher

Returns a new instance of Fetcher.



13
14
15
16
17
18
 
# File 'lib/relaton/plateau/fetcher.rb', line 13

def initialize(output, format)
  @output = output
  @format = format
  @ext = format.sub(/^bib/, "")
  @files = []
end

Class Method Details

.fetch(source, output: "data", format: "yaml") ⇒ Object 



24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
 
# File 'lib/relaton/plateau/fetcher.rb', line 24

def self.fetch(source, output: "data", format: "yaml")
  t1 = Time.now
  puts "Started at: #{t1}"
  FileUtils.mkdir_p output

  if source == "plateau-handbooks"
    new(output, format).extract_handbooks_data
  elsif source == "plateau-technical-reports"
    new(output, format).extract_technical_reports_data
  else
    puts "Invalid source: #{source}"
  end

  t2 = Time.now
  puts "Stopped at: #{t2}"
  puts "Done in: #{(t2 - t1).round} sec."
end

Instance Method Details

#create_request(uri) ⇒ Object

Create a GET request with custom headers to mimic a browser



55
56
57
58
59
60
61
62
63
64
65
66
 
# File 'lib/relaton/plateau/fetcher.rb', line 55

def create_request(uri)
  request = Net::HTTP::Get.new(uri)
  request["User-Agent"] = "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:127.0) Gecko/20100101 Firefox/127.0"
  request["Accept"] = "*/*"
  request["Accept-Language"] = "en-US,en;q=0.5"
  request["Accept-Encoding"] = "gzip, deflate, br, zstd"
  request["Referer"] = "https://www.mlit.go.jp/plateau/libraries/"
  request["purpose"] = "prefetch"
  request["x-nextjs-data"] = "1"
  request["Connection"] = "keep-alive"
  request
end

#extract_handbooks_data ⇒ Object

Extract data for handbooks



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
 
# File 'lib/relaton/plateau/fetcher.rb', line 112

def extract_handbooks_data
  data = fetch_json_data(HANDBOOKS_URL)
  Util.info "Extracting handbooks data..."
  data["pageProps"]["handbooks"]["nodes"].each do |entry|
    handbook = entry["handbook"]
    versions = handbook["versions"]

    description_parts = handbook["description"]&.split("<br />") || ["", ""]
    title_en = description_parts[0].strip if description_parts[0]
    abstract = description_parts[1].strip if description_parts[1]

    doctype = entry["slug"].match("-") ?  "annex" : "handbook"

    versions.each do |version|
      item = HandbookParser.new(
        version: version, entry: entry, title_en: title_en, abstract: abstract, doctype: doctype
      ).parse
      save_document(item)

      # ::Relaton::Plateau::BibItem.new(
      #   pubid: "PLATEAU Handbook ##{entry["slug"]}",
      #   title_jp: handbook["title"],
      #   title_en: title_en,
      #   abstract_jp: abstract_jp,
      #   cover: "https://www.mlit.go.jp/#{handbook["thumbnail"]["mediaItemUrl"]}",
      #   type: document_type,
      #   publication_date: Date.parse(version["date"].gsub(".", "-")),
      #   url_pdf: version["pdf"],
      #   url_html: version["html"],
      #   filesize: version["filesize"].to_i,
      #   edition_number: version["title"].match(/\d\.\d/)[0],
      #   edition_text: version["title"],
      #   # tags: [],
      # )
    end
  end
  index.save
end

#extract_technical_reports_data ⇒ Object

Extract data for technical reports



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
179
 
# File 'lib/relaton/plateau/fetcher.rb', line 154

def extract_technical_reports_data
  data = fetch_json_data(TECHNICAL_REPORTS_URL)
  Util.info "Extracting technical reports data..."
  data["pageProps"]["nodes"].map do |entry|
    save_document(TechnicalReportParser.new(entry).parse)

    # technical_report = entry["technicalReport"]

    # ::Relaton::Plateau::BibItem.new(
    #   title_jp: technical_report["title"],
    #   abstract_jp: technical_report["subtitle"],
    #   cover: "https://www.mlit.go.jp/#{technical_report["thumbnail"]["mediaItemUrl"]}",
    #   pubid: "PLATEAU Tech Report ##{entry["slug"]}",
    #   type: "technical-report",
    #   subtype: entry["technicalReportCategories"]["nodes"].map { |cat| cat["name"] },
    #   publication_date: Date.parse(entry["date"]),
    #   url_pdf: technical_report["pdf"],
    #   filesize: technical_report["filesize"].to_i,
    #   edition_number: "1.0",
    #   edition_text: "1.0",
    #   tags: entry["globalTags"]["nodes"].map { |tag| tag["name"] },
    # )

  end
  index.save
end

#fetch_json_data(url) ⇒ Hash

Fetch JSON data from a URL with custom headers

Parameters:

  • url (String) —

    The URL to fetch JSON data from

Returns:

  • (Hash) —

    The parsed JSON data



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
 
# File 'lib/relaton/plateau/fetcher.rb', line 83

def fetch_json_data(url)
  uri = URI(url)

  request = create_request(uri)

  # Send the request and get the response
  response = Net::HTTP.start(uri.hostname, uri.port, use_ssl: true) do |http|
    http.request(request)
  end

  # Check if the response is successful
  unless response.code.to_i == 200
    Util.warn "Failed to fetch data: #{response.code} #{response.message}"
    return {}
  end

  body = hadle_response(response)

  # Parse the JSON response
  JSON.parse(body)
rescue StandardError => e
  # Handle any errors during the fetching process
  Util.error "Error fetching JSON data from #{url}: #{e.message}"
  {}
end

#file_name(id) ⇒ Object 



200
201
202
203
204
205
206
207
208
 
# File 'lib/relaton/plateau/fetcher.rb', line 200

def file_name(id)
  name = id.gsub(/\s+/, "_").gsub(/\W+/, "").downcase
  if id.match?(/民間活用編/)
    name += "_private"
  elsif id.match?(/公共活用編/)
    name += "_public"
  end
  File.join(@output, "#{name}.#{@ext}")
end

#hadle_response(response) ⇒ Object

Handle different content encodings



69
70
71
72
73
74
75
76
77
 
# File 'lib/relaton/plateau/fetcher.rb', line 69

def hadle_response(response)
  if response["Content-Encoding"] == "gzip"
    Zlib::GzipReader.new(StringIO.new(response.body)).read
  elsif response["Content-Encoding"] == "deflate"
    Zlib::Inflate.inflate(response.body)
  else
    response.body
  end
end

#index ⇒ Object 



20
21
22
 
# File 'lib/relaton/plateau/fetcher.rb', line 20

def index
  @index ||= Relaton::Index.find_or_create :plateau, file: "index-v1.yaml"
end

#save_document(item) ⇒ Object

def self.save_to_yaml(data, filename)

File.open(filename, "w") do |file|
  file.write(data.to_yaml)
end
puts "Data saved to #{filename}."

end



188
189
190
191
192
193
194
195
196
197
198
 
# File 'lib/relaton/plateau/fetcher.rb', line 188

def save_document(item)
  id = item.docidentifier.first.id
  file = file_name id
  if @files.include?(file)
    Util.warn "File #{file} already exists, skipping.", key: id
  else
    File.write(file, serialize(item))
    @files << file
    index.add_or_update id, file
  end
end

#serialize(item) ⇒ Object 



210
211
212
213
214
215
216
 
# File 'lib/relaton/plateau/fetcher.rb', line 210

def serialize(item)
  case @format
  when "yaml" then item.to_hash.to_yaml
  when "xml" then item.to_xml bibdata: true
  else item.send("to_#{@format}")
  end
end