Class: Datahen::Scraper::RubyParserExecutor

Inherits:
Executor
  • Object
show all
Defined in:
lib/datahen/scraper/ruby_parser_executor.rb

Constant Summary collapse

FIND_OUTPUTS_RETRY_LIMIT = 
nil

Constants inherited from Executor

Executor::MAX_FIND_OUTPUTS_PER_PAGE

Instance Attribute Summary collapse

Attributes inherited from Executor

#filename, #gid, #job_id, #page

Class Method Summary collapse

Instance Method Summary collapse

Methods inherited from Executor

#clean_backtrace, #eval_with_context, #find_output, #find_outputs, #finish, #finisher_update, #get_content, #get_failed_content, #get_job_id, #init_global_page, #init_job_page, #init_page, #parsing_update, #remove_old_dups!, #remove_old_output_dups!, #remove_old_page_dups!, #save_outputs, #save_pages, #save_pages_and_outputs, #seeding_update

Methods included from Plugin::ContextExposer

#create_context, #expose_to, #exposed_env, #exposed_methods, #isolated_binding, #var_or_proc

Constructor Details

#initialize(options = {}) ⇒ RubyParserExecutor

Returns a new instance of RubyParserExecutor.



19
20
21
22
23
24
25
26
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 19

def initialize(options={})
  @filename = options.fetch(:filename) { raise "Filename is required"}
  @page = options.fetch(:page) { nil }
  @gid = (self.page || {})['gid'] || options.fetch(:gid) { raise "GID or a page with a GID is required"}
  @job_id = options.fetch(:job_id)
  @page_vars = options.fetch(:vars) { {} }
  @keep_outputs = !!(options.fetch(:keep_outputs) { false })
end

Instance Attribute Details

#limbo_self ⇒ Boollean

Limbo self page flag.

Returns:

  • (Boollean) 


15
16
17
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 15

def limbo_self
  @limbo_self
end

#refetch_self ⇒ Boollean

Note:

It is stronger than #reparse_self flag.

Refetch self page flag.

Returns:

  • (Boollean) 


8
9
10
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 8

def refetch_self
  @refetch_self
end

#reparse_self ⇒ Boollean

Note:

It is stronger than #limbo_self flag.

Reparse self page flag.

Returns:

  • (Boollean) 


12
13
14
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 12

def reparse_self
  @reparse_self
end

#save ⇒ Object

Returns the value of attribute save.



4
5
6
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 4

def save
  @save
end

Class Method Details

.exposed_methods ⇒ Object 



28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 28

def self.exposed_methods
  [
    :get_content,
    :get_failed_content,
    :content,
    :failed_content,
    :outputs,
    :pages,
    :save_pages,
    :save_outputs,
    :find_output,
    :find_outputs,
    :refetch,
    :reparse,
    :limbo,
    :finish,
    :still_alive
  ].freeze
end

Instance Method Details

#content ⇒ Object 



235
236
237
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 235

def content
  @content ||= get_content(job_id, gid)
end

#eval_parser_script(save = false) ⇒ Object 



185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 185

def eval_parser_script(save=false)
  update_parsing_starting_status

  proc = Proc.new do
    page = init_page
    outputs = []
    pages = []
    page = init_page_vars(page)
    self.refetch_self = false
    self.reparse_self = false
    self.limbo_self = false

    begin
      context = isolated_binding({
        outputs: outputs,
        pages: pages,
        page: page
      })
      eval_with_context filename, context
    rescue Error::SafeTerminateError => e
      # do nothing, this is fine
    rescue SyntaxError => e
      handle_error(e) if save
      raise e
    rescue => e
      handle_error(e) if save
      raise e
    end

    puts "=========== Parsing Executed ==========="
    begin
      save_pages_and_outputs(pages, outputs, :parsing) unless refetch_self
    rescue => e
      handle_error(e) if save
      raise e
    end

    if refetch_self
      update_parsing_status gid, :to_refetch
    elsif reparse_self
      update_parsing_status gid, :to_reparse
    elsif limbo_self
      update_parsing_status gid, :limbo
    else
      update_parsing_status gid, :done
    end
  end
  proc.call
end

#exec_parser(save = false) ⇒ Object 



48
49
50
51
52
53
54
55
56
57
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 48

def exec_parser(save=false)
  @save = save
  if save
    puts "Executing parser script"
  else
    puts "Trying parser script"
  end

  eval_parser_script(save)
end

#failed_content ⇒ Object 



239
240
241
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 239

def failed_content
  @failed_content ||= get_failed_content(job_id, gid)
end

#handle_error(e) ⇒ Object 



249
250
251
252
253
254
255
256
257
258
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 249

def handle_error(e)
  error = ["Parsing #{e.class}: #{e.to_s} (Job:#{job_id} GID:#{gid})",clean_backtrace(e.backtrace)].join("\n")

  parsing_update(
    job_id: job_id,
    gid: gid,
    parsing_status: :failed,
    log_error: error,
    parsing_try_limit: (page || {})['parsing_try_limit'])
end

#init_page_vars(page) ⇒ Object 



59
60
61
62
63
64
65
66
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 59

def init_page_vars(page)
  return self.page unless self.page.nil?

  if !@page_vars.nil? && !@page_vars.empty?
    page['vars'] = @page_vars
  end
  page
end

#limbo(page_gid) ⇒ Object

Raises:

  • (ArgumentError) 


176
177
178
179
180
181
182
183
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 176

def limbo page_gid
  raise ArgumentError.new("page_gid needs to be a String.") unless page_gid.is_a?(String)
  if page_gid == gid
    self.limbo_self = true
    raise Error::SafeTerminateError
  end
  limbo_page page_gid
end

#limbo_page(page_gid) ⇒ Object 



167
168
169
170
171
172
173
174
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 167

def limbo_page page_gid
  if save
    update_parsing_status page_gid, :limbo
    puts "Limbo page #{page_gid}"
  else
    puts "Would have limbo page #{page_gid}"
  end
end

#refetch(page_gid) ⇒ Object

Raises:

  • (ArgumentError) 


140
141
142
143
144
145
146
147
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 140

def refetch page_gid
  raise ArgumentError.new("page_gid needs to be a String.") unless page_gid.is_a?(String)
  if page_gid == gid
    self.refetch_self = true
    raise Error::SafeTerminateError
  end
  refetch_page page_gid
end

#refetch_page(page_gid) ⇒ Object 



131
132
133
134
135
136
137
138
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 131

def refetch_page page_gid
  if save
    update_parsing_status page_gid, :to_refetch
    puts "Refetch page #{page_gid}"
  else
    puts "Would have refetch page #{page_gid}"
  end
end

#reparse(page_gid) ⇒ Object

Raises:

  • (ArgumentError) 


158
159
160
161
162
163
164
165
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 158

def reparse page_gid
  raise ArgumentError.new("page_gid needs to be a String.") unless page_gid.is_a?(String)
  if page_gid == gid
    self.reparse_self = true
    raise Error::SafeTerminateError
  end
  reparse_page page_gid
end

#reparse_page(page_gid) ⇒ Object 



149
150
151
152
153
154
155
156
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 149

def reparse_page page_gid
  if save
    update_parsing_status page_gid, :to_reparse
    puts "Reparse page #{page_gid}"
  else
    puts "Would have reparse page #{page_gid}"
  end
end

#save_type ⇒ Object 



127
128
129
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 127

def save_type
  :parsing
end

#still_alive(page_gid = nil) ⇒ Object 



243
244
245
246
247
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 243

def still_alive page_gid = nil
  page_gid = gid if page_gid.nil?
  client = Client::JobPage.new()
  client.still_alive(job_id, page_gid)
end

#update_parsing_done_status ⇒ Object 



95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 95

def update_parsing_done_status
  return unless save

  response = parsing_update(
    job_id: job_id,
    gid: gid,
    parsing_status: :done)

  if response.code == 200
    puts "Page Parsing Done."
  else
    puts "Error: Unable to save Page Parsing Done Status to server: #{response.body}"
    raise "Unable to save Page Parsing Done Status to server: #{response.body}"
  end
end

#update_parsing_starting_status ⇒ Object 



77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 77

def update_parsing_starting_status
  return unless save

  response = parsing_update(
    job_id: job_id,
    gid: gid,
    parsing_status: :starting,
    keep_outputs: @keep_outputs
  )

  if response.code == 200
    puts "Page Parsing Status Updated."
  else
    puts "Error: Unable to save Page Parsing Status to server: #{response.body}"
    raise "Unable to save Page Parsing Status to server: #{response.body}"
  end
end

#update_parsing_status(page_gid, status) ⇒ Object 



111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 111

def update_parsing_status page_gid, status
  return unless save

  response = parsing_update(
    job_id: job_id,
    gid: page_gid,
    parsing_status: status)

  if response.code == 200
    puts "Page #{page_gid} status changed to #{status}."
  else
    puts "Error: Unable to change page  #{page_gid} status: #{response.body} to #{status}"
    raise "Unable to change page  #{page_gid} status: #{response.body} to #{status}"
  end
end

#update_to_server(opts = {}) ⇒ Object 



68
69
70
71
72
73
74
75
 
# File 'lib/datahen/scraper/ruby_parser_executor.rb', line 68

def update_to_server(opts = {})
  parsing_update(
    job_id: opts[:job_id],
    gid: opts[:gid],
    pages: opts[:pages],
    outputs: opts[:outputs],
    parsing_status: opts[:status])
end