Class: Forki::PostScraper

Inherits:
Scraper
  • Object
show all
Defined in:
lib/forki/scrapers/post_scraper.rb

Overview

rubocop:disable Metrics/ClassLength

Instance Method Summary collapse

Methods inherited from Scraper

#download_image, #find_graphql_data_closure_index, #find_graphql_data_strings, #initialize

Constructor Details

This class inherits a constructor from Forki::Scraper

Instance Method Details

#check_if_post_is_image(graphql_objects) ⇒ Object 



63
64
65
66
67
68
 
# File 'lib/forki/scrapers/post_scraper.rb', line 63

def check_if_post_is_image(graphql_objects)
  graphql_objects.any? do |graphql_object|  # if any GraphQL objects contain the top-level keys above, return true
    true unless graphql_object.fetch("image", nil).nil? # so long as the associated values are not nil
    true unless graphql_object.fetch("currMedia", nil).nil?
  end
end

#check_if_post_is_in_comment_stream(graphql_objects) ⇒ Object 



70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
 
# File 'lib/forki/scrapers/post_scraper.rb', line 70

def check_if_post_is_in_comment_stream(graphql_objects)
  graphql_objects.find do |graphql_object|
    next unless graphql_object.key?("nodes")

    begin
      type = graphql_object["nodes"].first["comet_sections"]["content"]["story"]["attachments"].first["styles"]["attachment"]["media"]["__typename"]
    rescue StandardError
      # if there's an error just return false, since the structure is so specific checking the whole thing is a lot
      next
    end

    return true if type == "Video"
  end

  false
end

#check_if_post_is_reel(graphql_object) ⇒ Object 



51
52
53
54
55
56
57
58
59
60
61
 
# File 'lib/forki/scrapers/post_scraper.rb', line 51

def check_if_post_is_reel(graphql_object)
  return false unless graphql_object.key?("node")

  begin
    style_infos = graphql_object["node"]["comet_sections"]["content"]["story"]["attachments"].first["styles"]["attachment"]["style_infos"].first
  rescue NoMethodError # if the object doesn't match the attribute chain above, the line above will try to operate on nil
    return false
  end

  style_infos.include?("fb_shorts_story")
end

#check_if_post_is_video(graphql_objects) ⇒ Object 



47
48
49
 
# File 'lib/forki/scrapers/post_scraper.rb', line 47

def check_if_post_is_video(graphql_objects)
  graphql_objects.any? { |graphql_object| graphql_object.key?("is_live_streaming") || graphql_object.key?("video") || check_if_post_is_reel(graphql_object) }
end

#extract_image_post_data(graphql_object_array) ⇒ Object

Extracts data from an image post by parsing GraphQL strings as seen in the video post scraper above



222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
 
# File 'lib/forki/scrapers/post_scraper.rb', line 222

def extract_image_post_data(graphql_object_array)
  graphql_object_array.find { |graphql_object| graphql_object.key?("viewer_actor") && graphql_object.key?("display_comments") }
  curr_media_object = graphql_object_array.find { |graphql_object| graphql_object.key?("currMedia") }
  creation_story_object = graphql_object_array.find { |graphql_object| graphql_object.key?("creation_story") && graphql_object.key?("message") }

  feedback_object = graphql_object_array.find { |graphql_object| graphql_object.has_key?("comet_ufi_summary_and_actions_renderer") }["comet_ufi_summary_and_actions_renderer"]["feedback"]
  share_count_object = feedback_object.fetch("share_count", {})

  poster = creation_story_object["creation_story"]["comet_sections"]["actor_photo"]["story"]["actors"][0]

  reaction_counts = extract_reaction_counts(feedback_object["cannot_see_top_custom_reactions"]["top_reactions"])
  post_details = {
    id: curr_media_object["currMedia"]["id"],
    num_comments: feedback_object["comments_count_summary_renderer"]["feedback"]["total_comment_count"],
    num_shares: share_count_object.fetch("count", nil),
    reshare_warning: feedback_object["should_show_reshare_warning"],
    image_url: curr_media_object["currMedia"]["image"]["uri"],
    text: (creation_story_object["message"] || {}).fetch("text", nil),
    profile_link: poster["url"],
    created_at: curr_media_object["currMedia"]["created_time"],
    has_video: false
  }
  post_details[:image_file] = Forki.retrieve_media(post_details[:image_url])
  post_details[:reactions] = reaction_counts
  post_details
end

#extract_live_video_post_data_from_watch_page(graphql_strings) ⇒ Object

Extract data from live video post on the watch page



280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
 
# File 'lib/forki/scrapers/post_scraper.rb', line 280

def extract_live_video_post_data_from_watch_page(graphql_strings)
  creation_story_object = JSON.parse(graphql_strings.find { |graphql| (graphql.include? "comment_count") && \
                                                   (graphql.include? "creation_story") })["video"]["creation_story"]
  media_object = JSON.parse(graphql_strings.find { |graphql| graphql.include? "playable_url" })["video"]["creation_story"]["attachments"][0]["media"]
  video_permalink = creation_story_object["shareable"]["url"].delete("\\")
  reaction_counts = extract_reaction_counts(creation_story_object["feedback_context"]["feedback_target_with_context"]["cannot_see_top_custom_reactions"]["top_reactions"])

  post_details = {
    id: creation_story_object["shareable"]["id"],
    num_comments: creation_story_object["feedback_context"]["feedback_target_with_context"]["total_comment_count"],
    num_shares: nil,
    num_views: find_number_of_views, # as far as I can tell, this is never present for live videos
    reshare_warning: creation_story_object["feedback_context"]["feedback_target_with_context"]["should_show_reshare_warning"],
    video_preview_image_url: creation_story_object["attachments"][0]["media"]["preferred_thumbnail"]["image"]["uri"],
    video_url: (media_object.fetch("playable_url_quality_hd", nil) || media_object.fetch("playable_url", nil)).delete("\\"),
    text: creation_story_object["attachments"][0]["media"]["savable_description"]["text"],
    created_at: creation_story_object["attachments"][0]["media"]["publish_time"],
    profile_link: video_permalink[..video_permalink.index("/videos")],
    has_video: true
  }

  post_details[:video_preview_image_file] = Forki.retrieve_media(post_details[:video_preview_image_url])
  post_details[:video_file] = Forki.retrieve_media(post_details[:video_url])
  post_details[:reactions] = reaction_counts
  post_details
end

#extract_post_data(graphql_strings) ⇒ Object

Raises:



20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
 
# File 'lib/forki/scrapers/post_scraper.rb', line 20

def extract_post_data(graphql_strings)
  # Bail out of the post otherwise it gets stuck
  raise ContentUnavailableError unless is_post_available?

  graphql_objects = get_graphql_objects(graphql_strings)
  post_has_video = check_if_post_is_video(graphql_objects)
  post_has_image = check_if_post_is_image(graphql_objects)

  # There's a chance it may be embedded in a comment chain like this:
  # https://www.facebook.com/PlandemicMovie/posts/588866298398729/
  post_has_video_in_comment_stream = check_if_post_is_in_comment_stream(graphql_objects) if post_has_video == false

  if post_has_video
    extract_video_post_data(graphql_strings)
  elsif post_has_video_in_comment_stream
    extract_video_comment_post_data(graphql_objects)
  elsif post_has_image
    extract_image_post_data(graphql_objects)
  else
    raise UnhandledContentError
  end
end

#extract_reaction_counts(reactions_object) ⇒ Object

Returns a hash containing counts of each reaction to a post Takes the edges list and creates a dictionary for each element that looks like: 1234 Then merges the dictionaries with the inject call



310
311
312
313
314
315
316
 
# File 'lib/forki/scrapers/post_scraper.rb', line 310

def extract_reaction_counts(reactions_object)
  reactions_object["edges"].map do |reaction|
    {
      "num_#{reaction["node"]["localized_name"].downcase}s".to_sym => reaction["reaction_count"]
    }
  end.inject { |emoji_counts, count| emoji_counts.merge(count) }
end

#extract_video_comment_post_data(graphql_objects) ⇒ Object 



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
 
# File 'lib/forki/scrapers/post_scraper.rb', line 102

def extract_video_comment_post_data(graphql_objects)
  graphql_nodes = nil
  graphql_objects.find do |graphql_object|
    next unless graphql_object.key?("nodes")
    graphql_nodes = graphql_object["nodes"]

    break
  end

  media = graphql_nodes.first["comet_sections"]["content"]["story"]["attachments"].first["styles"]["attachment"]["media"]
  inital_feedback_object = graphql_nodes.first["comet_sections"]["feedback"]["story"]["feedback_context"]["feedback_target_with_context"]["ufi_renderer"]["feedback"]
  feedback_object = inital_feedback_object["comet_ufi_summary_and_actions_renderer"]["feedback"]

  post_details = {
    id: media["id"],
    num_comments: feedback_object["comment_count"]["total_count"],
    num_shares: feedback_object["share_count"]["count"],
    num_views: feedback_object["video_view_count"],
    reshare_warning: feedback_object["should_show_reshare_warning"],
    video_preview_image_url: media["preferred_thumbnail"]["image"]["uri"],
    video_url: media["playable_url_quality_hd"] || media["playable_url"],
    text: graphql_nodes.first["comet_sections"]["content"]["story"]["comet_sections"]["message"]["story"]["message"]["text"],
    created_at: media["publish_time"],
    profile_link: graphql_nodes.first["comet_sections"]["context_layout"]["story"]["comet_sections"]["actor_photo"]["story"]["actors"].first["url"],
    has_video: true
  }

  post_details[:video_preview_image_file] = Forki.retrieve_media(post_details[:video_preview_image_url])
  post_details[:video_file] = Forki.retrieve_media(post_details[:video_url])
  post_details[:reactions] = inital_feedback_object["comet_ufi_summary_and_actions_renderer"]["feedback"]["i18n_reaction_count"]
  post_details
end

#extract_video_post_data(graphql_strings) ⇒ Object

Unfortunately, there’s a taxonomy of video post types, all of which require different parsing methods Specifically, there are normal video posts, video posts from the watch page, and live video posts from the watch page The general strategy for extracting information from each type, though, is to find which of the 30-odd GraphQL strings are relevant After finding those GraphQL strings, we parse them into hashes and extract the information we need



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
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
 
# File 'lib/forki/scrapers/post_scraper.rb', line 139

def extract_video_post_data(graphql_strings)
  unless all("h1").find { |h1| h1.text.strip == "Watch" }.nil?
    return extract_video_post_data_from_watch_page(graphql_strings)  # If this is a "watch page" video
  end

  graphql_object_array = graphql_strings.map { |graphql_string| JSON.parse(graphql_string) }
  story_node_object = graphql_object_array.find { |graphql_object| graphql_object.key? "node" }&.fetch("node", nil) # user posted video
  story_node_object = story_node_object || graphql_object_array.find { |graphql_object| graphql_object.key? "nodes" }&.fetch("nodes")&.first # page posted video

  return extract_video_post_data_alternative(graphql_object_array) if story_node_object.nil?

  if story_node_object["comet_sections"]["content"]["story"]["attachments"].first["styles"]["attachment"].key?("media")
    video_object = story_node_object["comet_sections"]["content"]["story"]["attachments"].first["styles"]["attachment"]["media"]
    creation_date = video_object["publish_time"]
    # creation_date = video_object["video"]["publish_time"]
  elsif story_node_object["comet_sections"]["content"]["story"]["attachments"].first["styles"]["attachment"].key?("style_infos")
    # For "Reels" we need a separate way to parse this
    video_object = story_node_object["comet_sections"]["content"]["story"]["attachments"].first["styles"]["attachment"]["style_infos"].first["fb_shorts_story"]["short_form_video_context"]["playback_video"]
    creation_date = story_node_object["comet_sections"]["content"]["story"]["attachments"].first["styles"]["attachment"]["style_infos"].first["fb_shorts_story"]["creation_time"]
  else
    raise "Unable to parse video object"
  end

  feedback_object = story_node_object["comet_sections"]["feedback"]["story"]["feedback_context"]["feedback_target_with_context"]["ufi_renderer"]["feedback"]
  reaction_counts = extract_reaction_counts(feedback_object["comet_ufi_summary_and_actions_renderer"]["feedback"]["cannot_see_top_custom_reactions"]["top_reactions"])
  share_count_object = feedback_object.fetch("share_count", {})

  if story_node_object["comet_sections"]["content"]["story"]["comet_sections"].key? "message"
    text = story_node_object["comet_sections"]["content"]["story"]["comet_sections"]["message"]["story"]["message"]["text"]
  else
    text = ""
  end

  feedback_object["comment_list_renderer"]["feedback"]["comment_count"]["total_count"]
  num_comments = feedback_object.has_key?("comment_list_renderer") ? feedback_object["comment_list_renderer"]["feedback"]["comment_count"]["total_count"] : feedback_object["comment_count"]["total_count"]

  post_details = {
    id: video_object["id"],
    num_comments: num_comments,
    num_shares: share_count_object.fetch("count", nil),
    num_views: feedback_object["comet_ufi_summary_and_actions_renderer"]["feedback"]["video_view_count"],
    reshare_warning: feedback_object["comet_ufi_summary_and_actions_renderer"]["feedback"]["should_show_reshare_warning"],
    video_preview_image_url: video_object["preferred_thumbnail"]["image"]["uri"],
    video_url: video_object["playable_url_quality_hd"] || video_object["playable_url"],
    text: text,
    created_at: creation_date,
    profile_link: story_node_object["comet_sections"]["context_layout"]["story"]["comet_sections"]["actor_photo"]["story"]["actors"][0]["url"],
    has_video: true
  }
  post_details[:video_preview_image_file] = Forki.retrieve_media(post_details[:video_preview_image_url])
  post_details[:video_file] = Forki.retrieve_media(post_details[:video_url])
  post_details[:reactions] = reaction_counts
  post_details
end

#extract_video_post_data_alternative(graphql_object_array) ⇒ Object 



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
 
# File 'lib/forki/scrapers/post_scraper.rb', line 194

def extract_video_post_data_alternative(graphql_object_array)
  sidepane_object = graphql_object_array.find { |graphql_object| graphql_object.key?("tahoe_sidepane_renderer") }
  video_object = graphql_object_array.find { |graphql_object| graphql_object.keys == ["video"] }
  feedback_object = sidepane_object["tahoe_sidepane_renderer"]["video"]["feedback"]
  reaction_counts = extract_reaction_counts(sidepane_object["tahoe_sidepane_renderer"]["video"]["feedback"]["cannot_see_top_custom_reactions"]["top_reactions"])
  share_count_object = feedback_object.fetch("share_count", {})

  post_details = {
    id: video_object["id"],
    num_comments: feedback_object["comments_count_summary_renderer"]["feedback"]["total_comment_count"],
    num_shares: share_count_object.fetch("count", nil),
    num_views: feedback_object["video_view_count"],
    reshare_warning: feedback_object["should_show_reshare_warning"],
    video_preview_image_url: video_object["video"]["preferred_thumbnail"]["image"]["uri"],
    video_url: video_object["video"]["playable_url_quality_hd"] || video_object["video"]["playable_url"],
    text: sidepane_object["tahoe_sidepane_renderer"]["video"]["creation_story"]["comet_sections"]["message"]["story"]["message"]["text"],
    created_at: video_object["video"]["publish_time"],
    profile_link: sidepane_object["tahoe_sidepane_renderer"]["video"]["creation_story"]["comet_sections"]["actor_photo"]["story"]["actors"][0]["url"],
    has_video: true
  }

  post_details[:video_preview_image_file] = Forki.retrieve_media(post_details[:video_preview_image_url])
  post_details[:video_file] = Forki.retrieve_media(post_details[:video_url])
  post_details[:reactions] = reaction_counts
  post_details
end

#extract_video_post_data_from_watch_page(graphql_strings) ⇒ Object

Extract data from a non-live video post on the watch page



250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
 
# File 'lib/forki/scrapers/post_scraper.rb', line 250

def extract_video_post_data_from_watch_page(graphql_strings)
  return extract_live_video_post_data_from_watch_page(graphql_strings) if current_url.include?("live")
  video_object = graphql_strings.map { |g| JSON.parse(g) }.find { |x| x.key?("video") }
  creation_story_object = JSON.parse(graphql_strings.find { |graphql_string| (graphql_string.include?("creation_story")) && \
                                                        (graphql_string.include?("live_status")) })
  video_permalink = creation_story_object["creation_story"]["shareable"]["url"].delete("\\")
  media_object = video_object["video"]["story"]["attachments"][0]["media"]
  reaction_counts = extract_reaction_counts(creation_story_object["feedback"]["cannot_see_top_custom_reactions"]["top_reactions"])

  post_details = {
    id: video_object["id"],
    num_comments: creation_story_object["feedback"]["total_comment_count"],
    num_shares: nil, # Not present for watch feed videos?
    num_views: creation_story_object["feedback"]["video_view_count_renderer"]["feedback"]["video_view_count"],
    reshare_warning: creation_story_object["feedback"]["should_show_reshare_warning"],
    video_preview_image_url: video_object["video"]["story"]["attachments"][0]["media"]["preferred_thumbnail"]["image"]["uri"],
    video_url: (media_object.fetch("playable_url_quality_hd", nil) || media_object.fetch("playable_url", nil)).delete("\\"),
    text: (creation_story_object["creation_story"]["message"] || {})["text"],
    created_at: video_object["video"]["story"]["attachments"][0]["media"]["publish_time"],
    profile_link: video_permalink[..video_permalink.index("/videos")],
    has_video: true
  }

  post_details[:video_preview_image_file] = Forki.retrieve_media(post_details[:video_preview_image_url])
  post_details[:video_file] = Forki.retrieve_media(post_details[:video_url])
  post_details[:reactions] = reaction_counts
  post_details
end

#find_number_of_views ⇒ Object

Searches the DOM to finds the number of times a (video) post has been viewed. Returns nil if it can’t find a DOM element with the view count



13
14
15
16
17
18
 
# File 'lib/forki/scrapers/post_scraper.rb', line 13

def find_number_of_views
  views_pattern = /[0-9MK, ]+Views/
  spans = all("span")
  views_span = spans.find { |s| s.text(:all) =~ views_pattern }
  extract_int_from_num_element(views_span)
end

#get_graphql_objects(graphql_strings) ⇒ Object 



43
44
45
 
# File 'lib/forki/scrapers/post_scraper.rb', line 43

def get_graphql_objects(graphql_strings)
  graphql_strings.map { |graphql_object| JSON.parse(graphql_object) }
end

#is_post_available? ⇒ Boolean

Returns:

  • (Boolean) 


87
88
89
90
91
92
93
94
95
96
97
98
99
100
 
# File 'lib/forki/scrapers/post_scraper.rb', line 87

def is_post_available?
  begin
    # This Video Isn't Available Anymore
    find("span", wait: 5, text: "content isn't available", exact_text: false)
  rescue Capybara::ElementNotFound, Selenium::WebDriver::Error::StaleElementReferenceError
    begin
      find("span", wait: 5, text: "This Video Isn't Available Anymore", exact_text: false)
    rescue Capybara::ElementNotFound, Selenium::WebDriver::Error::StaleElementReferenceError
      return true
    end
  end

  false
end

#parse(url) ⇒ Object

Uses GraphQL data and DOM elements to collect information about the current post



330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
 
# File 'lib/forki/scrapers/post_scraper.rb', line 330

def parse(url)
  validate_and_load_page(url)
  graphql_strings = find_graphql_data_strings(page.html)
  post_data = extract_post_data(graphql_strings)
  post_data[:url] = url
  user_url = post_data[:profile_link]

  5.times do
    begin
      post_data[:screenshot_file] = take_screenshot
      break
    rescue Net::ReadTimeout; end

    sleep(5)
  end

  # page.quit # Close browser between page navigations to prevent cache folder access issues

  post_data[:user] = User.lookup(user_url).first
  page.quit

  post_data
rescue Net::ReadTimeout
  # Eat it?
rescue StandardError => e
  raise e
ensure
  page.quit
end

#take_screenshot ⇒ Object 



318
319
320
321
322
323
324
325
326
327
 
# File 'lib/forki/scrapers/post_scraper.rb', line 318

def take_screenshot
  # First check whether post being scraped has a fact check overlay. If it does clear it.
  begin
    find('div[aria-label=" See Photo "]').click() || find('div[aria-label=" See Video "]').click()
  rescue Capybara::ElementNotFound
    # Do nothing if element not found
  end

  save_screenshot("#{Forki.temp_storage_location}/facebook_screenshot_#{SecureRandom.uuid}.png")
end