Class: GooglePlayDevScraper::Scraper

Inherits:
ScraperBase show all
Defined in:
lib/googleplay_dev_scraper/scraper.rb

Overview

Google Play and google checkout scraper

Instance Attribute Summary

Attributes inherited from ScraperBase

#agent, #config

Instance Method Summary collapse

Methods inherited from ScraperBase

#setup, #try_get

Constructor Details

#initialize ⇒ Scraper

Returns a new instance of Scraper.



19
20
21
 
# File 'lib/googleplay_dev_scraper/scraper.rb', line 19

def initialize
  super
end

Instance Method Details

#body_string ⇒ Object 



23
24
25
 
# File 'lib/googleplay_dev_scraper/scraper.rb', line 23

def body_string
  @agent.page.body.force_encoding("UTF-8")
end

#dump_csv(csv_string) ⇒ Object

dump CSV (util)



114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
 
# File 'lib/googleplay_dev_scraper/scraper.rb', line 114

def dump_csv(csv_string)
  headers = nil
  CSV.parse(csv_string) do |row|
    unless headers
      headers = row
      next
    end

    i = 0
    row.each do |column|
      puts "#{headers[i]} : #{column}"
      i = i + 1
    end
    puts
  end
end

#get_appstats(package, start_day, end_day) ⇒ Object

Get application statistics CSV in zip

package

package name

start_day

start date (yyyyMMdd)

end_day

end date (yyyyMMdd)

Return

application statics zip data



100
101
102
103
104
105
106
107
108
109
110
111
 
# File 'lib/googleplay_dev_scraper/scraper.rb', line 100

def get_appstats(package, start_day, end_day)
  dim = "overall,country,language,os_version,device,app_version,carrier&met=active_device_installs,daily_device_installs,daily_device_uninstalls,daily_device_upgrades,active_user_installs,total_user_installs,daily_user_installs,daily_user_uninstalls,daily_avg_rating,total_avg_rating"
  url = "https://play.google.com/apps/publish/v2/statistics/download"
  url += "?package=#{package}"
  url += "&sd=#{start_day}&ed=#{end_day}"
  url += "&dim=#{dim}"
  #url += "&dev_acc=#{@config.dev_acc}"

  STDERR.puts "URL = #{url}"
  try_get(url)
  @agent.page.body
end

#get_estimated_sales_report(year, month) ⇒ Object

Get estimated sales report (report_type = sales_report)

year

Year (ex. 2012)

month

Month (1 - 12)

Return

CSV string



52
53
54
55
56
57
58
 
# File 'lib/googleplay_dev_scraper/scraper.rb', line 52

def get_estimated_sales_report(year, month)
  #https://play.google.com/apps/publish/v2/salesreport/download?report_date=2013_03&report_type=sales_report&dev_acc=09924472108471074593
  url = sprintf('https://play.google.com/apps/publish/v2/salesreport/download?report_date=%04d_%02d&report_type=sales_report&dev_acc=%s', year, month, @config.dev_acc)
  try_get(url)

  body_string
end

#get_order_list(start_time, end_time) ⇒ Object

Get order list

start_date

start time (DateTime)

end_date

end time (DateTime)

Return

CSV string



68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
 
# File 'lib/googleplay_dev_scraper/scraper.rb', line 68

def get_order_list(start_time, end_time)
  # unix time in ms
  start_ut = start_time.to_time.to_i * 1000
  end_ut = end_time.to_time.to_i * 1000

  try_get("https://wallet.google.com/merchant/pages/")
  if @agent.page.uri.path =~ /(bcid-[^\/]+)\/(oid-[^\/]+)\/(cid-[^\/]+)\//
    bcid = $1
    oid = $2
    cid = $3

    # You can check the URL with your browser.
    # (download csv file, and check download history with chrome/firefox)
    try_get("https://wallet.google.com/merchant/pages/" +
            bcid + "/" + oid + "/" + cid +
            "/purchaseorderdownload?startTime=#{start_ut}" + 
            "&endTime=#{end_ut}")
    body_string
  end
end

#get_sales_report(year, month) ⇒ Object

Get sales report (report_type = payout_report)

year

Year (ex. 2012)

month

Month (1 - 12)

Return

CSV string



35
36
37
38
39
40
41
 
# File 'lib/googleplay_dev_scraper/scraper.rb', line 35

def get_sales_report(year, month)
  #url = sprintf('https://play.google.com/apps/publish/salesreport/download?report_date=%04d_%02d&report_type=payout_report&dev_acc=%s', year, month, @config.dev_acc)
  url = sprintf('https://play.google.com/apps/publish/v2/salesreport/download?report_date=%04d_%02d&report_type=payout_report&dev_acc=%s', year, month, @config.dev_acc)
  try_get(url)

  body_string
end

#get_wallet_orders ⇒ Object

Get order list from wallet html page



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
 
# File 'lib/googleplay_dev_scraper/scraper.rb', line 134

def get_wallet_orders
  try_get("https://wallet.google.com/merchant/pages/")
  html = body_string

  doc = Nokogiri::HTML(html)

  #doc.xpath("//table[@id='purchaseOrderListTable']")

  result = ""

  doc.xpath("//tr[@class='orderRow']").each do |e|
    order_id = e['id']

    date = nil
    desc = nil
    total = nil
    status = nil

    e.children.each do |e2|
      case e2['class']
      when /wallet-date-column/
        date = e2.content
      when /wallet-description-column/
        desc = e2.content
      when /wallet-total-column/
        total = e2.content
      when /wallet-status-column/
        e3 = e2.children.first
        status = e3['title'] unless e3.nil?
      end
    end

    result += [order_id, date, desc, status, total].join(",") + "\n"
  end

  result
end