Class: Stats

Inherits:
Object
  • Object
show all
Defined in:
lib/stats.rb

Instance Method Summary collapse

Constructor Details

#initialize(options) ⇒ Stats

Returns a new instance of Stats.



3
4
5
6
# File 'lib/stats.rb', line 3

def initialize(options)
  @full_redis = Redis.new(options[:redis_options])
  @redis = NamespacedRedis.new(options[:redis_options], "cobweb-#{options[:crawl_id]}")
end

Instance Method Details

#end_crawl(options) ⇒ Object



16
17
18
19
20
# File 'lib/stats.rb', line 16

def end_crawl(options)
  @full_redis.srem "cobweb_crawls", options[:crawl_id]
  @redis.hset "statistics", "current_status", "Crawl Stopped"
  @redis.del "crawl_details"
end

#get_statisticsObject



163
164
165
# File 'lib/stats.rb', line 163

def get_statistics
  @redis.hgetall "statistics"
end

#get_statusObject



171
172
173
# File 'lib/stats.rb', line 171

def get_status
  @redis.hget "statistics", "current_status"
end

#increment_time_stat(stat_name, type = "minute", duration = 60) ⇒ Object



143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
# File 'lib/stats.rb', line 143

def increment_time_stat(stat_name, type="minute", duration=60)
  key = DateTime.now.strftime("%Y-%m-%d %H:%M")
  if type == "hour"
    key = DateTime.now.strftime("%Y-%m-%d %H:00")
  end
  minute_count = @redis.hget(stat_name, key).to_i
  if minute_count.nil?
    @redis.hset stat_name, key, 1
  else
    @redis.hset stat_name, key, minute_count + 1
  end
  #clear up older data
  @redis.hgetall(stat_name).keys.each do |key|
    if DateTime.parse(key) < DateTime.now-(duration/1440.0)
      puts "Deleting #{stat_name} - #{key}"
      @redis.hdel(stat_name, key)
    end
  end
end

#record_time_stat(stat_name, value, type = "minute", duration = 60) ⇒ Object



126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
# File 'lib/stats.rb', line 126

def record_time_stat(stat_name, value, type="minute", duration=60)
  key = DateTime.now.strftime("%Y-%m-%d %H:%M")
  if type == "hour"
    key = DateTime.now.strftime("%Y-%m-%d %H:00")
  end
  stat_value = @redis.hget(stat_name, key).to_i
  stat_count = @redis.hget("#{stat_name}-count", key).to_i
  
  if minute_count.nil?
    @redis.hset stat_name, key, value
    @redis.hset "#{stat_name}-count", key, 1
  else
    @redis.hset stat_name, key, ((stat_value*stat_count) + value) / (stat_count+1)
    @redis.hset "#{stat_name}-count", key, stat_count+1
  end
end

#set_totalsObject



175
176
177
178
# File 'lib/stats.rb', line 175

def set_totals
  stats = get_statistics
  stats[:crawled] = @redis.smembers "crawled"
end

#start_crawl(options) ⇒ Object



8
9
10
11
12
13
14
# File 'lib/stats.rb', line 8

def start_crawl(options)
  @full_redis.sadd "cobweb_crawls", options[:crawl_id]
  options.keys.each do |key|
    @redis.hset "crawl_details", key, options[key]
  end
  @redis.hset "statistics", "current_status", "Crawl Starting..."
end

#update_statistics(content) ⇒ Object



22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
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
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
# File 'lib/stats.rb', line 22

def update_statistics(content)
  
  crawl_counter = @redis.scard("crawled").to_i
  queue_counter = @redis.scard("queued").to_i
  
  if @redis.hexists "statistics", "average_response_time"
    @redis.hset("statistics", "average_response_time", (((@redis.hget("statistics", "average_response_time").to_f*crawl_counter) + content[:response_time].to_f) / (crawl_counter + 1)))
  else
    @redis.hset("statistics", "average_response_time", content[:response_time].to_f)
  end
  @redis.hset "statistics", "maximum_response_time", content[:response_time].to_f if @redis.hget("statistics", "maximum_response_time").nil? or content[:response_time].to_f > @redis.hget("statistics", "maximum_response_time").to_f
  @redis.hset "statistics", "minimum_response_time", content[:response_time].to_f if @redis.hget("statistics", "minimum_response_time").nil? or content[:response_time].to_f < @redis.hget("statistics", "minimum_response_time").to_f
  if @redis.hexists "statistics", "average_length"
    @redis.hset("statistics", "average_length", (((@redis.hget("statistics", "average_length").to_i*crawl_counter) + content[:length].to_i) / (crawl_counter + 1)))
  else
    @redis.hset("statistics", "average_length", content[:length].to_i)
  end
  @redis.hset "statistics", "maximum_length", content[:length].to_i if @redis.hget("statistics", "maximum_length").nil? or content[:length].to_i > @redis.hget("statistics", "maximum_length").to_i
  @redis.hset "statistics", "minimum_length", content[:length].to_i if @redis.hget("statistics", "minimum_length").nil? or content[:length].to_i < @redis.hget("statistics", "minimum_length").to_i


  if content[:mime_type].include?("text/html") or content[:mime_type].include?("application/xhtml+xml")
    @redis.hset "statistics", "page_count", @redis.hget("statistics", "page_count").to_i + 1
    @redis.hset "statistics", "page_size", @redis.hget("statistics", "page_size").to_i + content[:length].to_i
    increment_time_stat("pages_count")
  else
    @redis.hset "statistics", "asset_count", @redis.hget("statistics", "asset_count").to_i + 1
    @redis.hset "statistics", "asset_size", @redis.hget("statistics", "asset_size").to_i + content[:length].to_i
    increment_time_stat("assets_count")
  end
  
  total_redirects = @redis.hget("statistics", "total_redirects").to_i
  @redis.hset "statistics", "total_redirects", 0 if total_redirects.nil?
  @redis.hset("statistics", "total_redirects", total_redirects += content[:redirect_through].count) unless content[:redirect_through].nil?

  @redis.hset "statistics", "crawl_counter", crawl_counter
  @redis.hset "statistics", "queue_counter", queue_counter
  
  total_length = @redis.hget("statistics", "total_length").to_i
  @redis.hset "statistics", "total_length", total_length + content[:length].to_i

  mime_counts = {}
  if @redis.hexists "statistics", "mime_counts"
    mime_counts = JSON.parse(@redis.hget("statistics", "mime_counts"))
    if mime_counts.has_key? content[:mime_type]
      mime_counts[content[:mime_type]] += 1
    else
      mime_counts[content[:mime_type]] = 1
    end
  else
    mime_counts = {content[:mime_type] => 1}
  end
  @redis.hset "statistics", "mime_counts", mime_counts.to_json

  # record mime categories stats
  if content[:mime_type].starts_with? "text"
    increment_time_stat("mime_text_count")
  elsif content[:mime_type].starts_with? "application"
    increment_time_stat("mime_application_count")
  elsif content[:mime_type].starts_with? "audio"
    increment_time_stat("mime_audio_count")
  elsif content[:mime_type].starts_with? "image"
    increment_time_stat("mime_image_count")
  elsif content[:mime_type].starts_with? "message"
    increment_time_stat("mime_message_count")
  elsif content[:mime_type].starts_with? "model"
    increment_time_stat("mime_model_count")
  elsif content[:mime_type].starts_with? "multipart"
    increment_time_stat("mime_multipart_count")
  elsif content[:mime_type].starts_with? "video"
    increment_time_stat("mime_video_count")
  end
  
  status_counts = {}
  if @redis.hexists "statistics", "status_counts"
    status_counts = HashUtil.deep_symbolize_keys(JSON.parse(@redis.hget("statistics", "status_counts")))
    status_code = content[:status_code].to_i.to_s.to_sym
    if status_counts.has_key? status_code
      status_counts[status_code] += 1
    else
      status_counts[status_code] = 1
    end      
  else
    status_counts = {status_code => 1}
  end
  
  # record statistics by status type
  if content[:status_code] >= 200 && content[:status_code] < 300
    increment_time_stat("status_200_count")
  elsif content[:status_code] >= 400 && content[:status_code] < 500
    increment_time_stat("status|_400_count")
  elsif content[:status_code] >= 500 && content[:status_code] < 600
    increment_time_stat("status|_500_count")
  end
  
  @redis.hset "statistics", "status_counts", status_counts.to_json
  
  
  ## time based statistics
  increment_time_stat("minute_totals", "minute", 60)    
  
  get_statistics
end

#update_status(status) ⇒ Object



167
168
169
# File 'lib/stats.rb', line 167

def update_status(status)
  @redis.hset "statistics", "current_status", status
end