Class: Tapsoob::DataStream

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

Direct Known Subclasses

DataStreamKeyed

Constant Summary collapse

DEFAULT_CHUNKSIZE = 
1000

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(db, state, opts = {}) ⇒ DataStream

Returns a new instance of DataStream.



11
12
13
14
15
16
17
18
19
20
21
22
 
# File 'lib/tapsoob/data_stream.rb', line 11

def initialize(db, state, opts = {})
  @db = db
  @state = {
    :offset          => 0,
    :avg_chunksize   => 0,
    :num_chunksize   => 0,
    :total_chunksize => 0
  }.merge(state)
  @state[:chunksize] ||= DEFAULT_CHUNKSIZE
  @options = opts
  @complete = false
end

Instance Attribute Details

#db ⇒ Object (readonly)

Returns the value of attribute db.



9
10
11
 
# File 'lib/tapsoob/data_stream.rb', line 9

def db
  @db
end

#options ⇒ Object (readonly)

Returns the value of attribute options.



9
10
11
 
# File 'lib/tapsoob/data_stream.rb', line 9

def options
  @options
end

#state ⇒ Object (readonly)

Returns the value of attribute state.



9
10
11
 
# File 'lib/tapsoob/data_stream.rb', line 9

def state
  @state
end

Class Method Details

.factory(db, state, opts) ⇒ Object 



270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
 
# File 'lib/tapsoob/data_stream.rb', line 270

def self.factory(db, state, opts)
  if defined?(Sequel::MySQL) && Sequel::MySQL.respond_to?(:convert_invalid_date_time=)
    Sequel::MySQL.convert_invalid_date_time = :nil
  end

  if state.has_key?(:klass)
    return eval(state[:klass]).new(db, state, opts)
  end

  if Tapsoob::Utils.single_integer_primary_key(db, state[:table_name].to_sym)
    DataStreamKeyed.new(db, state, opts)
  else
    DataStream.new(db, state, opts)
  end
end

.parse_json(json) ⇒ Object 



187
188
189
190
191
 
# File 'lib/tapsoob/data_stream.rb', line 187

def self.parse_json(json)
  hash = JSON.parse(json).symbolize_keys
  hash[:state].symbolize_keys! if hash.has_key?(:state)
  hash
end

Instance Method Details

#complete? ⇒ Boolean

Returns:

  • (Boolean) 


162
163
164
 
# File 'lib/tapsoob/data_stream.rb', line 162

def complete?
  state[:offset] >= state[:size]
end

#encode_rows(rows) ⇒ Object 



142
143
144
 
# File 'lib/tapsoob/data_stream.rb', line 142

def encode_rows(rows)
  Tapsoob::Utils.base64encode(Marshal.dump(rows))
end

#error ⇒ Object 



33
34
35
 
# File 'lib/tapsoob/data_stream.rb', line 33

def error
  state[:error] || false
end

#error=(val) ⇒ Object 



29
30
31
 
# File 'lib/tapsoob/data_stream.rb', line 29

def error=(val)
  state[:error] = val
end

#fetch(opts = {}) ⇒ Object 



146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
 
# File 'lib/tapsoob/data_stream.rb', line 146

def fetch(opts = {})
  opts = (opts.empty? ? { :type => "database", :source => db.uri } : opts)

  log.debug "DataStream#fetch state -> #{state.inspect}"

  t1 = Time.now
  rows = (opts[:type] == "file" ? fetch_file(opts[:source]) : fetch_rows)
  encoded_data = encode_rows(rows)
  t2 = Time.now
  elapsed_time = t2 - t1

  state[:offset] += (rows == {} ? 0 : rows[:data].size)

  [encoded_data, (rows == {} ? 0 : rows[:data].size), elapsed_time]
end

#fetch_data_from_database(params) {|rows| ... } ⇒ Object

Yields:

  • (rows) 


166
167
168
169
170
171
172
173
174
175
176
 
# File 'lib/tapsoob/data_stream.rb', line 166

def fetch_data_from_database(params)
  encoded_data = params[:encoded_data]

  rows = parse_encoded_data(encoded_data, params[:checksum])

  # update local state
  state.merge!(params[:state].merge(:chunksize => state[:chunksize]))

  yield rows if block_given?
  (rows == {} ? 0 : rows[:data].size)
end

#fetch_data_to_database(params) ⇒ Object 



178
179
180
181
182
183
184
185
 
# File 'lib/tapsoob/data_stream.rb', line 178

def fetch_data_to_database(params)
  encoded_data = params[:encoded_data]

  rows = parse_encoded_data(encoded_data, params[:checksum])
  
  import_rows(rows)
  (rows == {} ? 0 : rows[:data].size)
end

#fetch_file(dump_path) ⇒ Object 



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
 
# File 'lib/tapsoob/data_stream.rb', line 88

def fetch_file(dump_path)
  #state[:chunksize] = fetch_chunksize
  # Read NDJSON format - each line is a separate JSON chunk
  file_path = File.join(dump_path, "data", "#{table_name}.json")

  # Parse all chunks and combine them
  all_data = []
  table_name_val = nil
  header_val = nil
  types_val = nil

  File.readlines(file_path).each do |line|
    chunk = JSON.parse(line.strip)
    table_name_val ||= chunk["table_name"]
    header_val ||= chunk["header"]
    types_val ||= chunk["types"]
    all_data.concat(chunk["data"]) if chunk["data"]
  end

  # Apply skip-duplicates if needed
  all_data = all_data.uniq if @options[:"skip-duplicates"]

  state[:size] = all_data.size
  log.debug "DataStream#fetch_file"

  rows = {
    :table_name => table_name_val,
    :header     => header_val,
    :data       => (all_data[state[:offset], state[:chunksize]] || []),
    :types      => types_val
  }
  update_chunksize_stats
  rows
end

#fetch_rows ⇒ Object

keep a record of the average chunksize within the first few hundred thousand records, after chunksize goes below 100 or maybe if offset is > 1000



74
75
76
77
78
79
80
81
82
83
84
85
86
 
# File 'lib/tapsoob/data_stream.rb', line 74

def fetch_rows
  #state[:chunksize] = fetch_chunksize
  ds = table.order(*order_by).limit(state[:chunksize], state[:offset])
  state[:size] = table.count
  log.debug "DataStream#fetch_rows SQL -> #{ds.sql}"
  rows = Tapsoob::Utils.format_data(db, ds.all,
    :string_columns => string_columns,
    :schema => db.schema(table_name),
    :table => table_name
  )
  update_chunksize_stats
  rows
end

#import_rows(rows) ⇒ Object 



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
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
 
# File 'lib/tapsoob/data_stream.rb', line 207

def import_rows(rows)
  columns = rows[:header]
  data    = rows[:data]

  # Only import existing columns
  if table.columns.size != columns.size
    existing_columns        = table.columns.map(&:to_s)
    additional_columns      = columns - existing_columns
    additional_columns_idxs = additional_columns.map { |c| columns.index(c) }
    additional_columns_idxs.reverse.each do |idx|
      columns.delete_at(idx)
      rows[:types].delete_at(idx)
    end
    data.each_index { |didx| additional_columns_idxs.reverse.each { |idx| data[didx].delete_at(idx) } }
  end

  # Decode blobs
  if rows.has_key?(:types) && rows[:types].include?("blob")
    blob_indices = rows[:types].each_index.select { |idx| rows[:types][idx] == "blob" }
    data.each_index do |idx|
      blob_indices.each do |bi|
        data[idx][bi] = Sequel::SQL::Blob.new(Tapsoob::Utils.base64decode(data[idx][bi])) unless data[idx][bi].nil?
      end
    end
  end

  # Parse date/datetime/time columns
  if rows.has_key?(:types)
    %w(date datetime time).each do |type|
      if rows[:types].include?(type)
        type_indices = rows[:types].each_index.select { |idx| rows[:types][idx] == type }
        data.each_index do |idx|
          type_indices.each do |ti|
            data[idx][ti] = Sequel.send("string_to_#{type}".to_sym, data[idx][ti]) unless data[idx][ti].nil?
          end
        end
      end
    end
  end

  # Remove id column
  if @options[:"discard-identity"] && rows[:header].include?("id")
    columns = rows[:header] - ["id"]
    data    = data.map { |d| d[1..-1] }
  end

  table.import(columns, data, :commit_every => 100)
rescue Exception => ex
  case ex.message
  when /integer out of range/ then
    raise Tapsoob::InvalidData, <<-ERROR, []
  \nDetected integer data that exceeds the maximum allowable size for an integer type.
  This generally occurs when importing from SQLite due to the fact that SQLite does
  not enforce maximum values on integer types.
    ERROR
  else raise ex
  end
end

#increment(row_count) ⇒ Object 



68
69
70
 
# File 'lib/tapsoob/data_stream.rb', line 68

def increment(row_count)
  state[:offset] += row_count
end

#log ⇒ Object 



24
25
26
27
 
# File 'lib/tapsoob/data_stream.rb', line 24

def log
  Tapsoob.log.level = Logger::DEBUG if state[:debug]
  Tapsoob.log
end

#max_chunksize_training ⇒ Object 



123
124
125
 
# File 'lib/tapsoob/data_stream.rb', line 123

def max_chunksize_training
  20
end

#order_by(name = nil) ⇒ Object 



61
62
63
64
65
66
 
# File 'lib/tapsoob/data_stream.rb', line 61

def order_by(name=nil)
  @order_by ||= begin
    name ||= table_name
    Tapsoob::Utils.order_by(db, name)
  end
end

#parse_encoded_data(encoded_data, checksum) ⇒ Object

Raises:



193
194
195
196
197
198
199
200
201
202
203
204
205
 
# File 'lib/tapsoob/data_stream.rb', line 193

def parse_encoded_data(encoded_data, checksum)
  raise Tapsoob::CorruptedData.new("Checksum Failed") unless Tapsoob::Utils.valid_data?(encoded_data, checksum)

  begin
    return Marshal.load(Tapsoob::Utils.base64decode(encoded_data))
  rescue Object => e
    unless ENV['NO_DUMP_MARSHAL_ERRORS']
      puts "Error encountered loading data, wrote the data chunk to dump.#{Process.pid}.dat"
      File.open("dump.#{Process.pid}.dat", "w") { |f| f.write(encoded_data) }
    end
    raise e
  end
end

#string_columns ⇒ Object 



53
54
55
 
# File 'lib/tapsoob/data_stream.rb', line 53

def string_columns
  @string_columns ||= Tapsoob::Utils.incorrect_blobs(db, table_name)
end

#table ⇒ Object 



57
58
59
 
# File 'lib/tapsoob/data_stream.rb', line 57

def table
  @table ||= db[table_name_sql]
end

#table_name ⇒ Object 



37
38
39
 
# File 'lib/tapsoob/data_stream.rb', line 37

def table_name
  state[:table_name].to_sym
end

#table_name_sql ⇒ Object 



41
42
43
 
# File 'lib/tapsoob/data_stream.rb', line 41

def table_name_sql
  table_name
end

#to_hash ⇒ Object 



45
46
47
 
# File 'lib/tapsoob/data_stream.rb', line 45

def to_hash
  state.merge(:klass => self.class.to_s)
end

#to_json ⇒ Object 



49
50
51
 
# File 'lib/tapsoob/data_stream.rb', line 49

def to_json
  JSON.generate(to_hash)
end

#update_chunksize_stats ⇒ Object

def fetch_chunksize

chunksize = state[:chunksize]
return chunksize if state[:num_chunksize] < max_chunksize_training
return chunksize if state[:avg_chunksize] == 0
return chunksize if state[:error]
state[:avg_chunksize] > chunksize ? state[:avg_chunksize] : chunksize

end



135
136
137
138
139
140
 
# File 'lib/tapsoob/data_stream.rb', line 135

def update_chunksize_stats
  return if state[:num_chunksize] >= max_chunksize_training
  state[:total_chunksize] += state[:chunksize]
  state[:num_chunksize] += 1
  state[:avg_chunksize] = state[:total_chunksize] / state[:num_chunksize] rescue state[:chunksize]
end

#verify_stream ⇒ Object 



266
267
268
 
# File 'lib/tapsoob/data_stream.rb', line 266

def verify_stream
  state[:offset] = table.count
end