tapsoob 0.2.7-java
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +7 -0
- data/.gitignore +10 -0
- data/.rspec +2 -0
- data/Gemfile +44 -0
- data/README.md +73 -0
- data/Rakefile +4 -0
- data/bin/tapsoob +6 -0
- data/lib/tapsoob/chunksize.rb +53 -0
- data/lib/tapsoob/cli.rb +192 -0
- data/lib/tapsoob/config.rb +33 -0
- data/lib/tapsoob/data_stream.rb +350 -0
- data/lib/tapsoob/errors.rb +16 -0
- data/lib/tapsoob/log.rb +16 -0
- data/lib/tapsoob/operation.rb +482 -0
- data/lib/tapsoob/progress_bar.rb +236 -0
- data/lib/tapsoob/railtie.rb +11 -0
- data/lib/tapsoob/schema.rb +91 -0
- data/lib/tapsoob/utils.rb +184 -0
- data/lib/tapsoob/version.rb +4 -0
- data/lib/tapsoob.rb +6 -0
- data/lib/tasks/tapsoob.rake +95 -0
- data/spec/lib/tapsoob/chunksize_spec.rb +92 -0
- data/spec/lib/tapsoob/version_spec.rb +7 -0
- data/spec/spec_helper.rb +91 -0
- data/tapsoob.gemspec +37 -0
- metadata +143 -0
@@ -0,0 +1,350 @@
|
|
1
|
+
# -*- encoding : utf-8 -*-
|
2
|
+
require 'tapsoob/log'
|
3
|
+
require 'tapsoob/utils'
|
4
|
+
|
5
|
+
module Tapsoob
|
6
|
+
class DataStream
|
7
|
+
DEFAULT_CHUNKSIZE = 1000
|
8
|
+
|
9
|
+
attr_reader :db, :state
|
10
|
+
|
11
|
+
def initialize(db, state)
|
12
|
+
@db = db
|
13
|
+
@state = {
|
14
|
+
:offset => 0,
|
15
|
+
:avg_chunksize => 0,
|
16
|
+
:num_chunksize => 0,
|
17
|
+
:total_chunksize => 0
|
18
|
+
}.merge(state)
|
19
|
+
@state[:chunksize] ||= DEFAULT_CHUNKSIZE
|
20
|
+
@complete = false
|
21
|
+
end
|
22
|
+
|
23
|
+
def log
|
24
|
+
Tapsoob.log
|
25
|
+
end
|
26
|
+
|
27
|
+
def error=(val)
|
28
|
+
state[:error] = val
|
29
|
+
end
|
30
|
+
|
31
|
+
def error
|
32
|
+
state[:error] || false
|
33
|
+
end
|
34
|
+
|
35
|
+
def table_name
|
36
|
+
state[:table_name].to_sym
|
37
|
+
end
|
38
|
+
|
39
|
+
def table_name_sql
|
40
|
+
table_name
|
41
|
+
end
|
42
|
+
|
43
|
+
def to_hash
|
44
|
+
state.merge(:klass => self.class.to_s)
|
45
|
+
end
|
46
|
+
|
47
|
+
def to_json
|
48
|
+
JSON.generate(to_hash)
|
49
|
+
end
|
50
|
+
|
51
|
+
def string_columns
|
52
|
+
@string_columns ||= Tapsoob::Utils.incorrect_blobs(db, table_name)
|
53
|
+
end
|
54
|
+
|
55
|
+
def table
|
56
|
+
@table ||= db[table_name_sql]
|
57
|
+
end
|
58
|
+
|
59
|
+
def order_by(name=nil)
|
60
|
+
@order_by ||= begin
|
61
|
+
name ||= table_name
|
62
|
+
Tapsoob::Utils.order_by(db, name)
|
63
|
+
end
|
64
|
+
end
|
65
|
+
|
66
|
+
def increment(row_count)
|
67
|
+
state[:offset] += row_count
|
68
|
+
end
|
69
|
+
|
70
|
+
# keep a record of the average chunksize within the first few hundred thousand records, after chunksize
|
71
|
+
# goes below 100 or maybe if offset is > 1000
|
72
|
+
def fetch_rows
|
73
|
+
state[:chunksize] = fetch_chunksize
|
74
|
+
ds = table.order(*order_by).limit(state[:chunksize], state[:offset])
|
75
|
+
log.debug "DataStream#fetch_rows SQL -> #{ds.sql}"
|
76
|
+
rows = Tapsoob::Utils.format_data(ds.all,
|
77
|
+
:string_columns => string_columns,
|
78
|
+
:schema => db.schema(table_name),
|
79
|
+
:table => table_name
|
80
|
+
)
|
81
|
+
update_chunksize_stats
|
82
|
+
rows
|
83
|
+
end
|
84
|
+
|
85
|
+
def fetch_file(dump_path)
|
86
|
+
state[:chunksize] = fetch_chunksize
|
87
|
+
ds = JSON.parse(File.read(File.join(dump_path, "data", "#{table_name}.json")))
|
88
|
+
log.debug "DataStream#fetch_file"
|
89
|
+
rows = {
|
90
|
+
:header => ds["header"],
|
91
|
+
:data => ds["data"][state[:offset], (state[:offset] + state[:chunksize])] || [ ]
|
92
|
+
}
|
93
|
+
update_chunksize_stats
|
94
|
+
rows
|
95
|
+
end
|
96
|
+
|
97
|
+
def max_chunksize_training
|
98
|
+
20
|
99
|
+
end
|
100
|
+
|
101
|
+
def fetch_chunksize
|
102
|
+
chunksize = state[:chunksize]
|
103
|
+
return chunksize if state[:num_chunksize] < max_chunksize_training
|
104
|
+
return chunksize if state[:avg_chunksize] == 0
|
105
|
+
return chunksize if state[:error]
|
106
|
+
state[:avg_chunksize] > chunksize ? state[:avg_chunksize] : chunksize
|
107
|
+
end
|
108
|
+
|
109
|
+
def update_chunksize_stats
|
110
|
+
return if state[:num_chunksize] >= max_chunksize_training
|
111
|
+
state[:total_chunksize] += state[:chunksize]
|
112
|
+
state[:num_chunksize] += 1
|
113
|
+
state[:avg_chunksize] = state[:total_chunksize] / state[:num_chunksize] rescue state[:chunksize]
|
114
|
+
end
|
115
|
+
|
116
|
+
def encode_rows(rows)
|
117
|
+
Tapsoob::Utils.base64encode(Marshal.dump(rows))
|
118
|
+
end
|
119
|
+
|
120
|
+
def fetch(opts = {})
|
121
|
+
opts = (opts.empty? ? { :type => "database", :source => db.uri } : opts)
|
122
|
+
|
123
|
+
log.debug "DataStream#fetch state -> #{state.inspect}"
|
124
|
+
|
125
|
+
t1 = Time.now
|
126
|
+
rows = (opts[:type] == "file" ? fetch_file(opts[:source]) : fetch_rows)
|
127
|
+
encoded_data = encode_rows(rows)
|
128
|
+
t2 = Time.now
|
129
|
+
elapsed_time = t2 - t1
|
130
|
+
|
131
|
+
if opts[:type] == "file"
|
132
|
+
@complete = rows[:data] == [ ]
|
133
|
+
else
|
134
|
+
@complete = rows == { }
|
135
|
+
end
|
136
|
+
|
137
|
+
[encoded_data, (@complete ? 0 : rows[:data].size), elapsed_time]
|
138
|
+
end
|
139
|
+
|
140
|
+
def complete?
|
141
|
+
@complete
|
142
|
+
end
|
143
|
+
|
144
|
+
def fetch_database(dump_path)
|
145
|
+
params = fetch_from_database
|
146
|
+
encoded_data = params[:encoded_data]
|
147
|
+
json = params[:json]
|
148
|
+
|
149
|
+
rows = parse_encoded_data(encoded_data, json[:checksum])
|
150
|
+
|
151
|
+
@complete = rows == { }
|
152
|
+
|
153
|
+
# update local state
|
154
|
+
state.merge!(json[:state].merge(:chunksize => state[:chunksize]))
|
155
|
+
|
156
|
+
unless @complete
|
157
|
+
Tapsoob::Utils.export_rows(dump_path, table_name, rows)
|
158
|
+
state[:offset] += rows[:data].size
|
159
|
+
rows[:data].size
|
160
|
+
else
|
161
|
+
0
|
162
|
+
end
|
163
|
+
end
|
164
|
+
|
165
|
+
def fetch_from_database
|
166
|
+
res = nil
|
167
|
+
log.debug "DataStream#fetch_from_database state -> #{state.inspect}"
|
168
|
+
state[:chunksize] = Tapsoob::Utils.calculate_chunksize(state[:chunksize]) do |c|
|
169
|
+
state[:chunksize] = c.to_i
|
170
|
+
encoded_data = fetch.first
|
171
|
+
|
172
|
+
checksum = Tapsoob::Utils.checksum(encoded_data).to_s
|
173
|
+
|
174
|
+
res = {
|
175
|
+
:json => { :checksum => checksum, :state => to_hash },
|
176
|
+
:encoded_data => encoded_data
|
177
|
+
}
|
178
|
+
end
|
179
|
+
|
180
|
+
res
|
181
|
+
end
|
182
|
+
|
183
|
+
def fetch_data_in_database(params)
|
184
|
+
encoded_data = params[:encoded_data]
|
185
|
+
|
186
|
+
rows = parse_encoded_data(encoded_data, params[:checksum])
|
187
|
+
|
188
|
+
@complete = rows[:data] == [ ]
|
189
|
+
|
190
|
+
unless @complete
|
191
|
+
import_rows(rows)
|
192
|
+
rows[:data].size
|
193
|
+
else
|
194
|
+
0
|
195
|
+
end
|
196
|
+
end
|
197
|
+
|
198
|
+
def self.parse_json(json)
|
199
|
+
hash = JSON.parse(json).symbolize_keys
|
200
|
+
hash[:state].symbolize_keys! if hash.has_key?(:state)
|
201
|
+
hash
|
202
|
+
end
|
203
|
+
|
204
|
+
def parse_encoded_data(encoded_data, checksum)
|
205
|
+
raise Tapsoob::CorruptedData.new("Checksum Failed") unless Tapsoob::Utils.valid_data?(encoded_data, checksum)
|
206
|
+
|
207
|
+
begin
|
208
|
+
return Marshal.load(Tapsoob::Utils.base64decode(encoded_data))
|
209
|
+
rescue Object => e
|
210
|
+
unless ENV['NO_DUMP_MARSHAL_ERRORS']
|
211
|
+
puts "Error encountered loading data, wrote the data chunk to dump.#{Process.pid}.dat"
|
212
|
+
File.open("dump.#{Process.pid}.dat", "w") { |f| f.write(encoded_data) }
|
213
|
+
end
|
214
|
+
raise e
|
215
|
+
end
|
216
|
+
end
|
217
|
+
|
218
|
+
def import_rows(rows)
|
219
|
+
table.import(rows[:header], rows[:data], :commit_every => 100)
|
220
|
+
state[:offset] += rows[:data].size
|
221
|
+
rescue Exception => ex
|
222
|
+
case ex.message
|
223
|
+
when /integer out of range/ then
|
224
|
+
raise Tapsoob::InvalidData, <<-ERROR, []
|
225
|
+
\nDetected integer data that exceeds the maximum allowable size for an integer type.
|
226
|
+
This generally occurs when importing from SQLite due to the fact that SQLite does
|
227
|
+
not enforce maximum values on integer types.
|
228
|
+
ERROR
|
229
|
+
else raise ex
|
230
|
+
end
|
231
|
+
end
|
232
|
+
|
233
|
+
def verify_stream
|
234
|
+
state[:offset] = table.count
|
235
|
+
end
|
236
|
+
|
237
|
+
def self.factory(db, state)
|
238
|
+
if defined?(Sequel::MySQL) && Sequel::MySQL.respond_to?(:convert_invalid_date_time=)
|
239
|
+
Sequel::MySQL.convert_invalid_date_time = :nil
|
240
|
+
end
|
241
|
+
|
242
|
+
if state.has_key?(:klass)
|
243
|
+
return eval(state[:klass]).new(db, state)
|
244
|
+
end
|
245
|
+
|
246
|
+
if Tapsoob::Utils.single_integer_primary_key(db, state[:table_name].to_sym)
|
247
|
+
DataStreamKeyed.new(db, state)
|
248
|
+
else
|
249
|
+
DataStream.new(db, state)
|
250
|
+
end
|
251
|
+
end
|
252
|
+
end
|
253
|
+
|
254
|
+
class DataStreamKeyed < DataStream
|
255
|
+
attr_accessor :buffer
|
256
|
+
|
257
|
+
def initialize(db, state)
|
258
|
+
super(db, state)
|
259
|
+
@state = { :primary_key => order_by(state[:table_name]).first, :filter => 0 }.merge(@state)
|
260
|
+
@state[:chunksize] ||= DEFAULT_CHUNKSIZE
|
261
|
+
@buffer = []
|
262
|
+
end
|
263
|
+
|
264
|
+
def primary_key
|
265
|
+
state[:primary_key].to_sym
|
266
|
+
end
|
267
|
+
|
268
|
+
def buffer_limit
|
269
|
+
if state[:last_fetched] and state[:last_fetched] < state[:filter] and self.buffer.size == 0
|
270
|
+
state[:last_fetched]
|
271
|
+
else
|
272
|
+
state[:filter]
|
273
|
+
end
|
274
|
+
end
|
275
|
+
|
276
|
+
def calc_limit(chunksize)
|
277
|
+
# we want to not fetch more than is needed while we're
|
278
|
+
# inside sinatra but locally we can select more than
|
279
|
+
# is strictly needed
|
280
|
+
if defined?(Sinatra)
|
281
|
+
(chunksize * 1.1).ceil
|
282
|
+
else
|
283
|
+
(chunksize * 3).ceil
|
284
|
+
end
|
285
|
+
end
|
286
|
+
|
287
|
+
def load_buffer(chunksize)
|
288
|
+
# make sure BasicObject is not polluted by subsequent requires
|
289
|
+
Sequel::BasicObject.remove_methods!
|
290
|
+
|
291
|
+
num = 0
|
292
|
+
loop do
|
293
|
+
limit = calc_limit(chunksize)
|
294
|
+
# we have to use local variables in order for the virtual row filter to work correctly
|
295
|
+
key = primary_key
|
296
|
+
buf_limit = buffer_limit
|
297
|
+
ds = table.order(*order_by).filter { key.sql_number > buf_limit }.limit(limit)
|
298
|
+
log.debug "DataStreamKeyed#load_buffer SQL -> #{ds.sql}"
|
299
|
+
data = ds.all
|
300
|
+
self.buffer += data
|
301
|
+
num += data.size
|
302
|
+
if data.size > 0
|
303
|
+
# keep a record of the last primary key value in the buffer
|
304
|
+
state[:filter] = self.buffer.last[ primary_key ]
|
305
|
+
end
|
306
|
+
|
307
|
+
break if num >= chunksize or data.size == 0
|
308
|
+
end
|
309
|
+
end
|
310
|
+
|
311
|
+
def fetch_buffered(chunksize)
|
312
|
+
load_buffer(chunksize) if self.buffer.size < chunksize
|
313
|
+
rows = buffer.slice(0, chunksize)
|
314
|
+
state[:last_fetched] = if rows.size > 0
|
315
|
+
rows.last[ primary_key ]
|
316
|
+
else
|
317
|
+
nil
|
318
|
+
end
|
319
|
+
rows
|
320
|
+
end
|
321
|
+
|
322
|
+
#def import_rows(rows)
|
323
|
+
# table.import(rows[:header], rows[:data])
|
324
|
+
#end
|
325
|
+
|
326
|
+
#def fetch_rows
|
327
|
+
# chunksize = state[:chunksize]
|
328
|
+
# Tapsoob::Utils.format_data(fetch_buffered(chunksize) || [],
|
329
|
+
# :string_columns => string_columns)
|
330
|
+
#end
|
331
|
+
|
332
|
+
def increment(row_count)
|
333
|
+
# pop the rows we just successfully sent off the buffer
|
334
|
+
@buffer.slice!(0, row_count)
|
335
|
+
end
|
336
|
+
|
337
|
+
def verify_stream
|
338
|
+
key = primary_key
|
339
|
+
ds = table.order(*order_by)
|
340
|
+
current_filter = ds.max(key.sql_number)
|
341
|
+
|
342
|
+
# set the current filter to the max of the primary key
|
343
|
+
state[:filter] = current_filter
|
344
|
+
# clear out the last_fetched value so it can restart from scratch
|
345
|
+
state[:last_fetched] = nil
|
346
|
+
|
347
|
+
log.debug "DataStreamKeyed#verify_stream -> state: #{state.inspect}"
|
348
|
+
end
|
349
|
+
end
|
350
|
+
end
|
@@ -0,0 +1,16 @@
|
|
1
|
+
# -*- encoding : utf-8 -*-
|
2
|
+
module Tapsoob
|
3
|
+
class BaseError < StandardError
|
4
|
+
attr_reader :original_backtrace
|
5
|
+
|
6
|
+
def initialize(message, opts = {})
|
7
|
+
@original_backtrace = opts.delete(:backtrace)
|
8
|
+
super(message)
|
9
|
+
end
|
10
|
+
end
|
11
|
+
|
12
|
+
class NotImplemented < BaseError; end
|
13
|
+
class DuplicatePrimaryKeyError < BaseError; end
|
14
|
+
class CorruptedData < BaseError; end
|
15
|
+
class InvalidData < BaseError; end
|
16
|
+
end
|
data/lib/tapsoob/log.rb
ADDED
@@ -0,0 +1,16 @@
|
|
1
|
+
# -*- encoding : utf-8 -*-
|
2
|
+
module Tapsoob
|
3
|
+
def self.log=(log)
|
4
|
+
@@log = log
|
5
|
+
end
|
6
|
+
|
7
|
+
def self.log
|
8
|
+
@@log ||= begin
|
9
|
+
require 'logger'
|
10
|
+
log = Logger.new($stderr)
|
11
|
+
log.level = Logger::ERROR
|
12
|
+
log.datetime_format = "%Y-%m-%d %H:%M:%S"
|
13
|
+
log
|
14
|
+
end
|
15
|
+
end
|
16
|
+
end
|