1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
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
125
126
127
128
129
130
131
132
133
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
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
|
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
require_relative "streaming-reader"
require_relative "org/apache/arrow/flatbuf/block"
require_relative "org/apache/arrow/flatbuf/footer"
module ArrowFormat
class FileReader
include Enumerable
include Readable
MAGIC = "ARROW1".b.freeze
MAGIC_BUFFER = IO::Buffer.for(MAGIC)
START_MARKER_SIZE = MAGIC_BUFFER.size
END_MARKER_SIZE = MAGIC_BUFFER.size
# <magic number "ARROW1">
# <empty padding bytes [to 8 byte boundary]>
STREAMING_FORMAT_START_OFFSET = 8
CONTINUATION_BUFFER =
IO::Buffer.for(MessagePullReader::CONTINUATION_STRING)
FOOTER_SIZE_FORMAT = :s32
FOOTER_SIZE_SIZE = IO::Buffer.size_of(FOOTER_SIZE_FORMAT)
def initialize(input)
case input
when IO
@buffer = IO::Buffer.map(input, nil, 0, IO::Buffer::READONLY)
when String
@buffer = IO::Buffer.for(input)
else
@buffer = input
end
validate
@footer = read_footer
@record_batch_blocks = @footer.record_batches
@schema = read_schema(@footer.schema)
@dictionaries = read_dictionaries
end
def n_record_batches
@record_batch_blocks.size
end
def read(i)
fb_message, body = read_block(@record_batch_blocks[i])
fb_header = fb_message.header
unless fb_header.is_a?(Org::Apache::Arrow::Flatbuf::RecordBatch)
raise FileReadError.new(@buffer,
"Not a record batch message: #{i}: " +
fb_header.class.name)
end
read_record_batch(fb_header, @schema, body)
end
def each
return to_enum(__method__) {n_record_batches} unless block_given?
@record_batch_blocks.size.times do |i|
yield(read(i))
end
end
private
def validate
minimum_size = STREAMING_FORMAT_START_OFFSET +
FOOTER_SIZE_SIZE +
END_MARKER_SIZE
if @buffer.size < minimum_size
raise FileReadError.new(@buffer,
"Input must be larger than or equal to " +
"#{minimum_size}: #{@buffer.size}")
end
start_marker = @buffer.slice(0, START_MARKER_SIZE)
if start_marker != MAGIC_BUFFER
raise FileReadError.new(@buffer, "No start marker")
end
end_marker = @buffer.slice(@buffer.size - END_MARKER_SIZE,
END_MARKER_SIZE)
if end_marker != MAGIC_BUFFER
raise FileReadError.new(@buffer, "No end marker")
end
end
def read_footer
footer_size_offset = @buffer.size - END_MARKER_SIZE - FOOTER_SIZE_SIZE
footer_size = @buffer.get_value(FOOTER_SIZE_FORMAT, footer_size_offset)
footer_data = @buffer.slice(footer_size_offset - footer_size,
footer_size)
Org::Apache::Arrow::Flatbuf::Footer.new(footer_data)
end
def read_block(block)
offset = block.offset
# If we can report property error information, we can use
# MessagePullReader here.
#
# message_pull_reader = MessagePullReader.new do |message, body|
# return read_record_batch(message.header, @schema, body)
# end
# chunk = @buffer.slice(offset,
# MessagePullReader::CONTINUATION_SIZE +
# MessagePullReader::METADATA_LENGTH_SIZE +
# block.meta_data_length +
# block.body_length)
# message_pull_reader.consume(chunk)
continuation_size = CONTINUATION_BUFFER.size
continuation = @buffer.slice(offset, continuation_size)
unless continuation == CONTINUATION_BUFFER
raise FileReadError.new(@buffer,
"Invalid continuation: #{i}: " +
continuation.inspect)
end
offset += continuation_size
metadata_length_type = MessagePullReader::METADATA_LENGTH_TYPE
metadata_length_size = MessagePullReader::METADATA_LENGTH_SIZE
metadata_length = @buffer.get_value(metadata_length_type, offset)
expected_metadata_length =
block.meta_data_length -
continuation_size -
metadata_length_size
unless metadata_length == expected_metadata_length
raise FileReadError.new(@buffer,
"Invalid metadata length #{i}: " +
"expected:#{expected_metadata_length} " +
"actual:#{metadata_length}")
end
offset += metadata_length_size
metadata = @buffer.slice(offset, metadata_length)
fb_message = Org::Apache::Arrow::Flatbuf::Message.new(metadata)
offset += metadata_length
body = @buffer.slice(offset, block.body_length)
[fb_message, body]
end
def read_dictionaries
dictionary_blocks = @footer.dictionaries
return nil if dictionary_blocks.nil?
dictionary_fields = {}
@schema.fields.each do |field|
next unless field.type.is_a?(DictionaryType)
dictionary_fields[field.dictionary_id] = field
end
dictionaries = {}
dictionary_blocks.each do |block|
fb_message, body = read_block(block)
fb_header = fb_message.header
unless fb_header.is_a?(Org::Apache::Arrow::Flatbuf::DictionaryBatch)
raise FileReadError.new(@buffer,
"Not a dictionary batch message: " +
fb_header.inspect)
end
id = fb_header.id
if fb_header.delta?
unless dictionaries.key?(id)
raise FileReadError.new(@buffer,
"A delta dictionary batch message " +
"must exist after a non delta " +
"dictionary batch message: " +
fb_header.inspect)
end
else
if dictionaries.key?(id)
raise FileReadError.new(@buffer,
"Multiple non delta dictionary batch " +
"messages for the same ID is invalid: " +
fb_header.inspect)
end
end
value_type = dictionary_fields[id].type.value_type
schema = Schema.new([Field.new("dummy", value_type, true, nil)])
record_batch = read_record_batch(fb_header.data, schema, body)
if fb_header.delta?
dictionaries[id] << record_batch.columns[0]
else
dictionaries[id] = [record_batch.columns[0]]
end
end
dictionaries
end
def find_dictionary(id)
@dictionaries[id]
end
end
end
|
