deimos-temp-fork 0.0.1
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/.circleci/config.yml +83 -0
- data/.gitignore +41 -0
- data/.gitmodules +0 -0
- data/.rspec +1 -0
- data/.rubocop.yml +333 -0
- data/.ruby-gemset +1 -0
- data/.ruby-version +1 -0
- data/CHANGELOG.md +349 -0
- data/CODE_OF_CONDUCT.md +77 -0
- data/Dockerfile +23 -0
- data/Gemfile +6 -0
- data/Gemfile.lock +286 -0
- data/Guardfile +22 -0
- data/LICENSE.md +195 -0
- data/README.md +1099 -0
- data/Rakefile +13 -0
- data/bin/deimos +4 -0
- data/deimos-ruby.gemspec +44 -0
- data/docker-compose.yml +71 -0
- data/docs/ARCHITECTURE.md +140 -0
- data/docs/CONFIGURATION.md +236 -0
- data/docs/DATABASE_BACKEND.md +147 -0
- data/docs/INTEGRATION_TESTS.md +52 -0
- data/docs/PULL_REQUEST_TEMPLATE.md +35 -0
- data/docs/UPGRADING.md +128 -0
- data/lib/deimos-temp-fork.rb +95 -0
- data/lib/deimos/active_record_consume/batch_consumption.rb +164 -0
- data/lib/deimos/active_record_consume/batch_slicer.rb +27 -0
- data/lib/deimos/active_record_consume/message_consumption.rb +79 -0
- data/lib/deimos/active_record_consume/schema_model_converter.rb +52 -0
- data/lib/deimos/active_record_consumer.rb +67 -0
- data/lib/deimos/active_record_producer.rb +87 -0
- data/lib/deimos/backends/base.rb +32 -0
- data/lib/deimos/backends/db.rb +41 -0
- data/lib/deimos/backends/kafka.rb +33 -0
- data/lib/deimos/backends/kafka_async.rb +33 -0
- data/lib/deimos/backends/test.rb +20 -0
- data/lib/deimos/batch_consumer.rb +7 -0
- data/lib/deimos/config/configuration.rb +381 -0
- data/lib/deimos/config/phobos_config.rb +137 -0
- data/lib/deimos/consume/batch_consumption.rb +150 -0
- data/lib/deimos/consume/message_consumption.rb +94 -0
- data/lib/deimos/consumer.rb +104 -0
- data/lib/deimos/instrumentation.rb +76 -0
- data/lib/deimos/kafka_message.rb +60 -0
- data/lib/deimos/kafka_source.rb +128 -0
- data/lib/deimos/kafka_topic_info.rb +102 -0
- data/lib/deimos/message.rb +79 -0
- data/lib/deimos/metrics/datadog.rb +47 -0
- data/lib/deimos/metrics/mock.rb +39 -0
- data/lib/deimos/metrics/provider.rb +36 -0
- data/lib/deimos/monkey_patches/phobos_cli.rb +35 -0
- data/lib/deimos/monkey_patches/phobos_producer.rb +51 -0
- data/lib/deimos/poll_info.rb +9 -0
- data/lib/deimos/producer.rb +224 -0
- data/lib/deimos/railtie.rb +8 -0
- data/lib/deimos/schema_backends/avro_base.rb +140 -0
- data/lib/deimos/schema_backends/avro_local.rb +30 -0
- data/lib/deimos/schema_backends/avro_schema_coercer.rb +119 -0
- data/lib/deimos/schema_backends/avro_schema_registry.rb +34 -0
- data/lib/deimos/schema_backends/avro_validation.rb +21 -0
- data/lib/deimos/schema_backends/base.rb +150 -0
- data/lib/deimos/schema_backends/mock.rb +42 -0
- data/lib/deimos/shared_config.rb +63 -0
- data/lib/deimos/test_helpers.rb +360 -0
- data/lib/deimos/tracing/datadog.rb +35 -0
- data/lib/deimos/tracing/mock.rb +40 -0
- data/lib/deimos/tracing/provider.rb +29 -0
- data/lib/deimos/utils/db_poller.rb +150 -0
- data/lib/deimos/utils/db_producer.rb +243 -0
- data/lib/deimos/utils/deadlock_retry.rb +68 -0
- data/lib/deimos/utils/inline_consumer.rb +150 -0
- data/lib/deimos/utils/lag_reporter.rb +175 -0
- data/lib/deimos/utils/schema_controller_mixin.rb +115 -0
- data/lib/deimos/version.rb +5 -0
- data/lib/generators/deimos/active_record/templates/migration.rb.tt +28 -0
- data/lib/generators/deimos/active_record/templates/model.rb.tt +5 -0
- data/lib/generators/deimos/active_record_generator.rb +79 -0
- data/lib/generators/deimos/db_backend/templates/migration +25 -0
- data/lib/generators/deimos/db_backend/templates/rails3_migration +31 -0
- data/lib/generators/deimos/db_backend_generator.rb +48 -0
- data/lib/generators/deimos/db_poller/templates/migration +11 -0
- data/lib/generators/deimos/db_poller/templates/rails3_migration +16 -0
- data/lib/generators/deimos/db_poller_generator.rb +48 -0
- data/lib/tasks/deimos.rake +34 -0
- data/spec/active_record_batch_consumer_spec.rb +481 -0
- data/spec/active_record_consume/batch_slicer_spec.rb +42 -0
- data/spec/active_record_consume/schema_model_converter_spec.rb +105 -0
- data/spec/active_record_consumer_spec.rb +154 -0
- data/spec/active_record_producer_spec.rb +85 -0
- data/spec/backends/base_spec.rb +10 -0
- data/spec/backends/db_spec.rb +54 -0
- data/spec/backends/kafka_async_spec.rb +11 -0
- data/spec/backends/kafka_spec.rb +11 -0
- data/spec/batch_consumer_spec.rb +256 -0
- data/spec/config/configuration_spec.rb +248 -0
- data/spec/consumer_spec.rb +209 -0
- data/spec/deimos_spec.rb +169 -0
- data/spec/generators/active_record_generator_spec.rb +56 -0
- data/spec/handlers/my_batch_consumer.rb +10 -0
- data/spec/handlers/my_consumer.rb +10 -0
- data/spec/kafka_listener_spec.rb +55 -0
- data/spec/kafka_source_spec.rb +381 -0
- data/spec/kafka_topic_info_spec.rb +111 -0
- data/spec/message_spec.rb +19 -0
- data/spec/phobos.bad_db.yml +73 -0
- data/spec/phobos.yml +77 -0
- data/spec/producer_spec.rb +498 -0
- data/spec/rake_spec.rb +19 -0
- data/spec/schema_backends/avro_base_shared.rb +199 -0
- data/spec/schema_backends/avro_local_spec.rb +32 -0
- data/spec/schema_backends/avro_schema_registry_spec.rb +32 -0
- data/spec/schema_backends/avro_validation_spec.rb +24 -0
- data/spec/schema_backends/base_spec.rb +33 -0
- data/spec/schemas/com/my-namespace/Generated.avsc +71 -0
- data/spec/schemas/com/my-namespace/MyNestedSchema.avsc +62 -0
- data/spec/schemas/com/my-namespace/MySchema-key.avsc +13 -0
- data/spec/schemas/com/my-namespace/MySchema.avsc +18 -0
- data/spec/schemas/com/my-namespace/MySchemaCompound-key.avsc +18 -0
- data/spec/schemas/com/my-namespace/MySchemaWithBooleans.avsc +18 -0
- data/spec/schemas/com/my-namespace/MySchemaWithDateTimes.avsc +33 -0
- data/spec/schemas/com/my-namespace/MySchemaWithId.avsc +28 -0
- data/spec/schemas/com/my-namespace/MySchemaWithUniqueId.avsc +32 -0
- data/spec/schemas/com/my-namespace/Wibble.avsc +43 -0
- data/spec/schemas/com/my-namespace/Widget.avsc +27 -0
- data/spec/schemas/com/my-namespace/WidgetTheSecond.avsc +27 -0
- data/spec/schemas/com/my-namespace/request/CreateTopic.avsc +11 -0
- data/spec/schemas/com/my-namespace/request/Index.avsc +11 -0
- data/spec/schemas/com/my-namespace/request/UpdateRequest.avsc +11 -0
- data/spec/schemas/com/my-namespace/response/CreateTopic.avsc +11 -0
- data/spec/schemas/com/my-namespace/response/Index.avsc +11 -0
- data/spec/schemas/com/my-namespace/response/UpdateResponse.avsc +11 -0
- data/spec/spec_helper.rb +267 -0
- data/spec/utils/db_poller_spec.rb +320 -0
- data/spec/utils/db_producer_spec.rb +514 -0
- data/spec/utils/deadlock_retry_spec.rb +74 -0
- data/spec/utils/inline_consumer_spec.rb +31 -0
- data/spec/utils/lag_reporter_spec.rb +76 -0
- data/spec/utils/platform_schema_validation_spec.rb +0 -0
- data/spec/utils/schema_controller_mixin_spec.rb +84 -0
- data/support/deimos-solo.png +0 -0
- data/support/deimos-with-name-next.png +0 -0
- data/support/deimos-with-name.png +0 -0
- data/support/flipp-logo.png +0 -0
- metadata +551 -0
@@ -0,0 +1,164 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require 'deimos/active_record_consume/batch_slicer'
|
4
|
+
require 'deimos/utils/deadlock_retry'
|
5
|
+
require 'deimos/message'
|
6
|
+
|
7
|
+
module Deimos
|
8
|
+
module ActiveRecordConsume
|
9
|
+
# Methods for consuming batches of messages and saving them to the database
|
10
|
+
# in bulk ActiveRecord operations.
|
11
|
+
module BatchConsumption
|
12
|
+
# Handle a batch of Kafka messages. Batches are split into "slices",
|
13
|
+
# which are groups of independent messages that can be processed together
|
14
|
+
# in a single database operation.
|
15
|
+
# If two messages in a batch have the same key, we cannot process them
|
16
|
+
# in the same operation as they would interfere with each other. Thus
|
17
|
+
# they are split
|
18
|
+
# @param payloads [Array<Hash>] Decoded payloads.
|
19
|
+
# @param metadata [Hash] Information about batch, including keys.
|
20
|
+
def consume_batch(payloads, metadata)
|
21
|
+
messages = payloads.
|
22
|
+
zip(metadata[:keys]).
|
23
|
+
map { |p, k| Deimos::Message.new(p, nil, key: k) }
|
24
|
+
|
25
|
+
tags = %W(topic:#{metadata[:topic]})
|
26
|
+
|
27
|
+
Deimos.instrument('ar_consumer.consume_batch', tags) do
|
28
|
+
# The entire batch should be treated as one transaction so that if
|
29
|
+
# any message fails, the whole thing is rolled back or retried
|
30
|
+
# if there is deadlock
|
31
|
+
Deimos::Utils::DeadlockRetry.wrap(tags) do
|
32
|
+
if @compacted || self.class.config[:no_keys]
|
33
|
+
update_database(compact_messages(messages))
|
34
|
+
else
|
35
|
+
uncompacted_update(messages)
|
36
|
+
end
|
37
|
+
end
|
38
|
+
end
|
39
|
+
end
|
40
|
+
|
41
|
+
# Get unique key for the ActiveRecord instance from the incoming key.
|
42
|
+
# Override this method (with super) to customize the set of attributes that
|
43
|
+
# uniquely identifies each record in the database.
|
44
|
+
# @param key [String] The encoded key.
|
45
|
+
# @return [Hash] The key attributes.
|
46
|
+
def record_key(key)
|
47
|
+
decoded_key = decode_key(key)
|
48
|
+
|
49
|
+
if decoded_key.nil?
|
50
|
+
{}
|
51
|
+
elsif decoded_key.is_a?(Hash)
|
52
|
+
@key_converter.convert(decoded_key)
|
53
|
+
else
|
54
|
+
{ @klass.primary_key => decoded_key }
|
55
|
+
end
|
56
|
+
end
|
57
|
+
|
58
|
+
protected
|
59
|
+
|
60
|
+
# Perform database operations for a batch of messages without compaction.
|
61
|
+
# All messages are split into slices containing only unique keys, and
|
62
|
+
# each slice is handles as its own batch.
|
63
|
+
# @param messages [Array<Message>] List of messages.
|
64
|
+
def uncompacted_update(messages)
|
65
|
+
BatchSlicer.
|
66
|
+
slice(messages).
|
67
|
+
each(&method(:update_database))
|
68
|
+
end
|
69
|
+
|
70
|
+
# Perform database operations for a group of messages.
|
71
|
+
# All messages with payloads are passed to upsert_records.
|
72
|
+
# All tombstones messages are passed to remove_records.
|
73
|
+
# @param messages [Array<Message>] List of messages.
|
74
|
+
def update_database(messages)
|
75
|
+
# Find all upserted records (i.e. that have a payload) and all
|
76
|
+
# deleted record (no payload)
|
77
|
+
removed, upserted = messages.partition(&:tombstone?)
|
78
|
+
|
79
|
+
upsert_records(upserted) if upserted.any?
|
80
|
+
remove_records(removed) if removed.any?
|
81
|
+
end
|
82
|
+
|
83
|
+
# Upsert any non-deleted records
|
84
|
+
# @param messages [Array<Message>] List of messages for a group of
|
85
|
+
# records to either be updated or inserted.
|
86
|
+
def upsert_records(messages)
|
87
|
+
key_cols = key_columns(messages)
|
88
|
+
|
89
|
+
# Create payloads with payload + key attributes
|
90
|
+
upserts = messages.map do |m|
|
91
|
+
attrs = if self.method(:record_attributes).parameters.size == 2
|
92
|
+
record_attributes(m.payload, m.key)
|
93
|
+
else
|
94
|
+
record_attributes(m.payload)
|
95
|
+
end
|
96
|
+
|
97
|
+
attrs&.merge(record_key(m.key))
|
98
|
+
end
|
99
|
+
|
100
|
+
# If overridden record_attributes indicated no record, skip
|
101
|
+
upserts.compact!
|
102
|
+
|
103
|
+
options = if key_cols.empty?
|
104
|
+
{} # Can't upsert with no key, just do regular insert
|
105
|
+
else
|
106
|
+
{
|
107
|
+
on_duplicate_key_update: {
|
108
|
+
# conflict_target must explicitly list the columns for
|
109
|
+
# Postgres and SQLite. Not required for MySQL, but this
|
110
|
+
# ensures consistent behaviour.
|
111
|
+
conflict_target: key_cols,
|
112
|
+
columns: :all
|
113
|
+
}
|
114
|
+
}
|
115
|
+
end
|
116
|
+
|
117
|
+
@klass.import!(upserts, options)
|
118
|
+
end
|
119
|
+
|
120
|
+
# Delete any records with a tombstone.
|
121
|
+
# @param messages [Array<Message>] List of messages for a group of
|
122
|
+
# deleted records.
|
123
|
+
def remove_records(messages)
|
124
|
+
clause = deleted_query(messages)
|
125
|
+
|
126
|
+
clause.delete_all
|
127
|
+
end
|
128
|
+
|
129
|
+
# Create an ActiveRecord relation that matches all of the passed
|
130
|
+
# records. Used for bulk deletion.
|
131
|
+
# @param records [Array<Message>] List of messages.
|
132
|
+
# @return ActiveRecord::Relation Matching relation.
|
133
|
+
def deleted_query(records)
|
134
|
+
keys = records.
|
135
|
+
map { |m| record_key(m.key)[@klass.primary_key] }.
|
136
|
+
reject(&:nil?)
|
137
|
+
|
138
|
+
@klass.unscoped.where(@klass.primary_key => keys)
|
139
|
+
end
|
140
|
+
|
141
|
+
# Get the set of attribute names that uniquely identify messages in the
|
142
|
+
# batch. Requires at least one record.
|
143
|
+
# @param records [Array<Message>] Non-empty list of messages.
|
144
|
+
# @return [Array<String>] List of attribute names.
|
145
|
+
# @raise If records is empty.
|
146
|
+
def key_columns(records)
|
147
|
+
raise 'Cannot determine key from empty batch' if records.empty?
|
148
|
+
|
149
|
+
first_key = records.first.key
|
150
|
+
record_key(first_key).keys
|
151
|
+
end
|
152
|
+
|
153
|
+
# Compact a batch of messages, taking only the last message for each
|
154
|
+
# unique key.
|
155
|
+
# @param batch [Array<Message>] Batch of messages.
|
156
|
+
# @return [Array<Message>] Compacted batch.
|
157
|
+
def compact_messages(batch)
|
158
|
+
return batch unless batch.first&.key.present?
|
159
|
+
|
160
|
+
batch.reverse.uniq(&:key).reverse!
|
161
|
+
end
|
162
|
+
end
|
163
|
+
end
|
164
|
+
end
|
@@ -0,0 +1,27 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Deimos
|
4
|
+
module ActiveRecordConsume
|
5
|
+
# Helper class for breaking down batches into independent groups for
|
6
|
+
# processing
|
7
|
+
class BatchSlicer
|
8
|
+
# Split the batch into a series of independent slices. Each slice contains
|
9
|
+
# messages that can be processed in any order (i.e. they have distinct
|
10
|
+
# keys). Messages with the same key will be separated into different
|
11
|
+
# slices that maintain the correct order.
|
12
|
+
# E.g. Given messages A1, A2, B1, C1, C2, C3, they will be sliced as:
|
13
|
+
# [[A1, B1, C1], [A2, C2], [C3]]
|
14
|
+
def self.slice(messages)
|
15
|
+
ops = messages.group_by(&:key)
|
16
|
+
|
17
|
+
# Find maximum depth
|
18
|
+
depth = ops.values.map(&:length).max || 0
|
19
|
+
|
20
|
+
# Generate slices for each depth
|
21
|
+
depth.times.map do |i|
|
22
|
+
ops.values.map { |arr| arr.dig(i) }.compact
|
23
|
+
end
|
24
|
+
end
|
25
|
+
end
|
26
|
+
end
|
27
|
+
end
|
@@ -0,0 +1,79 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Deimos
|
4
|
+
module ActiveRecordConsume
|
5
|
+
# Methods for consuming individual messages and saving them to the database
|
6
|
+
# as ActiveRecord instances.
|
7
|
+
module MessageConsumption
|
8
|
+
# Find the record specified by the given payload and key.
|
9
|
+
# Default is to use the primary key column and the value of the first
|
10
|
+
# field in the key.
|
11
|
+
# @param klass [Class < ActiveRecord::Base]
|
12
|
+
# @param _payload [Hash]
|
13
|
+
# @param key [Object]
|
14
|
+
# @return [ActiveRecord::Base]
|
15
|
+
def fetch_record(klass, _payload, key)
|
16
|
+
klass.unscoped.where(klass.primary_key => key).first
|
17
|
+
end
|
18
|
+
|
19
|
+
# Assign a key to a new record.
|
20
|
+
# @param record [ActiveRecord::Base]
|
21
|
+
# @param _payload [Hash]
|
22
|
+
# @param key [Object]
|
23
|
+
def assign_key(record, _payload, key)
|
24
|
+
record[record.class.primary_key] = key
|
25
|
+
end
|
26
|
+
|
27
|
+
# :nodoc:
|
28
|
+
def consume(payload, metadata)
|
29
|
+
unless self.process_message?(payload)
|
30
|
+
Deimos.config.logger.debug(
|
31
|
+
message: 'Skipping processing of message',
|
32
|
+
payload: payload,
|
33
|
+
metadata: metadata
|
34
|
+
)
|
35
|
+
return
|
36
|
+
end
|
37
|
+
|
38
|
+
key = metadata.with_indifferent_access[:key]
|
39
|
+
klass = self.class.config[:record_class]
|
40
|
+
record = fetch_record(klass, (payload || {}).with_indifferent_access, key)
|
41
|
+
if payload.nil?
|
42
|
+
destroy_record(record)
|
43
|
+
return
|
44
|
+
end
|
45
|
+
if record.blank?
|
46
|
+
record = klass.new
|
47
|
+
assign_key(record, payload, key)
|
48
|
+
end
|
49
|
+
|
50
|
+
# for backwards compatibility
|
51
|
+
# TODO next major release we should deprecate this
|
52
|
+
attrs = if self.method(:record_attributes).parameters.size == 2
|
53
|
+
record_attributes(payload.with_indifferent_access, key)
|
54
|
+
else
|
55
|
+
record_attributes(payload.with_indifferent_access)
|
56
|
+
end
|
57
|
+
# don't use attributes= - bypass Rails < 5 attr_protected
|
58
|
+
attrs.each do |k, v|
|
59
|
+
record.send("#{k}=", v)
|
60
|
+
end
|
61
|
+
save_record(record)
|
62
|
+
end
|
63
|
+
|
64
|
+
# @param record [ActiveRecord::Base]
|
65
|
+
def save_record(record)
|
66
|
+
record.created_at ||= Time.zone.now if record.respond_to?(:created_at)
|
67
|
+
record.updated_at = Time.zone.now if record.respond_to?(:updated_at)
|
68
|
+
record.save!
|
69
|
+
end
|
70
|
+
|
71
|
+
# Destroy a record that received a null payload. Override if you need
|
72
|
+
# to do something other than a straight destroy (e.g. mark as archived).
|
73
|
+
# @param record [ActiveRecord::Base]
|
74
|
+
def destroy_record(record)
|
75
|
+
record&.destroy
|
76
|
+
end
|
77
|
+
end
|
78
|
+
end
|
79
|
+
end
|
@@ -0,0 +1,52 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Deimos
|
4
|
+
module ActiveRecordConsume
|
5
|
+
# Convert a message with a schema to an ActiveRecord model
|
6
|
+
class SchemaModelConverter
|
7
|
+
# Create new converter
|
8
|
+
# @param decoder [SchemaBackends::Base] Incoming message schema.
|
9
|
+
# @param klass [ActiveRecord::Base] Model to map to.
|
10
|
+
def initialize(decoder, klass)
|
11
|
+
@decoder = decoder
|
12
|
+
@klass = klass
|
13
|
+
end
|
14
|
+
|
15
|
+
# Convert a message from a decoded hash to a set of ActiveRecord
|
16
|
+
# attributes. Attributes that don't exist in the model will be ignored.
|
17
|
+
# @param payload [Hash] Decoded message payload.
|
18
|
+
# @return [Hash] Model attributes.
|
19
|
+
def convert(payload)
|
20
|
+
attributes = {}
|
21
|
+
@decoder.schema_fields.each do |field|
|
22
|
+
column = @klass.columns.find { |c| c.name == field.name }
|
23
|
+
next if column.nil?
|
24
|
+
next if %w(updated_at created_at).include?(field.name)
|
25
|
+
|
26
|
+
attributes[field.name] = _coerce_field(column, payload[field.name])
|
27
|
+
end
|
28
|
+
attributes
|
29
|
+
end
|
30
|
+
|
31
|
+
private
|
32
|
+
|
33
|
+
# @param column [ActiveRecord::ConnectionAdapters::Column]
|
34
|
+
# @param val [Object]
|
35
|
+
def _coerce_field(column, val)
|
36
|
+
return nil if val.nil?
|
37
|
+
|
38
|
+
if column.type == :datetime
|
39
|
+
int_val = begin
|
40
|
+
val.is_a?(Integer) ? val : (val.is_a?(String) && Integer(val))
|
41
|
+
rescue StandardError
|
42
|
+
nil
|
43
|
+
end
|
44
|
+
|
45
|
+
return Time.zone.at(int_val) if int_val
|
46
|
+
end
|
47
|
+
|
48
|
+
val
|
49
|
+
end
|
50
|
+
end
|
51
|
+
end
|
52
|
+
end
|
@@ -0,0 +1,67 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require 'deimos/active_record_consume/batch_consumption'
|
4
|
+
require 'deimos/active_record_consume/message_consumption'
|
5
|
+
require 'deimos/active_record_consume/schema_model_converter'
|
6
|
+
require 'deimos/consumer'
|
7
|
+
|
8
|
+
module Deimos
|
9
|
+
# Basic ActiveRecord consumer class. Consumes messages and upserts them to
|
10
|
+
# the database. For tombstones (null payloads), deletes corresponding
|
11
|
+
# records from the database. Can operate in either message-by-message mode
|
12
|
+
# or in batch mode.
|
13
|
+
#
|
14
|
+
# In batch mode, ActiveRecord callbacks will be skipped and messages will
|
15
|
+
# be batched to minimize database calls.
|
16
|
+
|
17
|
+
# To configure batch vs. message mode, change the delivery mode of your
|
18
|
+
# Phobos listener.
|
19
|
+
# Message-by-message -> use `delivery: message` or `delivery: batch`
|
20
|
+
# Batch -> use `delivery: inline_batch`
|
21
|
+
class ActiveRecordConsumer < Consumer
|
22
|
+
include ActiveRecordConsume::MessageConsumption
|
23
|
+
include ActiveRecordConsume::BatchConsumption
|
24
|
+
|
25
|
+
class << self
|
26
|
+
# param klass [Class < ActiveRecord::Base] the class used to save to the
|
27
|
+
# database.
|
28
|
+
def record_class(klass)
|
29
|
+
config[:record_class] = klass
|
30
|
+
end
|
31
|
+
|
32
|
+
# param val [Boolean] Turn pre-compaction of the batch on or off. If true,
|
33
|
+
# only the last message for each unique key in a batch is processed.
|
34
|
+
def compacted(val)
|
35
|
+
config[:compacted] = val
|
36
|
+
end
|
37
|
+
end
|
38
|
+
|
39
|
+
# Setup
|
40
|
+
def initialize
|
41
|
+
@klass = self.class.config[:record_class]
|
42
|
+
@converter = ActiveRecordConsume::SchemaModelConverter.new(self.class.decoder, @klass)
|
43
|
+
|
44
|
+
if self.class.config[:key_schema]
|
45
|
+
@key_converter = ActiveRecordConsume::SchemaModelConverter.new(self.class.key_decoder, @klass)
|
46
|
+
end
|
47
|
+
|
48
|
+
@compacted = self.class.config[:compacted] != false
|
49
|
+
end
|
50
|
+
|
51
|
+
# Override this method (with `super`) if you want to add/change the default
|
52
|
+
# attributes set to the new/existing record.
|
53
|
+
# @param payload [Hash]
|
54
|
+
# @param _key [String]
|
55
|
+
def record_attributes(payload, _key=nil)
|
56
|
+
@converter.convert(payload)
|
57
|
+
end
|
58
|
+
|
59
|
+
# Override this message to conditionally save records
|
60
|
+
# @param payload [Hash] The kafka message as a hash
|
61
|
+
# @return [Boolean] if true, record is created/update.
|
62
|
+
# If false, record processing is skipped but message offset is still committed.
|
63
|
+
def process_message?(_payload)
|
64
|
+
true
|
65
|
+
end
|
66
|
+
end
|
67
|
+
end
|
@@ -0,0 +1,87 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require 'deimos/producer'
|
4
|
+
|
5
|
+
module Deimos
|
6
|
+
# Class which automatically produces a record when given an ActiveRecord
|
7
|
+
# instance or a list of them. Just call `send_events` on a list of records
|
8
|
+
# and they will be auto-published. You can override `generate_payload`
|
9
|
+
# to make changes to the payload before it's published.
|
10
|
+
#
|
11
|
+
# You can also call this with a list of hashes representing attributes.
|
12
|
+
# This is common when using activerecord-import.
|
13
|
+
class ActiveRecordProducer < Producer
|
14
|
+
class << self
|
15
|
+
# Indicate the class this producer is working on.
|
16
|
+
# @param klass [Class]
|
17
|
+
# @param refetch [Boolean] if true, and we are given a hash instead of
|
18
|
+
# a record object, refetch the record to pass into the `generate_payload`
|
19
|
+
# method.
|
20
|
+
def record_class(klass, refetch: true)
|
21
|
+
config[:record_class] = klass
|
22
|
+
config[:refetch_record] = refetch
|
23
|
+
end
|
24
|
+
|
25
|
+
# @param record [ActiveRecord::Base]
|
26
|
+
# @param force_send [Boolean]
|
27
|
+
def send_event(record, force_send: false)
|
28
|
+
send_events([record], force_send: force_send)
|
29
|
+
end
|
30
|
+
|
31
|
+
# @param records [Array<ActiveRecord::Base>]
|
32
|
+
# @param force_send [Boolean]
|
33
|
+
def send_events(records, force_send: false)
|
34
|
+
primary_key = config[:record_class]&.primary_key
|
35
|
+
messages = records.map do |record|
|
36
|
+
if record.respond_to?(:attributes)
|
37
|
+
attrs = record.attributes.with_indifferent_access
|
38
|
+
else
|
39
|
+
attrs = record.with_indifferent_access
|
40
|
+
if config[:refetch_record] && attrs[primary_key]
|
41
|
+
record = config[:record_class].find(attrs[primary_key])
|
42
|
+
end
|
43
|
+
end
|
44
|
+
generate_payload(attrs, record).with_indifferent_access
|
45
|
+
end
|
46
|
+
self.publish_list(messages, force_send: force_send)
|
47
|
+
end
|
48
|
+
|
49
|
+
# Generate the payload, given a list of attributes or a record..
|
50
|
+
# Can be overridden or added to by subclasses.
|
51
|
+
# @param attributes [Hash]
|
52
|
+
# @param _record [ActiveRecord::Base] May be nil if refetch_record
|
53
|
+
# is not set.
|
54
|
+
# @return [Hash]
|
55
|
+
def generate_payload(attributes, _record)
|
56
|
+
fields = self.encoder.schema_fields
|
57
|
+
payload = attributes.stringify_keys
|
58
|
+
payload.delete_if do |k, _|
|
59
|
+
k.to_sym != :payload_key && !fields.map(&:name).include?(k)
|
60
|
+
end
|
61
|
+
end
|
62
|
+
|
63
|
+
# Query to use when polling the database with the DbPoller. Add
|
64
|
+
# includes, joins, or wheres as necessary, or replace entirely.
|
65
|
+
# @param time_from [Time] the time to start the query from.
|
66
|
+
# @param time_to [Time] the time to end the query.
|
67
|
+
# @param column_name [Symbol] the column name to look for.
|
68
|
+
# @param min_id [Numeric] the minimum ID (i.e. all IDs must be greater
|
69
|
+
# than this value).
|
70
|
+
# @return [ActiveRecord::Relation]
|
71
|
+
def poll_query(time_from:, time_to:, column_name: :updated_at, min_id:)
|
72
|
+
klass = config[:record_class]
|
73
|
+
table = ActiveRecord::Base.connection.quote_table_name(klass.table_name)
|
74
|
+
column = ActiveRecord::Base.connection.quote_column_name(column_name)
|
75
|
+
primary = ActiveRecord::Base.connection.quote_column_name(klass.primary_key)
|
76
|
+
klass.where(
|
77
|
+
"((#{table}.#{column} = ? AND #{table}.#{primary} > ?) \
|
78
|
+
OR #{table}.#{column} > ?) AND #{table}.#{column} <= ?",
|
79
|
+
time_from,
|
80
|
+
min_id,
|
81
|
+
time_from,
|
82
|
+
time_to
|
83
|
+
)
|
84
|
+
end
|
85
|
+
end
|
86
|
+
end
|
87
|
+
end
|