multiwoven-integrations 0.1.66 → 0.1.67
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 +4 -4
- data/lib/multiwoven/integrations/rollout.rb +2 -1
- data/lib/multiwoven/integrations/source/aws_athena/client.rb +109 -0
- data/lib/multiwoven/integrations/source/aws_athena/config/meta.json +15 -0
- data/lib/multiwoven/integrations/source/aws_athena/config/spec.json +63 -0
- data/lib/multiwoven/integrations/source/aws_athena/icon.svg +22 -0
- data/lib/multiwoven/integrations.rb +2 -0
- data/multiwoven-integrations.gemspec +1 -0
- metadata +20 -2
checksums.yaml
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
---
|
|
2
2
|
SHA256:
|
|
3
|
-
metadata.gz:
|
|
4
|
-
data.tar.gz:
|
|
3
|
+
metadata.gz: da70c8d91dea6167ebb4103cfe0e1999d663ee2ac420d9ae272a5dbd70924380
|
|
4
|
+
data.tar.gz: bc6fd6fd71c8c02dcdcf4e1f3eaf69dee37f3df6ba5f6fb1d4af49aa64220119
|
|
5
5
|
SHA512:
|
|
6
|
-
metadata.gz:
|
|
7
|
-
data.tar.gz:
|
|
6
|
+
metadata.gz: 2040a6232add940bd41cffa131c5738d64a4187df2615f3625910aa6129ae089ffd89027ff78df15d753a27f5eb10cec0357f4b977f25d86c9663480dce88c71
|
|
7
|
+
data.tar.gz: cfe7e24682df19fd0aa10284891d073324933838ceb10bfab3a8dc7259705af422b86d4b5fc85665af580a2097625fb0feafdd05152ea9d2f5bf6d0d0e51fff4
|
|
@@ -2,7 +2,7 @@
|
|
|
2
2
|
|
|
3
3
|
module Multiwoven
|
|
4
4
|
module Integrations
|
|
5
|
-
VERSION = "0.1.
|
|
5
|
+
VERSION = "0.1.67"
|
|
6
6
|
|
|
7
7
|
ENABLED_SOURCES = %w[
|
|
8
8
|
Snowflake
|
|
@@ -11,6 +11,7 @@ module Multiwoven
|
|
|
11
11
|
Postgresql
|
|
12
12
|
Databricks
|
|
13
13
|
SalesforceConsumerGoodsCloud
|
|
14
|
+
AwsAthena
|
|
14
15
|
].freeze
|
|
15
16
|
|
|
16
17
|
ENABLED_DESTINATIONS = %w[
|
|
@@ -0,0 +1,109 @@
|
|
|
1
|
+
# frozen_string_literal: true
|
|
2
|
+
|
|
3
|
+
module Multiwoven::Integrations::Source
|
|
4
|
+
module AwsAthena
|
|
5
|
+
include Multiwoven::Integrations::Core
|
|
6
|
+
class Client < SourceConnector
|
|
7
|
+
def check_connection(connection_config)
|
|
8
|
+
connection_config = connection_config.with_indifferent_access
|
|
9
|
+
athena_client = create_connection(connection_config)
|
|
10
|
+
athena_client.list_work_groups
|
|
11
|
+
ConnectionStatus.new(status: ConnectionStatusType["succeeded"]).to_multiwoven_message
|
|
12
|
+
rescue StandardError => e
|
|
13
|
+
ConnectionStatus.new(status: ConnectionStatusType["failed"], message: e.message).to_multiwoven_message
|
|
14
|
+
end
|
|
15
|
+
|
|
16
|
+
def discover(connection_config)
|
|
17
|
+
connection_config = connection_config.with_indifferent_access
|
|
18
|
+
query = "SELECT table_name, column_name, data_type, is_nullable FROM information_schema.columns WHERE table_schema = '#{connection_config[:schema]}' ORDER BY table_name, ordinal_position;"
|
|
19
|
+
db = create_connection(connection_config)
|
|
20
|
+
results = query_execution(db, query)
|
|
21
|
+
catalog = Catalog.new(streams: create_streams(results))
|
|
22
|
+
catalog.to_multiwoven_message
|
|
23
|
+
rescue StandardError => e
|
|
24
|
+
handle_exception(
|
|
25
|
+
"AWS:ATHENA:DISCOVER:EXCEPTION",
|
|
26
|
+
"error",
|
|
27
|
+
e
|
|
28
|
+
)
|
|
29
|
+
end
|
|
30
|
+
|
|
31
|
+
def read(sync_config)
|
|
32
|
+
connection_config = sync_config.source.connection_specification
|
|
33
|
+
connection_config = connection_config.with_indifferent_access
|
|
34
|
+
query = sync_config.model.query
|
|
35
|
+
query = batched_query(query, sync_config.limit, sync_config.offset) unless sync_config.limit.nil? && sync_config.offset.nil?
|
|
36
|
+
db = create_connection(connection_config)
|
|
37
|
+
query(db, query)
|
|
38
|
+
rescue StandardError => e
|
|
39
|
+
handle_exception(
|
|
40
|
+
"AWS:ATHENA:READ:EXCEPTION",
|
|
41
|
+
"error",
|
|
42
|
+
e
|
|
43
|
+
)
|
|
44
|
+
end
|
|
45
|
+
|
|
46
|
+
private
|
|
47
|
+
|
|
48
|
+
def create_connection(connection_config)
|
|
49
|
+
Aws.config.update({ credentials: Aws::Credentials.new(connection_config[:access_key], connection_config[:secret_access_key]), region: connection_config[:region] })
|
|
50
|
+
@database = connection_config[:schema]
|
|
51
|
+
@output_location = connection_config[:output_location]
|
|
52
|
+
Aws::Athena::Client.new
|
|
53
|
+
end
|
|
54
|
+
|
|
55
|
+
def query_execution(db, query)
|
|
56
|
+
response = db.start_query_execution(
|
|
57
|
+
query_string: query,
|
|
58
|
+
query_execution_context: { database: @database },
|
|
59
|
+
result_configuration: { output_location: @output_location }
|
|
60
|
+
)
|
|
61
|
+
query_execution_id = response[:query_execution_id]
|
|
62
|
+
loop do
|
|
63
|
+
response = db.get_query_execution(query_execution_id: query_execution_id)
|
|
64
|
+
status = response.query_execution.status.state
|
|
65
|
+
break if %w[SUCCEEDED FAILED CANCELLED].include?(status)
|
|
66
|
+
end
|
|
67
|
+
transform_results(db.get_query_results(query_execution_id: query_execution_id))
|
|
68
|
+
end
|
|
69
|
+
|
|
70
|
+
def create_streams(records)
|
|
71
|
+
group_by_table(records).map do |_, r|
|
|
72
|
+
Multiwoven::Integrations::Protocol::Stream.new(name: r[:tablename], action: StreamAction["fetch"], json_schema: convert_to_json_schema(r[:columns]))
|
|
73
|
+
end
|
|
74
|
+
end
|
|
75
|
+
|
|
76
|
+
def transform_results(results)
|
|
77
|
+
columns = results.result_set.result_set_metadata.column_info.map(&:name)
|
|
78
|
+
rows = results.result_set.rows.map do |row|
|
|
79
|
+
row.data.map(&:var_char_value)
|
|
80
|
+
end
|
|
81
|
+
rows.map { |row| columns.zip(row).to_h }
|
|
82
|
+
end
|
|
83
|
+
|
|
84
|
+
def query(db, query)
|
|
85
|
+
records = []
|
|
86
|
+
query_execution(db, query).map do |row|
|
|
87
|
+
records << RecordMessage.new(data: row, emitted_at: Time.now.to_i).to_multiwoven_message
|
|
88
|
+
end
|
|
89
|
+
records
|
|
90
|
+
end
|
|
91
|
+
|
|
92
|
+
def group_by_table(records)
|
|
93
|
+
result = {}
|
|
94
|
+
records.each_with_index do |entry, index|
|
|
95
|
+
table_name = entry["table_name"]
|
|
96
|
+
column_data = {
|
|
97
|
+
column_name: entry["column_name"],
|
|
98
|
+
data_type: entry["data_type"],
|
|
99
|
+
is_nullable: entry["is_nullable"] == "YES"
|
|
100
|
+
}
|
|
101
|
+
result[index] ||= {}
|
|
102
|
+
result[index][:tablename] = table_name
|
|
103
|
+
result[index][:columns] = [column_data]
|
|
104
|
+
end
|
|
105
|
+
result
|
|
106
|
+
end
|
|
107
|
+
end
|
|
108
|
+
end
|
|
109
|
+
end
|
|
@@ -0,0 +1,15 @@
|
|
|
1
|
+
{
|
|
2
|
+
"data": {
|
|
3
|
+
"name": "AwsAthena",
|
|
4
|
+
"title": "AWS Athena",
|
|
5
|
+
"connector_type": "source",
|
|
6
|
+
"category": "Data Warehouse",
|
|
7
|
+
"documentation_url": "https://docs.mutliwoven.com",
|
|
8
|
+
"github_issue_label": "source-aws-athena",
|
|
9
|
+
"icon": "icon.svg",
|
|
10
|
+
"license": "MIT",
|
|
11
|
+
"release_stage": "alpha",
|
|
12
|
+
"support_level": "community",
|
|
13
|
+
"tags": ["language:ruby", "multiwoven"]
|
|
14
|
+
}
|
|
15
|
+
}
|
|
@@ -0,0 +1,63 @@
|
|
|
1
|
+
{
|
|
2
|
+
"documentation_url": "https://docs.multiwoven.com/integrations/sources/athena",
|
|
3
|
+
"stream_type": "dynamic",
|
|
4
|
+
"connector_query_type": "raw_sql",
|
|
5
|
+
"connection_specification": {
|
|
6
|
+
"$schema": "http://json-schema.org/draft-07/schema#",
|
|
7
|
+
"title": "AWS Athena",
|
|
8
|
+
"type": "object",
|
|
9
|
+
"required": ["access_key", "secret_access_key", "region", "workgroup", "catalog", "output_location"],
|
|
10
|
+
"properties": {
|
|
11
|
+
"access_key": {
|
|
12
|
+
"description": "The AWS Access Key ID to use for authentication.",
|
|
13
|
+
"examples": ["AWSATHENAACCESS"],
|
|
14
|
+
"type": "string",
|
|
15
|
+
"title": "Personal Access Key",
|
|
16
|
+
"order": 0
|
|
17
|
+
},
|
|
18
|
+
"secret_access_key": {
|
|
19
|
+
"description": "The AWS Secret Access Key to use for authentication.",
|
|
20
|
+
"examples": ["AWSATHENASECRET"],
|
|
21
|
+
"type": "string",
|
|
22
|
+
"multiwoven_secret": true,
|
|
23
|
+
"title": "Secret Access Key",
|
|
24
|
+
"order": 1
|
|
25
|
+
},
|
|
26
|
+
"region": {
|
|
27
|
+
"description": "AWS region where Athena is located.",
|
|
28
|
+
"examples": ["ATHENA_REGION"],
|
|
29
|
+
"type": "string",
|
|
30
|
+
"title": "Secret Access Key",
|
|
31
|
+
"order": 2
|
|
32
|
+
},
|
|
33
|
+
"workgroup": {
|
|
34
|
+
"description": "The Athena workgroup you previously set up in AWS.",
|
|
35
|
+
"examples": ["ATHENA_WORKGROUP"],
|
|
36
|
+
"type": "string",
|
|
37
|
+
"title": "Workgroup",
|
|
38
|
+
"order": 3
|
|
39
|
+
},
|
|
40
|
+
"catalog": {
|
|
41
|
+
"description": "The Data catalog name within Athena.",
|
|
42
|
+
"examples": ["ATHENA_CATALOG"],
|
|
43
|
+
"type": "string",
|
|
44
|
+
"title": "Catalog",
|
|
45
|
+
"order": 4
|
|
46
|
+
},
|
|
47
|
+
"schema": {
|
|
48
|
+
"description": "The specific Athena database/schema to connect to.",
|
|
49
|
+
"examples": ["ATHENA_DB"],
|
|
50
|
+
"type": "string",
|
|
51
|
+
"title": "Database",
|
|
52
|
+
"order": 5
|
|
53
|
+
},
|
|
54
|
+
"output_location": {
|
|
55
|
+
"description": "S3 path for query output.",
|
|
56
|
+
"examples": ["s3://example-bucket-name/query-results/"],
|
|
57
|
+
"type": "string",
|
|
58
|
+
"title": "Query",
|
|
59
|
+
"order": 6
|
|
60
|
+
}
|
|
61
|
+
}
|
|
62
|
+
}
|
|
63
|
+
}
|
|
@@ -0,0 +1,22 @@
|
|
|
1
|
+
<svg height="100" viewBox="0 0 100 100" width="100" xmlns="http://www.w3.org/2000/svg">
|
|
2
|
+
<path d="m84 64.68-34-4.19-34 4.19 34 26.62z" fill="#fcbf92"/>
|
|
3
|
+
<path d="m16 64.68 34 10.17v16.45l-34-17.05z" fill="#9d5025"/>
|
|
4
|
+
<path d="m84 64.68-34 10.17v16.45l34-17.05z" fill="#f58534"/>
|
|
5
|
+
<path d="m20.39 50.22h-4.39v9.67l4.39.8 4.88-5.18z" fill="#9d5025"/>
|
|
6
|
+
<path d="m25.27 60.29-4.88.4v-10.47h4.88z" fill="#f58534"/>
|
|
7
|
+
<path d="m28.56 38.16-5.78 1.09v21.94l5.78 1.2 5.29-12.17z" fill="#9d5025"/>
|
|
8
|
+
<path d="m33.85 61.79-5.29.6v-24.23l5.29.6z" fill="#f58534"/>
|
|
9
|
+
<path d="m39.83 43.04-7.98.8v19.24l7.98 1.6 8.47-10.87z" fill="#9d5025"/>
|
|
10
|
+
<path d="m45.21 63.78-5.38.9v-21.64l5.38.4z" fill="#f58534"/>
|
|
11
|
+
<g fill="#9d5025">
|
|
12
|
+
<path d="m50 25.59-5.38 1.7v38.29l5.38 1.09 5.38-20.54z"/>
|
|
13
|
+
<path d="m54.98 63.78 5.19.9 7.98-14.46-7.98-14.36-5.19.8z"/>
|
|
14
|
+
<path d="m68.15 50.22-7.98-14.36-5.19.8"/>
|
|
15
|
+
<path d="m66.05 61.79 5.39.6 5.78-18.25-5.78-18.15-5.39 1.3z"/>
|
|
16
|
+
<path d="m74.73 60.29 4.88.4 4.39-18.05-4.39-18.64-4.88 1.2z"/>
|
|
17
|
+
</g>
|
|
18
|
+
<path d="m50 25.59 5.38 1.7v38.29l-5.38 1.09z" fill="#f58534"/>
|
|
19
|
+
<path d="m68.15 37.46-7.98-1.6v28.82l7.98-1.6z" fill="#f58534"/>
|
|
20
|
+
<path d="m77.22 28.29-5.78-2.3v36.4l5.78-1.2z" fill="#f58534"/>
|
|
21
|
+
<path d="m84 26.19-4.39-2.19v36.69l4.39-.8z" fill="#f58534"/>
|
|
22
|
+
</svg>
|
|
@@ -22,6 +22,7 @@ require "stripe"
|
|
|
22
22
|
require "net/sftp"
|
|
23
23
|
require "csv"
|
|
24
24
|
require "securerandom"
|
|
25
|
+
require "aws-sdk-athena"
|
|
25
26
|
require "zip"
|
|
26
27
|
require "zendesk_api"
|
|
27
28
|
|
|
@@ -49,6 +50,7 @@ require_relative "integrations/source/bigquery/client"
|
|
|
49
50
|
require_relative "integrations/source/postgresql/client"
|
|
50
51
|
require_relative "integrations/source/databricks/client"
|
|
51
52
|
require_relative "integrations/source/salesforce_consumer_goods_cloud/client"
|
|
53
|
+
require_relative "integrations/source/aws_athena/client"
|
|
52
54
|
|
|
53
55
|
# Destination
|
|
54
56
|
require_relative "integrations/destination/klaviyo/client"
|
|
@@ -35,6 +35,7 @@ Gem::Specification.new do |spec|
|
|
|
35
35
|
|
|
36
36
|
spec.add_runtime_dependency "activesupport"
|
|
37
37
|
spec.add_runtime_dependency "async-websocket"
|
|
38
|
+
spec.add_runtime_dependency "aws-sdk-athena"
|
|
38
39
|
spec.add_runtime_dependency "csv"
|
|
39
40
|
spec.add_runtime_dependency "dry-schema"
|
|
40
41
|
spec.add_runtime_dependency "dry-struct"
|
metadata
CHANGED
|
@@ -1,14 +1,14 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: multiwoven-integrations
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 0.1.
|
|
4
|
+
version: 0.1.67
|
|
5
5
|
platform: ruby
|
|
6
6
|
authors:
|
|
7
7
|
- Subin T P
|
|
8
8
|
autorequire:
|
|
9
9
|
bindir: exe
|
|
10
10
|
cert_chain: []
|
|
11
|
-
date: 2024-05-
|
|
11
|
+
date: 2024-05-15 00:00:00.000000000 Z
|
|
12
12
|
dependencies:
|
|
13
13
|
- !ruby/object:Gem::Dependency
|
|
14
14
|
name: activesupport
|
|
@@ -38,6 +38,20 @@ dependencies:
|
|
|
38
38
|
- - ">="
|
|
39
39
|
- !ruby/object:Gem::Version
|
|
40
40
|
version: '0'
|
|
41
|
+
- !ruby/object:Gem::Dependency
|
|
42
|
+
name: aws-sdk-athena
|
|
43
|
+
requirement: !ruby/object:Gem::Requirement
|
|
44
|
+
requirements:
|
|
45
|
+
- - ">="
|
|
46
|
+
- !ruby/object:Gem::Version
|
|
47
|
+
version: '0'
|
|
48
|
+
type: :runtime
|
|
49
|
+
prerelease: false
|
|
50
|
+
version_requirements: !ruby/object:Gem::Requirement
|
|
51
|
+
requirements:
|
|
52
|
+
- - ">="
|
|
53
|
+
- !ruby/object:Gem::Version
|
|
54
|
+
version: '0'
|
|
41
55
|
- !ruby/object:Gem::Dependency
|
|
42
56
|
name: csv
|
|
43
57
|
requirement: !ruby/object:Gem::Requirement
|
|
@@ -486,6 +500,10 @@ files:
|
|
|
486
500
|
- lib/multiwoven/integrations/protocol/protocol.rb
|
|
487
501
|
- lib/multiwoven/integrations/rollout.rb
|
|
488
502
|
- lib/multiwoven/integrations/service.rb
|
|
503
|
+
- lib/multiwoven/integrations/source/aws_athena/client.rb
|
|
504
|
+
- lib/multiwoven/integrations/source/aws_athena/config/meta.json
|
|
505
|
+
- lib/multiwoven/integrations/source/aws_athena/config/spec.json
|
|
506
|
+
- lib/multiwoven/integrations/source/aws_athena/icon.svg
|
|
489
507
|
- lib/multiwoven/integrations/source/bigquery/client.rb
|
|
490
508
|
- lib/multiwoven/integrations/source/bigquery/config/meta.json
|
|
491
509
|
- lib/multiwoven/integrations/source/bigquery/config/spec.json
|