multiwoven-integrations 0.1.76 → 0.2.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/lib/multiwoven/integrations/core/source_connector.rb +6 -0
- data/lib/multiwoven/integrations/rollout.rb +2 -1
- data/lib/multiwoven/integrations/source/amazon_s3/client.rb +151 -0
- data/lib/multiwoven/integrations/source/amazon_s3/config/meta.json +15 -0
- data/lib/multiwoven/integrations/source/amazon_s3/config/spec.json +51 -0
- data/lib/multiwoven/integrations/source/amazon_s3/icon.svg +34 -0
- data/lib/multiwoven/integrations.rb +3 -0
- data/multiwoven-integrations.gemspec +2 -0
- metadata +34 -2
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 91cf72711d231942521cb83dd735b0a4b9067bd30d2aaeecf2be27d69d163da2
|
4
|
+
data.tar.gz: 5f2cc42a72e86cb0d8d1ed590c6fc337aa4539892305fc4f0502c74c7bb5dfa7
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 392ac958aeb012d4e1bf48daadbe168cf8e0cabd095fd79d155ed8fa4cc14cd8967974d9ee0085e4cf36ac9b78111d09da6e631fe21f97fd5be2136200fd56af
|
7
|
+
data.tar.gz: e3e76a87549ad98f9bed153746035fc1c6672ec30754eb1943a54d43cf2dbe253cbc3bb4235c4b11c9b98544856ff432cbd067af36bada6a12422d835426acca
|
@@ -13,6 +13,12 @@ module Multiwoven
|
|
13
13
|
|
14
14
|
private
|
15
15
|
|
16
|
+
# This needs to be implemented as private method
|
17
|
+
# In every source connector. This will be used for model preview
|
18
|
+
def create_connection(connector_config)
|
19
|
+
# return a connection to the client's source
|
20
|
+
end
|
21
|
+
|
16
22
|
# This needs to be implemented as private method
|
17
23
|
# In every source connector. This will be used for model preview
|
18
24
|
def query(connection, query)
|
@@ -2,7 +2,7 @@
|
|
2
2
|
|
3
3
|
module Multiwoven
|
4
4
|
module Integrations
|
5
|
-
VERSION = "0.
|
5
|
+
VERSION = "0.2.0"
|
6
6
|
|
7
7
|
ENABLED_SOURCES = %w[
|
8
8
|
Snowflake
|
@@ -13,6 +13,7 @@ module Multiwoven
|
|
13
13
|
SalesforceConsumerGoodsCloud
|
14
14
|
AwsAthena
|
15
15
|
Clickhouse
|
16
|
+
AmazonS3
|
16
17
|
].freeze
|
17
18
|
|
18
19
|
ENABLED_DESTINATIONS = %w[
|
@@ -0,0 +1,151 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Multiwoven::Integrations::Source
|
4
|
+
module AmazonS3
|
5
|
+
include Multiwoven::Integrations::Core
|
6
|
+
class Client < SourceConnector
|
7
|
+
DISCOVER_QUERY = "SELECT * FROM S3Object LIMIT 1;"
|
8
|
+
|
9
|
+
def check_connection(connection_config)
|
10
|
+
connection_config = connection_config.with_indifferent_access
|
11
|
+
client = config_aws(connection_config)
|
12
|
+
client.get_bucket_policy_status({ bucket: connection_config[:bucket] })
|
13
|
+
ConnectionStatus.new(status: ConnectionStatusType["succeeded"]).to_multiwoven_message
|
14
|
+
rescue StandardError => e
|
15
|
+
ConnectionStatus.new(status: ConnectionStatusType["failed"], message: e.message).to_multiwoven_message
|
16
|
+
end
|
17
|
+
|
18
|
+
def discover(connection_config)
|
19
|
+
connection_config = connection_config.with_indifferent_access
|
20
|
+
conn = create_connection(connection_config)
|
21
|
+
# If pulling from multiple files, all files must have the same schema
|
22
|
+
path = build_path(connection_config[:path])
|
23
|
+
full_path = "s3://#{connection_config[:bucket]}/#{path}*.#{connection_config[:file_type]}"
|
24
|
+
records = get_results(conn, "DESCRIBE SELECT * FROM '#{full_path}';")
|
25
|
+
columns = build_discover_columns(records)
|
26
|
+
streams = [Multiwoven::Integrations::Protocol::Stream.new(name: full_path, action: StreamAction["fetch"], json_schema: convert_to_json_schema(columns))]
|
27
|
+
catalog = Catalog.new(streams: streams)
|
28
|
+
catalog.to_multiwoven_message
|
29
|
+
rescue StandardError => e
|
30
|
+
handle_exception(e, { context: "AMAZONS3:DISCOVER:EXCEPTION", type: "error" })
|
31
|
+
end
|
32
|
+
|
33
|
+
def read(sync_config)
|
34
|
+
connection_config = sync_config.source.connection_specification.with_indifferent_access
|
35
|
+
conn = create_connection(connection_config)
|
36
|
+
query = sync_config.model.query
|
37
|
+
query = batched_query(query, sync_config.limit, sync_config.offset) unless sync_config.limit.nil? && sync_config.offset.nil?
|
38
|
+
query(conn, query)
|
39
|
+
rescue StandardError => e
|
40
|
+
handle_exception(e, {
|
41
|
+
context: "AMAZONS3:READ:EXCEPTION",
|
42
|
+
type: "error",
|
43
|
+
sync_id: sync_config.sync_id,
|
44
|
+
sync_run_id: sync_config.sync_run_id
|
45
|
+
})
|
46
|
+
end
|
47
|
+
|
48
|
+
private
|
49
|
+
|
50
|
+
# DuckDB
|
51
|
+
def create_connection(connection_config)
|
52
|
+
conn = DuckDB::Database.open.connect
|
53
|
+
# Set up S3 configuration
|
54
|
+
secret_query = "
|
55
|
+
CREATE SECRET amazons3_source (
|
56
|
+
TYPE S3,
|
57
|
+
KEY_ID '#{connection_config[:access_id]}',
|
58
|
+
SECRET '#{connection_config[:secret_access]}',
|
59
|
+
REGION '#{connection_config[:region]}'
|
60
|
+
);
|
61
|
+
"
|
62
|
+
get_results(conn, secret_query)
|
63
|
+
conn
|
64
|
+
end
|
65
|
+
|
66
|
+
def build_path(path)
|
67
|
+
path = "#{path}/" if !path.to_s.strip.empty? && path[-1] != "/"
|
68
|
+
path
|
69
|
+
end
|
70
|
+
|
71
|
+
def get_results(conn, query)
|
72
|
+
results = conn.query(query)
|
73
|
+
hash_array_values(results)
|
74
|
+
end
|
75
|
+
|
76
|
+
def query(conn, query)
|
77
|
+
records = get_results(conn, query)
|
78
|
+
records.map do |row|
|
79
|
+
RecordMessage.new(data: row, emitted_at: Time.now.to_i).to_multiwoven_message
|
80
|
+
end
|
81
|
+
end
|
82
|
+
|
83
|
+
def hash_array_values(describe)
|
84
|
+
keys = describe.columns.map(&:name)
|
85
|
+
describe.map do |row|
|
86
|
+
Hash[keys.zip(row)]
|
87
|
+
end
|
88
|
+
end
|
89
|
+
|
90
|
+
def build_discover_columns(describe_results)
|
91
|
+
describe_results.map do |row|
|
92
|
+
type = column_schema_helper(row["column_type"])
|
93
|
+
{
|
94
|
+
column_name: row["column_name"],
|
95
|
+
type: type
|
96
|
+
}
|
97
|
+
end
|
98
|
+
end
|
99
|
+
|
100
|
+
def column_schema_helper(column_type)
|
101
|
+
case column_type
|
102
|
+
when "VARCHAR", "BIT", "DATE", "TIME", "TIMESTAMP", "UUID"
|
103
|
+
"string"
|
104
|
+
when "DOUBLE"
|
105
|
+
"number"
|
106
|
+
when "BIGINT", "HUGEINT", "INTEGER", "SMALLINT"
|
107
|
+
"integer"
|
108
|
+
when "BOOLEAN"
|
109
|
+
"boolean"
|
110
|
+
end
|
111
|
+
end
|
112
|
+
|
113
|
+
# AWS SDK
|
114
|
+
def config_aws(config)
|
115
|
+
config = config.with_indifferent_access
|
116
|
+
Aws.config.update({
|
117
|
+
region: config[:region],
|
118
|
+
credentials: Aws::Credentials.new(config[:access_id], config[:secret_access])
|
119
|
+
})
|
120
|
+
config.with_indifferent_access
|
121
|
+
Aws::S3::Client.new
|
122
|
+
end
|
123
|
+
|
124
|
+
def build_select_content_options(config, query)
|
125
|
+
config = config.with_indifferent_access
|
126
|
+
bucket_name = config[:bucket]
|
127
|
+
file_key = config[:file_key]
|
128
|
+
file_type = config[:file_type]
|
129
|
+
options = {
|
130
|
+
bucket: bucket_name,
|
131
|
+
key: file_key,
|
132
|
+
expression_type: "SQL",
|
133
|
+
expression: query,
|
134
|
+
output_serialization: {
|
135
|
+
json: {}
|
136
|
+
}
|
137
|
+
}
|
138
|
+
if file_type == "parquet"
|
139
|
+
options[:input_serialization] = {
|
140
|
+
parquet: {}
|
141
|
+
}
|
142
|
+
elsif file_type == "csv"
|
143
|
+
options[:input_serialization] = {
|
144
|
+
csv: { file_header_info: "USE" }
|
145
|
+
}
|
146
|
+
end
|
147
|
+
options
|
148
|
+
end
|
149
|
+
end
|
150
|
+
end
|
151
|
+
end
|
@@ -0,0 +1,15 @@
|
|
1
|
+
{
|
2
|
+
"data": {
|
3
|
+
"name": "AmazonS3",
|
4
|
+
"title": "Amazon S3",
|
5
|
+
"connector_type": "source",
|
6
|
+
"category": "Data Lake",
|
7
|
+
"documentation_url": "https://docs.mutliwoven.com",
|
8
|
+
"github_issue_label": "source-amazons3",
|
9
|
+
"icon": "icon.svg",
|
10
|
+
"license": "MIT",
|
11
|
+
"release_stage": "alpha",
|
12
|
+
"support_level": "community",
|
13
|
+
"tags": ["language:ruby", "multiwoven"]
|
14
|
+
}
|
15
|
+
}
|
@@ -0,0 +1,51 @@
|
|
1
|
+
{
|
2
|
+
"documentation_url": "https://docs.multiwoven.com/integrations/sources/amazons3",
|
3
|
+
"stream_type": "dynamic",
|
4
|
+
"connector_query_type": "raw_sql",
|
5
|
+
"connection_specification": {
|
6
|
+
"$schema": "http://json-schema.org/draft-07/schema#",
|
7
|
+
"title": "AmazonS3",
|
8
|
+
"type": "object",
|
9
|
+
"required": ["region", "bucket", "access_id", "secret_access", "file_type"],
|
10
|
+
"properties": {
|
11
|
+
"region": {
|
12
|
+
"description": "AWS region",
|
13
|
+
"examples": ["us-east-2"],
|
14
|
+
"type": "string",
|
15
|
+
"title": "Region",
|
16
|
+
"order": 1
|
17
|
+
},
|
18
|
+
"access_id": {
|
19
|
+
"type": "string",
|
20
|
+
"title": "Access Id",
|
21
|
+
"order": 2
|
22
|
+
},
|
23
|
+
"secret_access": {
|
24
|
+
"type": "string",
|
25
|
+
"title": "Secret Access",
|
26
|
+
"multiwoven_secret": true,
|
27
|
+
"order": 3
|
28
|
+
},
|
29
|
+
"bucket": {
|
30
|
+
"description": "Bucket Name",
|
31
|
+
"type": "string",
|
32
|
+
"title": "Bucket",
|
33
|
+
"order": 4
|
34
|
+
},
|
35
|
+
"path": {
|
36
|
+
"description": "Path to csv or parquet files",
|
37
|
+
"examples": ["/path/to/files"],
|
38
|
+
"type": "string",
|
39
|
+
"title": "Path",
|
40
|
+
"order": 5
|
41
|
+
},
|
42
|
+
"file_type": {
|
43
|
+
"description": "The type of file to read",
|
44
|
+
"type": "string",
|
45
|
+
"title": "File Type",
|
46
|
+
"enum": ["csv", "parquet"],
|
47
|
+
"order": 6
|
48
|
+
}
|
49
|
+
}
|
50
|
+
}
|
51
|
+
}
|
@@ -0,0 +1,34 @@
|
|
1
|
+
<svg xmlns="http://www.w3.org/2000/svg" width="428" height="512" viewBox="0 0 428 512">
|
2
|
+
<defs>
|
3
|
+
<style>
|
4
|
+
.cls-1 {
|
5
|
+
fill: #e25444;
|
6
|
+
}
|
7
|
+
|
8
|
+
.cls-1, .cls-2, .cls-3 {
|
9
|
+
fill-rule: evenodd;
|
10
|
+
}
|
11
|
+
|
12
|
+
.cls-2 {
|
13
|
+
fill: #7b1d13;
|
14
|
+
}
|
15
|
+
|
16
|
+
.cls-3 {
|
17
|
+
fill: #58150d;
|
18
|
+
}
|
19
|
+
</style>
|
20
|
+
</defs>
|
21
|
+
<path class="cls-1" d="M378,99L295,257l83,158,34-19V118Z"/>
|
22
|
+
<path class="cls-2" d="M378,99L212,118,127.5,257,212,396l166,19V99Z"/>
|
23
|
+
<path class="cls-3" d="M43,99L16,111V403l27,12L212,257Z"/>
|
24
|
+
<path class="cls-1" d="M42.637,98.667l169.587,47.111V372.444L42.637,415.111V98.667Z"/>
|
25
|
+
<path class="cls-3" d="M212.313,170.667l-72.008-11.556,72.008-81.778,71.83,81.778Z"/>
|
26
|
+
<path class="cls-3" d="M284.143,159.111l-71.919,11.733-71.919-11.733V77.333"/>
|
27
|
+
<path class="cls-3" d="M212.313,342.222l-72.008,13.334,72.008,70.222,71.83-70.222Z"/>
|
28
|
+
<path class="cls-2" d="M212,16L140,54V159l72.224-20.333Z"/>
|
29
|
+
<path class="cls-2" d="M212.224,196.444l-71.919,7.823V309.105l71.919,8.228V196.444Z"/>
|
30
|
+
<path class="cls-2" d="M212.224,373.333L140.305,355.3V458.363L212.224,496V373.333Z"/>
|
31
|
+
<path class="cls-1" d="M284.143,355.3l-71.919,18.038V496l71.919-37.637V355.3Z"/>
|
32
|
+
<path class="cls-1" d="M212.224,196.444l71.919,7.823V309.105l-71.919,8.228V196.444Z"/>
|
33
|
+
<path class="cls-1" d="M212,16l72,38V159l-72-20V16Z"/>
|
34
|
+
</svg>
|
@@ -27,6 +27,8 @@ require "zip"
|
|
27
27
|
require "zendesk_api"
|
28
28
|
require "faraday"
|
29
29
|
require "base64"
|
30
|
+
require "aws-sdk-s3"
|
31
|
+
require "duckdb"
|
30
32
|
require "iterable-api-client"
|
31
33
|
|
32
34
|
# Service
|
@@ -55,6 +57,7 @@ require_relative "integrations/source/databricks/client"
|
|
55
57
|
require_relative "integrations/source/salesforce_consumer_goods_cloud/client"
|
56
58
|
require_relative "integrations/source/aws_athena/client"
|
57
59
|
require_relative "integrations/source/clickhouse/client"
|
60
|
+
require_relative "integrations/source/amazon_s3/client"
|
58
61
|
|
59
62
|
# Destination
|
60
63
|
require_relative "integrations/destination/klaviyo/client"
|
@@ -36,10 +36,12 @@ Gem::Specification.new do |spec|
|
|
36
36
|
spec.add_runtime_dependency "activesupport"
|
37
37
|
spec.add_runtime_dependency "async-websocket"
|
38
38
|
spec.add_runtime_dependency "aws-sdk-athena"
|
39
|
+
spec.add_runtime_dependency "aws-sdk-s3"
|
39
40
|
spec.add_runtime_dependency "csv"
|
40
41
|
spec.add_runtime_dependency "dry-schema"
|
41
42
|
spec.add_runtime_dependency "dry-struct"
|
42
43
|
spec.add_runtime_dependency "dry-types"
|
44
|
+
spec.add_runtime_dependency "duckdb"
|
43
45
|
spec.add_runtime_dependency "git"
|
44
46
|
spec.add_runtime_dependency "google-apis-sheets_v4"
|
45
47
|
spec.add_runtime_dependency "google-cloud-bigquery"
|
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.
|
4
|
+
version: 0.2.0
|
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: 2024-06-
|
11
|
+
date: 2024-06-18 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: activesupport
|
@@ -52,6 +52,20 @@ dependencies:
|
|
52
52
|
- - ">="
|
53
53
|
- !ruby/object:Gem::Version
|
54
54
|
version: '0'
|
55
|
+
- !ruby/object:Gem::Dependency
|
56
|
+
name: aws-sdk-s3
|
57
|
+
requirement: !ruby/object:Gem::Requirement
|
58
|
+
requirements:
|
59
|
+
- - ">="
|
60
|
+
- !ruby/object:Gem::Version
|
61
|
+
version: '0'
|
62
|
+
type: :runtime
|
63
|
+
prerelease: false
|
64
|
+
version_requirements: !ruby/object:Gem::Requirement
|
65
|
+
requirements:
|
66
|
+
- - ">="
|
67
|
+
- !ruby/object:Gem::Version
|
68
|
+
version: '0'
|
55
69
|
- !ruby/object:Gem::Dependency
|
56
70
|
name: csv
|
57
71
|
requirement: !ruby/object:Gem::Requirement
|
@@ -108,6 +122,20 @@ dependencies:
|
|
108
122
|
- - ">="
|
109
123
|
- !ruby/object:Gem::Version
|
110
124
|
version: '0'
|
125
|
+
- !ruby/object:Gem::Dependency
|
126
|
+
name: duckdb
|
127
|
+
requirement: !ruby/object:Gem::Requirement
|
128
|
+
requirements:
|
129
|
+
- - ">="
|
130
|
+
- !ruby/object:Gem::Version
|
131
|
+
version: '0'
|
132
|
+
type: :runtime
|
133
|
+
prerelease: false
|
134
|
+
version_requirements: !ruby/object:Gem::Requirement
|
135
|
+
requirements:
|
136
|
+
- - ">="
|
137
|
+
- !ruby/object:Gem::Version
|
138
|
+
version: '0'
|
111
139
|
- !ruby/object:Gem::Dependency
|
112
140
|
name: git
|
113
141
|
requirement: !ruby/object:Gem::Requirement
|
@@ -519,6 +547,10 @@ files:
|
|
519
547
|
- lib/multiwoven/integrations/protocol/protocol.rb
|
520
548
|
- lib/multiwoven/integrations/rollout.rb
|
521
549
|
- lib/multiwoven/integrations/service.rb
|
550
|
+
- lib/multiwoven/integrations/source/amazon_s3/client.rb
|
551
|
+
- lib/multiwoven/integrations/source/amazon_s3/config/meta.json
|
552
|
+
- lib/multiwoven/integrations/source/amazon_s3/config/spec.json
|
553
|
+
- lib/multiwoven/integrations/source/amazon_s3/icon.svg
|
522
554
|
- lib/multiwoven/integrations/source/aws_athena/client.rb
|
523
555
|
- lib/multiwoven/integrations/source/aws_athena/config/meta.json
|
524
556
|
- lib/multiwoven/integrations/source/aws_athena/config/spec.json
|