egis 1.4.0 → 1.5.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4fa3323e3c02d8537df83f33799f580fb196c255315c9923eb70c0e87bd1f50e
-  data.tar.gz: 9b4054a2201d98f501603cc137c1c7df93fccf105c448b6d4261515c3f53d959
+  metadata.gz: 924727f52ac809df7ab4457a6a0c05469cd6c128adaff339018e964dc9ff4ea0
+  data.tar.gz: c8ba9d090641c899e832b73b062314bd76fa4fb39041fb175d7e3a717e943856
 SHA512:
-  metadata.gz: 3da9b098de6948b584244db63eda8ef4ba3d26ff29806be0e1a3b3469a321bb72d0761bc4bcb415759646a2397c71d32221477cc2dabeb9233dd8cad22300cef
-  data.tar.gz: 5c50463ef9585563b3aa6fefbb89b4df101f93693b7e5d5f423e98fa803e614d7541639db9afe06fc1a4544205d6e3b0cbf3c166112f556f58b8c798b528571d
+  metadata.gz: fa429f550a68475a75f9dd9c0610b0d41c9a8017f0e46c0f9e467cfaeddfb7d6587d6e7df5b4440bbd75e4d4d9a12d5966f40e4ce72cb8abf70e5dc2897f91d8
+  data.tar.gz: 3beac29ce82c267bfc2621e730a8b631f76a94f7426ac7d7901b31d204cda3320d1ab934e9c3e156ee52e5fca315728412ffd5a007bf4623dd2ac077d85399ba

data/lib/egis.rb

@@ -50,8 +50,8 @@ require 'egis/s3_location_parser'
 #
 module Egis
   class << self
-    def configure
-      yield(configuration)
+    def configure(&block)
+      configuration.configure(&block)
     end
 
     # @!visibility private

data/lib/egis/aws_client_provider.rb

@@ -6,25 +6,23 @@ require 'aws-sdk-athena'
 module Egis
   # @!visibility private
   class AwsClientProvider
-    def s3_client
-      Aws::S3::Client.new(client_config)
+    def s3_client(configuration)
+      Aws::S3::Client.new(client_config(configuration))
     end
 
-    def athena_client
-      Aws::Athena::Client.new(client_config)
+    def athena_client(configuration)
+      Aws::Athena::Client.new(client_config(configuration))
     end
 
     private
 
-    def client_config
-      configuration = Egis.configuration
-
-      config = {}
-      config[:region] = configuration.aws_region if configuration.aws_region
-      config[:access_key_id] = configuration.aws_access_key_id if configuration.aws_access_key_id
-      config[:secret_access_key] = configuration.aws_secret_access_key if configuration.aws_secret_access_key
-      config[:profile] = configuration.aws_profile if configuration.aws_profile
-      config
+    def client_config(configuration)
+      {
+        region: configuration.aws_region,
+        access_key_id: configuration.aws_access_key_id,
+        secret_access_key: configuration.aws_secret_access_key,
+        profile: configuration.aws_profile
+      }.compact
     end
   end
 end

data/lib/egis/client.rb

@@ -4,6 +4,9 @@ module Egis
   ##
   # The most fundamental {Egis} class. Provides an interface for executing Athena queries.
   #
+  # @yieldparam config [Egis::Configuration] Egis configuration block, if missing Egis will use global configuration
+  #   provided by {Egis.configure}
+  #
   # See configuration instructions {Egis.configure}.
   #
   # @see Egis.configure
@@ -33,14 +36,17 @@ module Egis
       'CANCELLED' => Egis::QueryStatus::CANCELLED
     }.freeze
 
-    DEFAULT_QUERY_STATUS_BACKOFF = ->(attempt) { 1.5**attempt - 1 }
+    private_constant :QUERY_STATUS_MAPPING
 
-    private_constant :QUERY_STATUS_MAPPING, :DEFAULT_QUERY_STATUS_BACKOFF
+    attr_reader :aws_s3_client
 
-    def initialize(aws_client_provider: Egis::AwsClientProvider.new, s3_location_parser: Egis::S3LocationParser.new)
-      @aws_athena_client = aws_client_provider.athena_client
+    def initialize(aws_client_provider: Egis::AwsClientProvider.new,
+                   s3_location_parser: Egis::S3LocationParser.new,
+                   &block)
+      @configuration = block_given? ? Egis.configuration.dup.configure(&block) : Egis.configuration
+      @aws_athena_client = aws_client_provider.athena_client(configuration)
+      @aws_s3_client = aws_client_provider.s3_client(configuration)
       @s3_location_parser = s3_location_parser
-      @query_status_backoff = Egis.configuration.query_status_backoff || DEFAULT_QUERY_STATUS_BACKOFF
     end
 
     ##
@@ -99,16 +105,17 @@ module Egis
         query_execution.query_execution_id,
         QUERY_STATUS_MAPPING.fetch(query_status),
         query_execution.status.state_change_reason,
-        parse_output_location(query_execution)
+        parse_output_location(query_execution),
+        client: self
       )
     end
 
     private
 
-    attr_reader :aws_athena_client, :s3_location_parser, :query_status_backoff
+    attr_reader :configuration, :aws_athena_client, :s3_location_parser
 
     def query_execution_params(query, work_group, database, output_location)
-      work_group_params = work_group || Egis.configuration.work_group
+      work_group_params = work_group || configuration.work_group
 
       params = {query_string: query}
       params[:work_group] = work_group_params if work_group_params
@@ -128,7 +135,7 @@ module Egis
     def wait_for_query_to_finish(query_id)
       attempt = 1
       loop do
-        sleep(query_status_backoff.call(attempt))
+        sleep(configuration.query_status_backoff.call(attempt))
         status = query_status(query_id)
 
         return status unless status.queued? || status.running?

data/lib/egis/configuration.rb

@@ -8,6 +8,12 @@ module Egis
 
     def initialize
       @logger = Logger.new(STDOUT, level: :info)
+      @query_status_backoff = ->(attempt) { 1.5**attempt - 1 }
+    end
+
+    def configure
+      yield(self)
+      self
     end
   end
 end

data/lib/egis/database.rb

@@ -14,10 +14,10 @@ module Egis
   #   @return [String] Athena database name
   #
   class Database
-    def initialize(name, client: Egis::Client.new, output_downloader: Egis::OutputDownloader.new)
+    def initialize(name, client: Egis::Client.new, output_downloader: Egis::OutputDownloader.new(client.aws_s3_client))
       @client = client
-      @name = name
       @output_downloader = output_downloader
+      @name = name
     end
 
     attr_reader :name
@@ -33,7 +33,7 @@ module Egis
     # @return [Egis::Table]
 
     def table(table_name, table_schema, table_location, **options)
-      Table.new(self, table_name, table_schema, table_location, options: options)
+      Table.new(self, table_name, table_schema, table_location, client: client, options: options)
     end
 
     ##

data/lib/egis/output_downloader.rb

@@ -5,8 +5,8 @@ require 'csv'
 module Egis
   # @!visibility private
   class OutputDownloader
-    def initialize(aws_client_provider: Egis::AwsClientProvider.new)
-      @s3_client = aws_client_provider.s3_client
+    def initialize(aws_s3_client)
+      @s3_client = aws_s3_client
     end
 
     def download(output_location)

data/lib/egis/query_status.rb

@@ -23,7 +23,8 @@ module Egis
     attr_reader :id, :status, :message, :output_location
 
     def initialize(id, status, message, output_location,
-                   output_downloader: Egis::OutputDownloader.new,
+                   client: Egis::Client.new,
+                   output_downloader: Egis::OutputDownloader.new(client.aws_s3_client),
                    output_parser: Egis::OutputParser.new)
       raise ArgumentError, "Unsupported status #{status}" unless STATUSES.include?(status)
 
@@ -51,6 +52,10 @@ module Egis
       status == RUNNING
     end
 
+    def cancelled?
+      status == CANCELLED
+    end
+
     def in_progress?
       [RUNNING, QUEUED].include?(status)
     end

data/lib/egis/s3_cleaner.rb

@@ -3,8 +3,8 @@
 module Egis
   # @!visibility private
   class S3Cleaner
-    def initialize(aws_client_provider: Egis::AwsClientProvider.new)
-      @s3_client = aws_client_provider.s3_client
+    def initialize(aws_s3_client)
+      @s3_client = aws_s3_client
     end
 
     def delete(bucket, prefix)

data/lib/egis/table.rb

@@ -17,11 +17,13 @@ module Egis
     DEFAULT_OPTIONS = {format: :tsv}.freeze
 
     def initialize(database, name, schema, location, options: {},
+                   client: Egis::Client.new,
                    partitions_generator: Egis::PartitionsGenerator.new,
                    table_ddl_generator: Egis::TableDDLGenerator.new,
-                   output_downloader: Egis::OutputDownloader.new,
+                   output_downloader: Egis::OutputDownloader.new(client.aws_s3_client),
                    output_parser: Egis::OutputParser.new,
-                   table_data_wiper: Egis::TableDataWiper.new)
+                   s3_cleaner: Egis::S3Cleaner.new(client.aws_s3_client),
+                   table_data_wiper: Egis::TableDataWiper.new(s3_cleaner: s3_cleaner))
       @database = database
       @name = name
       @schema = schema

data/lib/egis/testing/testing_mode.rb

@@ -6,7 +6,7 @@ module Egis
     class TestingMode
       def initialize(test_id, s3_bucket,
                      client: Egis::Client.new,
-                     output_downloader: Egis::OutputDownloader.new,
+                     output_downloader: Egis::OutputDownloader.new(client.aws_s3_client),
                      s3_location_parser: Egis::S3LocationParser.new)
         @test_id = test_id
         @s3_bucket = s3_bucket

data/lib/egis/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Egis
-  VERSION = '1.4.0'
+  VERSION = '1.5.0'
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: egis
 version: !ruby/object:Gem::Version
-  version: 1.4.0
+  version: 1.5.0
 platform: ruby
 authors:
 - Agnieszka Czereba
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-03-12 00:00:00.000000000 Z
+date: 2021-05-14 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: aws-sdk-athena
@@ -99,7 +99,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.1.4
+rubygems_version: 3.1.6
 signing_key: 
 specification_version: 4
 summary: A handy wrapper for AWS Athena Ruby SDK.