avro_turf 0.8.1 → 0.11.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 8e58564680b9399ae8df438412385f23bdabd46cee8deafc0dfa1c8b827d7792
-  data.tar.gz: 1df38f38434777fab06fddec69a8834442a4814e269edbed154fc74153f6b198
+  metadata.gz: 1e2ee4d9598bcaa2ec5738a1130fae0b19be7b5e7250f27540313b7471f97e23
+  data.tar.gz: 0b441cb30a153958c2ea283300a1a05c26218e2a97cb807fce36f8ad9d0240da
 SHA512:
-  metadata.gz: 6e47f299a673911614be989feefb56f2cd48be6a556e240919ad23b13b55928c3cd0837d5d5f43aa4c2f72e1f416465f811de24ae7e344f858dc147dd23be136
-  data.tar.gz: 187c4f087cf7ed656ef3bfed6bf0593938f57da698b0b99550e476a059e529cf355f33ce5bc1839680c1c0513c730e285a117f33ec9e6947c476f0301ab3c597
+  metadata.gz: 12779eac5c325752cfa1be34da94ef5f332490cda4ff0aef29529a00557008cecf39592396a3f3525a2a12cb67a46744800781e0da69d4bf02511f5a2284e5e7
+  data.tar.gz: a2e4c84fb338d62296aefb8ae8c206c262d756475ec11bd9cead17bfe6015ea069ff36a891df3381385fa650cb4b5ea1584855fccc5294f5910ab34563ad973a

data/.circleci/config.yml

@@ -0,0 +1,36 @@
+version: 2
+jobs:
+  build:
+    environment:
+      CIRCLE_ARTIFACTS: /tmp/circleci-artifacts
+      CIRCLE_TEST_REPORTS: /tmp/circleci-test-results
+    docker:
+    - image: circleci/ruby:2.6.2
+    steps:
+    - checkout
+    - run: mkdir -p $CIRCLE_ARTIFACTS $CIRCLE_TEST_REPORTS
+    - restore_cache:
+        keys:
+        # This branch if available
+        - v1-dep-{{ .Branch }}-
+        # Default branch if not
+        - v1-dep-master-
+        # Any branch if there are none on the default branch - this should be unnecessary if you have your default branch configured correctly
+        - v1-dep-
+    - run: gem install bundler --no-document
+    - run: 'bundle check --path=vendor/bundle || bundle install --path=vendor/bundle --jobs=4 --retry=3'
+    # Save dependency cache
+    - save_cache:
+        key: v1-dep-{{ .Branch }}-{{ epoch }}
+        paths:
+        - vendor/bundle
+        - ~/.bundle
+    - run: mkdir -p $CIRCLE_TEST_REPORTS/rspec
+    - run:
+        command: bundle exec rspec --color --require spec_helper --format progress
+    - store_test_results:
+        path: /tmp/circleci-test-results
+    - store_artifacts:
+        path: /tmp/circleci-artifacts
+    - store_artifacts:
+        path: /tmp/circleci-test-results

data/.github/workflows/ruby.yml

@@ -0,0 +1,20 @@
+name: Ruby
+
+on: [push]
+
+jobs:
+  build:
+
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v1
+    - name: Set up Ruby 2.6
+      uses: actions/setup-ruby@v1
+      with:
+        ruby-version: 2.6.x
+    - name: Build and test with RSpec
+      run: |
+        gem install bundler
+        bundle install --jobs 4 --retry 3
+        bundle exec rspec

data/CHANGELOG.md

@@ -1,5 +1,25 @@
 # avro_turf
 
+## Unreleased
+
+## v0.11.0
+
+- Add proxy support (#107)
+- Adding support for client certs (#109)
+
+## v0.10.0
+
+- Add more disk caching (#103)
+- Include schema information when decoding (#100, #101, #104)
+
+## v0.9.0
+
+- Compatibility with Avro v1.9.0 (#94)
+- Disable the auto registeration of schema (#95)
+- abstracted caching from CachedConfluentSchemaRegistry (#74)
+- Load avro-patches if installed to silence deprecation errors (#85)
+- Make schema store to be thread safe (#92)
+
 ## v0.8.1
 
 - Allow accessing schema store from outside AvroTurf (#68).

data/Gemfile

@@ -2,6 +2,3 @@ source 'https://rubygems.org'
 
 # Specify your gem's dependencies in avro_turf.gemspec
 gemspec
-
-# Used by CircleCI to format RSpec results.
-gem 'rspec_junit_formatter', :git => 'git@github.com:circleci/rspec_junit_formatter.git'

data/README.md

@@ -124,9 +124,29 @@ avro = AvroTurf::Messaging.new(registry_url: "http://my-registry:8081/")
 # time a schema is used.
 data = avro.encode({ "title" => "hello, world" }, schema_name: "greeting")
 
+# If you don't want to automatically register new schemas, you can pass explicitly
+# subject and version to specify which schema should be used for encoding.
+# It will fetch that schema from the registry and cache it. Subsequent instances
+# of the same schema version will be served by the cache.
+data = avro.encode({ "title" => "hello, world" }, subject: 'greeting', version: 1)
+
+# You can also pass explicitly schema_id to specify which schema
+# should be used for encoding.
+# It will fetch that schema from the registry and cache it. Subsequent instances
+# of the same schema version will be served by the cache.
+data = avro.encode({ "title" => "hello, world" }, schema_id: 2)
+
 # When decoding, the schema will be fetched from the registry and cached. Subsequent
 # instances of the same schema id will be served by the cache.
 avro.decode(data) #=> { "title" => "hello, world" }
+
+# If you want to get decoded message as well as the schema used to encode the message,
+# you can use `#decode_message` method.
+result = avro.decode_message(data)
+result.message       #=> { "title" => "hello, world" }
+result.schema_id     #=> 3
+result.writer_schema #=> #<Avro::Schema: ...>
+result.reader_schema #=> nil
 ```
 
 ### Confluent Schema Registry Client

data/avro_turf.gemspec

@@ -17,16 +17,17 @@ Gem::Specification.new do |spec|
   spec.test_files    = spec.files.grep(%r{^(test|spec|features)/})
   spec.require_paths = ["lib"]
 
-  spec.add_dependency "avro", ">= 1.7.7", "< 1.9"
+  spec.add_dependency "avro", ">= 1.7.7", "< 1.10"
   spec.add_dependency "excon", "~> 0.45"
 
-  spec.add_development_dependency "bundler", "~> 1.7"
-  spec.add_development_dependency "rake", "~> 10.0"
-  spec.add_development_dependency "rspec", "~> 3.2.0"
-  spec.add_development_dependency "fakefs", "~> 0.6.7"
+  spec.add_development_dependency "bundler", "~> 2.0"
+  spec.add_development_dependency "rake", "~> 13.0"
+  spec.add_development_dependency "rspec", "~> 3.2"
+  spec.add_development_dependency "fakefs", "~> 0.20.0"
   spec.add_development_dependency "webmock"
   spec.add_development_dependency "sinatra"
   spec.add_development_dependency "json_spec"
+  spec.add_development_dependency "rack-test"
 
   spec.post_install_message = %{
 avro_turf v0.8.0 deprecates the names AvroTurf::SchemaRegistry,

data/lib/avro_turf.rb

@@ -1,9 +1,18 @@
+begin
+  require 'avro-patches'
+rescue LoadError
+  false
+end
 require 'avro_turf/version'
 require 'avro'
 require 'json'
 require 'avro_turf/schema_store'
 require 'avro_turf/core_ext'
-require 'avro_turf/schema_to_avro_patch'
+
+# check for something that indicates Avro v1.9.0 or later
+unless defined?(::Avro::LogicalTypes)
+  require 'avro_turf/schema_to_avro_patch'
+end
 
 class AvroTurf
   class Error < StandardError; end

data/lib/avro_turf/cached_confluent_schema_registry.rb

@@ -1,16 +1,23 @@
 require 'avro_turf/confluent_schema_registry'
+require 'avro_turf/in_memory_cache'
+require 'avro_turf/disk_cache'
 
 # Caches registrations and lookups to the schema registry in memory.
 class AvroTurf::CachedConfluentSchemaRegistry
 
-  def initialize(upstream)
+  # Instantiate a new CachedConfluentSchemaRegistry instance with the given configuration.
+  # By default, uses a provided InMemoryCache to prevent repeated calls to the upstream registry.
+  #
+  # upstream  - The upstream schema registry object that fully responds to all methods in the
+  #             AvroTurf::ConfluentSchemaRegistry interface.
+  # cache     - Optional user provided Cache object that responds to all methods in the AvroTurf::InMemoryCache interface.
+  def initialize(upstream, cache: nil)
     @upstream = upstream
-    @schemas_by_id = {}
-    @ids_by_schema = {}
+    @cache = cache || AvroTurf::InMemoryCache.new()
   end
 
   # Delegate the following methods to the upstream
-  %i(subjects subject_versions subject_version check compatible?
+  %i(subjects subject_versions check compatible?
      global_config update_global_config subject_config update_subject_config).each do |name|
     define_method(name) do |*args|
       instance_variable_get(:@upstream).send(name, *args)
@@ -18,10 +25,15 @@ class AvroTurf::CachedConfluentSchemaRegistry
   end
 
   def fetch(id)
-    @schemas_by_id[id] ||= @upstream.fetch(id)
+    @cache.lookup_by_id(id) || @cache.store_by_id(id, @upstream.fetch(id))
   end
 
   def register(subject, schema)
-    @ids_by_schema[subject + schema.to_s] ||= @upstream.register(subject, schema)
+    @cache.lookup_by_schema(subject, schema) || @cache.store_by_schema(subject, schema, @upstream.register(subject, schema))
+  end
+
+  def subject_version(subject, version = 'latest')
+    @cache.lookup_by_version(subject, version) ||
+      @cache.store_by_version(subject, version, @upstream.subject_version(subject, version))
   end
 end

data/lib/avro_turf/confluent_schema_registry.rb

@@ -3,11 +3,30 @@ require 'excon'
 class AvroTurf::ConfluentSchemaRegistry
   CONTENT_TYPE = "application/vnd.schemaregistry.v1+json".freeze
 
-  def initialize(url, logger: Logger.new($stdout))
+  def initialize(
+    url,
+    logger: Logger.new($stdout),
+    proxy: nil,
+    client_cert: nil,
+    client_key: nil,
+    client_key_pass: nil,
+    client_cert_data: nil,
+    client_key_data: nil
+  )
     @logger = logger
-    @connection = Excon.new(url, headers: {
-      "Content-Type" => CONTENT_TYPE,
-    })
+    headers = {
+      "Content-Type" => CONTENT_TYPE
+    }
+    headers[:proxy] = proxy if proxy&.present?
+    @connection = Excon.new(
+      url,
+      headers: headers,
+      client_cert: client_cert,
+      client_key: client_key,
+      client_key_pass: client_key_pass,
+      client_cert_data: client_cert_data,
+      client_key_data: client_key_data
+    )
   end
 
   def fetch(id)

data/lib/avro_turf/disk_cache.rb

@@ -0,0 +1,83 @@
+# A cache for the CachedConfluentSchemaRegistry.
+# Extends the InMemoryCache to provide a write-thru to disk for persistent cache.
+class AvroTurf::DiskCache < AvroTurf::InMemoryCache
+
+  def initialize(disk_path)
+    super()
+
+    # load the write-thru cache on startup, if it exists
+    @schemas_by_id_path = File.join(disk_path, 'schemas_by_id.json')
+    @schemas_by_id = JSON.parse(File.read(@schemas_by_id_path)) if File.exist?(@schemas_by_id_path)
+
+    @ids_by_schema_path = File.join(disk_path, 'ids_by_schema.json')
+    @ids_by_schema = JSON.parse(File.read(@ids_by_schema_path)) if File.exist?(@ids_by_schema_path)
+
+    @schemas_by_subject_version_path = File.join(disk_path, 'schemas_by_subject_version.json')
+    @schemas_by_subject_version = {}
+  end
+
+  # override
+  # the write-thru cache (json) does not store keys in numeric format
+  # so, convert id to a string for caching purposes
+  def lookup_by_id(id)
+    super(id.to_s)
+  end
+
+  # override to include write-thru cache after storing result from upstream
+  def store_by_id(id, schema)
+    # must return the value from storing the result (i.e. do not return result from file write)
+    value = super(id.to_s, schema)
+    File.write(@schemas_by_id_path, JSON.pretty_generate(@schemas_by_id))
+    return value
+  end
+
+  # override to include write-thru cache after storing result from upstream
+  def store_by_schema(subject, schema, id)
+    # must return the value from storing the result (i.e. do not return result from file write)
+    value = super
+    File.write(@ids_by_schema_path, JSON.pretty_generate(@ids_by_schema))
+    return value
+  end
+
+  # checks instance var (in-memory cache) for schema
+  # checks disk cache if in-memory cache doesn't exists
+  # if file exists but no in-memory cache, read from file and sync in-memory cache
+  # finally, if file doesn't exist return nil
+  def lookup_by_version(subject, version)
+    key = "#{subject}#{version}"
+    schema = @schemas_by_subject_version[key]
+
+    return schema unless schema.nil?
+
+    hash = JSON.parse(File.read(@schemas_by_subject_version_path)) if File.exist?(@schemas_by_subject_version_path)
+    if hash
+      @schemas_by_subject_version = hash
+      @schemas_by_subject_version[key]
+    end
+  end
+
+  # check if file exists and parse json into a hash
+  # if file exists take json and overwite/insert schema at key
+  # if file doesn't exist create new hash
+  # write the new/updated hash to file
+  # update instance var (in memory-cache) to match
+  def store_by_version(subject, version, schema)
+    key = "#{subject}#{version}"
+    hash = JSON.parse(File.read(@schemas_by_subject_version_path)) if File.exist?(@schemas_by_subject_version_path)
+    hash = if hash
+             hash[key] = schema
+             hash
+           else
+             {key => schema}
+           end
+
+    write_to_disk_cache(@schemas_by_subject_version_path, hash)
+
+    @schemas_by_subject_version = hash
+    @schemas_by_subject_version[key]
+  end
+
+  private def write_to_disk_cache(path, hash)
+    File.write(path, JSON.pretty_generate(hash))
+  end
+end

data/lib/avro_turf/in_memory_cache.rb

@@ -0,0 +1,38 @@
+# A cache for the CachedConfluentSchemaRegistry.
+# Simply stores the schemas and ids in in-memory hashes.
+class AvroTurf::InMemoryCache
+
+  def initialize
+    @schemas_by_id = {}
+    @ids_by_schema = {}
+    @schema_by_subject_version = {}
+  end
+
+  def lookup_by_id(id)
+    @schemas_by_id[id]
+  end
+
+  def store_by_id(id, schema)
+    @schemas_by_id[id] = schema
+  end
+
+  def lookup_by_schema(subject, schema)
+    key = subject + schema.to_s
+    @ids_by_schema[key]
+  end
+
+  def store_by_schema(subject, schema, id)
+    key = subject + schema.to_s
+    @ids_by_schema[key] = id
+  end
+
+  def lookup_by_version(subject, version)
+    key = "#{subject}#{version}"
+    @schema_by_subject_version[key]
+  end
+
+  def store_by_version(subject, version, schema)
+    key = "#{subject}#{version}"
+    @schema_by_subject_version[key] = schema
+  end
+end

data/lib/avro_turf/messaging.rb

@@ -21,21 +21,53 @@ class AvroTurf
   # 1: https://github.com/confluentinc/schema-registry
   class Messaging
     MAGIC_BYTE = [0].pack("C").freeze
+    DecodedMessage = Struct.new(:schema_id, :writer_schema, :reader_schema, :message)
+    private_constant(:DecodedMessage)
 
     # Instantiate a new Messaging instance with the given configuration.
     #
-    # registry     - A schema registry object that responds to all methods in the
-    #                AvroTurf::ConfluentSchemaRegistry interface.
-    # registry_url - The String URL of the schema registry that should be used.
-    # schema_store - A schema store object that responds to #find(schema_name, namespace).
-    # schemas_path - The String file system path where local schemas are stored.
-    # namespace    - The String default schema namespace.
-    # logger       - The Logger that should be used to log information (optional).
-    def initialize(registry: nil, registry_url: nil, schema_store: nil, schemas_path: nil, namespace: nil, logger: nil)
+    # registry          - A schema registry object that responds to all methods in the
+    #                     AvroTurf::ConfluentSchemaRegistry interface.
+    # registry_url      - The String URL of the schema registry that should be used.
+    # schema_store      - A schema store object that responds to #find(schema_name, namespace).
+    # schemas_path      - The String file system path where local schemas are stored.
+    # namespace         - The String default schema namespace.
+    # logger            - The Logger that should be used to log information (optional).
+    # proxy             - Forward the request via  proxy (optional).
+    # client_cert       - Name of file containing client certificate (optional).
+    # client_key        - Name of file containing client private key to go with client_cert (optional).
+    # client_key_pass   - Password to go with client_key (optional).
+    # client_cert_data  - In-memory client certificate (optional).
+    # client_key_data   - In-memory client private key to go with client_cert_data (optional).
+    def initialize(
+      registry: nil,
+      registry_url: nil,
+      schema_store: nil,
+      schemas_path: nil,
+      namespace: nil,
+      logger: nil,
+      proxy: nil,
+      client_cert: nil,
+      client_key: nil,
+      client_key_pass: nil,
+      client_cert_data: nil,
+      client_key_data: nil
+    )
       @logger = logger || Logger.new($stderr)
       @namespace = namespace
       @schema_store = schema_store || SchemaStore.new(path: schemas_path || DEFAULT_SCHEMAS_PATH)
-      @registry = registry || CachedConfluentSchemaRegistry.new(ConfluentSchemaRegistry.new(registry_url, logger: @logger))
+      @registry = registry || CachedConfluentSchemaRegistry.new(
+        ConfluentSchemaRegistry.new(
+          registry_url,
+          logger: @logger,
+          proxy: proxy,
+          client_cert: client_cert,
+          client_key: client_key,
+          client_key_pass: client_key_pass,
+          client_cert_data: client_cert_data,
+          client_key_data: client_key_data
+        )
+      )
       @schemas_by_id = {}
     end
 
@@ -46,14 +78,24 @@ class AvroTurf
     # schema_name - The String name of the schema that should be used to encode
     #               the data.
     # namespace   - The namespace of the schema (optional).
+    # subject     - The subject name the schema should be registered under in
+    #               the schema registry (optional).
+    # version     - The integer version of the schema that should be used to decode
+    #               the data. Must match the schema used when encoding (optional).
+    # schema_id   - The integer id of the schema that should be used to encode
+    #               the data.
     #
     # Returns the encoded data as a String.
-    def encode(message, schema_name: nil, namespace: @namespace, subject: nil)
-      schema = @schema_store.find(schema_name, namespace)
-
-      # Schemas are registered under the full name of the top level Avro record
-      # type, or `subject` if it's provided.
-      schema_id = @registry.register(subject || schema.fullname, schema)
+    def encode(message, schema_name: nil, namespace: @namespace, subject: nil, version: nil, schema_id: nil)
+      schema_id, schema = if schema_id
+        fetch_schema_by_id(schema_id)
+      elsif subject && version
+        fetch_schema(subject, version)
+      elsif schema_name
+        register_schema(subject, schema_name, namespace)
+      else
+        raise ArgumentError.new('Neither schema_name nor schema_id nor subject + version provided to determine the schema.')
+      end
 
       stream = StringIO.new
       writer = Avro::IO::DatumWriter.new(schema)
@@ -69,6 +111,12 @@ class AvroTurf
       writer.write(message, encoder)
 
       stream.string
+    rescue Excon::Error::NotFound
+      if schema_id
+        raise SchemaNotFoundError.new("Schema with id: #{schema_id} is not found on registry")
+      else
+        raise SchemaNotFoundError.new("Schema with subject: `#{subject}` version: `#{version}` is not found on registry")
+      end
     end
 
     # Decodes data into the original message.
@@ -80,6 +128,20 @@ class AvroTurf
     #
     # Returns the decoded message.
     def decode(data, schema_name: nil, namespace: @namespace)
+      decode_message(data, schema_name: schema_name, namespace: namespace).message
+    end
+
+    # Decodes data into the original message.
+    #
+    # data        - A String containing encoded data.
+    # schema_name - The String name of the schema that should be used to decode
+    #               the data. Must match the schema used when encoding (optional).
+    # namespace   - The namespace of the schema (optional).
+    #
+    # Returns Struct with the next attributes:
+    #   schema_id  - The integer id of schema used to encode the message
+    #   message    - The decoded message
+    def decode_message(data, schema_name: nil, namespace: @namespace)
       readers_schema = schema_name && @schema_store.find(schema_name, namespace)
       stream = StringIO.new(data)
       decoder = Avro::IO::BinaryDecoder.new(stream)
@@ -100,7 +162,38 @@ class AvroTurf
       end
 
       reader = Avro::IO::DatumReader.new(writers_schema, readers_schema)
-      reader.read(decoder)
+      message = reader.read(decoder)
+
+      DecodedMessage.new(schema_id, writers_schema, readers_schema, message)
+    rescue Excon::Error::NotFound
+      raise SchemaNotFoundError.new("Schema with id: #{schema_id} is not found on registry")
+    end
+
+    private
+
+    # Providing subject and version to determine the schema,
+    # which skips the auto registeration of schema on the schema registry.
+    # Fetch the schema from registry with the provided subject name and version.
+    def fetch_schema(subject, version)
+      schema_data = @registry.subject_version(subject, version)
+      schema_id = schema_data.fetch('id')
+      schema = Avro::Schema.parse(schema_data.fetch('schema'))
+      [schema_id, schema]
+    end
+
+    # Fetch the schema from registry with the provided schema_id.
+    def fetch_schema_by_id(schema_id)
+      schema_json = @registry.fetch(schema_id)
+      schema = Avro::Schema.parse(schema_json)
+      [schema_id, schema]
+    end
+
+    # Schemas are registered under the full name of the top level Avro record
+    # type, or `subject` if it's provided.
+    def register_schema(subject, schema_name, namespace)
+      schema = @schema_store.find(schema_name, namespace)
+      schema_id = @registry.register(subject || schema.fullname, schema)
+      [schema_id, schema]
     end
   end
 end