avro_turf 0.11.0 → 1.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 1e2ee4d9598bcaa2ec5738a1130fae0b19be7b5e7250f27540313b7471f97e23
-  data.tar.gz: 0b441cb30a153958c2ea283300a1a05c26218e2a97cb807fce36f8ad9d0240da
+  metadata.gz: 8fc2d29f1112e649cc2cd5ec96f2126a0ea5fa8f46ebee01d908cdceeac8da0d
+  data.tar.gz: e1989abd9d9b0e5db24f360e96e7bce7ce18e5f488cc794d25a2b3fae6102e66
 SHA512:
-  metadata.gz: 12779eac5c325752cfa1be34da94ef5f332490cda4ff0aef29529a00557008cecf39592396a3f3525a2a12cb67a46744800781e0da69d4bf02511f5a2284e5e7
-  data.tar.gz: a2e4c84fb338d62296aefb8ae8c206c262d756475ec11bd9cead17bfe6015ea069ff36a891df3381385fa650cb4b5ea1584855fccc5294f5910ab34563ad973a
+  metadata.gz: f562a66ba746d2c0e4bee8b8f3a0e7fcf65900f261eca444986c8ff6c68231e67aa9c4dc659213721d0d09c65fd9273e12e158938c0773baa48cf86598d6bef9
+  data.tar.gz: 53e3640900a3b64038fe063f5dadab2049417490be1833bb5f17124307abe628af64ea0c68b5bdc2bd4ef2f6e58bbfac287696eaf5ece9a35ddb7c2ab8e5127e

data/.github/workflows/ruby.yml

@@ -1,6 +1,6 @@
 name: Ruby
 
-on: [push]
+on: [push, pull_request]
 
 jobs:
   build:

data/.github/workflows/stale.yml

@@ -0,0 +1,19 @@
+name: Mark stale issues and pull requests
+
+on:
+  schedule:
+  - cron: "0 0 * * *"
+
+jobs:
+  stale:
+
+    runs-on: ubuntu-latest
+
+    steps:
+    - uses: actions/stale@v1
+      with:
+        repo-token: ${{ secrets.GITHUB_TOKEN }}
+        stale-issue-message: 'Stale issue message'
+        stale-pr-message: 'Stale pull request message'
+        stale-issue-label: 'no-issue-activity'
+        stale-pr-label: 'no-pr-activity'

data/CHANGELOG.md

@@ -1,7 +1,20 @@
-# avro_turf
+# AvroTurf
 
 ## Unreleased
 
+## v1.2.0
+
+- Expose `fetch_schema`, `fetch_schema_by_id` and `register_schema` schema in `Messaging` interface (#117, #119)
+- Add ability to validate message before encoding in `Messaging#encode` interface (#116, #118)
+
+## v1.1.0
+
+- Compatibility with Avro v1.10.x.
+
+## v1.0.0
+
+- Stop caching nested sub-schemas (#111)
+
 ## v0.11.0
 
 - Add proxy support (#107)

data/README.md

@@ -16,6 +16,48 @@ These classes have been renamed to `AvroTurf::ConfluentSchemaRegistry`,
 
 The aliases for the original names will be removed in a future release.
 
+## Note about finding nested schemas
+
+As of AvroTurf version 0.12.0, only top-level schemas that have their own .avsc file will be loaded and resolvable by the `AvroTurf::SchemaStore#find` method. This change will likely not affect most users. However, if you use `AvroTurf::SchemaStore#load_schemas!` to pre-cache all your schemas and then rely on `AvroTurf::SchemaStore#find` to access nested schemas that are not defined by their own .avsc files, your code may stop working when you upgrade to v0.12.0.
+
+As an example, if you have a `person` schema (defined in `my/schemas/contacts/person.avsc`) that defines a nested `address` schema like this:
+
+```json
+{
+  "name": "person",
+  "namespace": "contacts",
+  "type": "record",
+  "fields": [
+    {
+      "name": "address",
+      "type": {
+        "name": "address",
+        "type": "record",
+        "fields": [
+          { "name": "addr1", "type": "string" },
+          { "name": "addr2", "type": "string" },
+          { "name": "city", "type": "string" },
+          { "name": "zip", "type": "string" }
+        ]
+      }
+    }
+  ]
+}
+```
+...this will no longer work in v0.12.0:
+```ruby
+store = AvroTurf::SchemaStore.new(path: 'my/schemas')
+store.load_schemas!
+
+# Accessing 'person' is correct and works fine.
+person = store.find('person', 'contacts') # my/schemas/contacts/person.avsc exists
+
+# Trying to access 'address' raises AvroTurf::SchemaNotFoundError
+address = store.find('address', 'contacts') # my/schemas/contacts/address.avsc is not found
+```
+
+For details and context, see [this pull request](https://github.com/dasch/avro_turf/pull/111).
+
 ## Installation
 
 Add this line to your application's Gemfile:
@@ -136,6 +178,10 @@ data = avro.encode({ "title" => "hello, world" }, subject: 'greeting', version:
 # of the same schema version will be served by the cache.
 data = avro.encode({ "title" => "hello, world" }, schema_id: 2)
 
+# Message can be validated before encoding to get a description of problem through
+# Avro::SchemaValidator::ValidationError exception
+data = avro.encode({ "titl" => "hello, world" }, schema_name: "greeting", validate: true)
+
 # 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" }
@@ -147,6 +193,20 @@ result.message       #=> { "title" => "hello, world" }
 result.schema_id     #=> 3
 result.writer_schema #=> #<Avro::Schema: ...>
 result.reader_schema #=> nil
+
+# You can also work with schema through this interface:
+# Fetch latest schema for subject from registry
+schema, schema_id = avro.fetch_schema(subject: 'greeting')
+# Fetch specific version
+schema, schema_id = avro.fetch_schema(subject: 'greeting', version: 1)
+# Fetch schema by id
+schema, schema_id = avro.fetch_schema_by_id(3)
+# Register schema fetched from store by name
+schema, schema_id = avro.register_schema(schema_name: 'greeting')
+# Specify namespace (same as schema_name: 'somewhere.greeting')
+schema, schema_id = avro.register_schema(schema_name: 'greeting', namespace: 'somewhere')
+# Customize subject under which to register schema
+schema, schema_id = avro.register_schema(schema_name: 'greeting', namespace: 'somewhere', subject: 'test')
 ```
 
 ### Confluent Schema Registry Client

data/avro_turf.gemspec

@@ -17,8 +17,8 @@ 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", "< 1.10"
-  spec.add_dependency "excon", "~> 0.45"
+  spec.add_dependency "avro", ">= 1.7.7", "< 1.11"
+  spec.add_dependency "excon", "~> 0.71"
 
   spec.add_development_dependency "bundler", "~> 2.0"
   spec.add_development_dependency "rake", "~> 13.0"

data/lib/avro_turf/messaging.rb

@@ -84,19 +84,26 @@ class AvroTurf
     #               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.
+    # validate    - The boolean for performing complete message validation before
+    #               encoding it, Avro::SchemaValidator::ValidationError with
+    #               a descriptive message will be raised in case of invalid message.
     #
     # Returns the encoded data as a String.
-    def encode(message, schema_name: nil, namespace: @namespace, subject: nil, version: nil, schema_id: nil)
-      schema_id, schema = if schema_id
+    def encode(message, schema_name: nil, namespace: @namespace, subject: nil, version: nil, schema_id: nil, validate: false)
+      schema, schema_id = if schema_id
         fetch_schema_by_id(schema_id)
       elsif subject && version
-        fetch_schema(subject, version)
+        fetch_schema(subject: subject, version: version)
       elsif schema_name
-        register_schema(subject, schema_name, namespace)
+        register_schema(subject: subject, schema_name: schema_name, namespace: namespace)
       else
         raise ArgumentError.new('Neither schema_name nor schema_id nor subject + version provided to determine the schema.')
       end
 
+      if validate
+        Avro::SchemaValidator.validate!(schema, message, recursive: true, encoded: false, fail_on_extra_fields: true)
+      end
+
       stream = StringIO.new
       writer = Avro::IO::DatumWriter.new(schema)
       encoder = Avro::IO::BinaryEncoder.new(stream)
@@ -169,31 +176,29 @@ class AvroTurf
       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)
+    def fetch_schema(subject:, version: 'latest')
       schema_data = @registry.subject_version(subject, version)
       schema_id = schema_data.fetch('id')
       schema = Avro::Schema.parse(schema_data.fetch('schema'))
-      [schema_id, schema]
+      [schema, schema_id]
     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]
+      [schema, schema_id]
     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)
+    def register_schema(schema_name:, subject: nil, namespace: nil)
       schema = @schema_store.find(schema_name, namespace)
       schema_id = @registry.register(subject || schema.fullname, schema)
-      [schema_id, schema]
+      [schema, schema_id]
     end
   end
 end

data/lib/avro_turf/schema_store.rb

@@ -46,26 +46,45 @@ class AvroTurf::SchemaStore
 
   # Loads single schema
   # Such method is not thread-safe, do not call it of from mutex synchronization routine
-  def load_schema!(fullname, namespace = nil)
+  def load_schema!(fullname, namespace = nil, local_schemas_cache = {})
     *namespace, schema_name = fullname.split(".")
     schema_path = File.join(@path, *namespace, schema_name + ".avsc")
     schema_json = JSON.parse(File.read(schema_path))
-    schema = Avro::Schema.real_parse(schema_json, @schemas)
 
+    schema = Avro::Schema.real_parse(schema_json, local_schemas_cache)
+
+    # Don't cache the parsed schema until after its fullname is validated
     if schema.respond_to?(:fullname) && schema.fullname != fullname
       raise AvroTurf::SchemaError, "expected schema `#{schema_path}' to define type `#{fullname}'"
     end
 
+    # Cache only this new top-level schema by its fullname. It's critical
+    # not to make every sub-schema resolvable at the top level here because
+    # multiple different avsc files may define the same sub-schema, and
+    # if we share the @schemas cache across all parsing contexts, the Avro
+    # gem will raise an Avro::SchemaParseError when parsing another avsc
+    # file that contains a subschema with the same fullname as one
+    # encountered previously in a different file:
+    # <Avro::SchemaParseError: The name "foo.bar" is already in use.>
+    # Essentially, the only schemas that should be resolvable in @schemas
+    # are those that have their own .avsc files on disk.
+    @schemas[fullname] = schema
+
     schema
   rescue ::Avro::SchemaParseError => e
     # This is a hack in order to figure out exactly which type was missing. The
     # Avro gem ought to provide this data directly.
     if e.to_s =~ /"([\w\.]+)" is not a schema we know about/
-      load_schema!($1)
+      # Try to first resolve a referenced schema from disk.
+      # If this is successful, the Avro gem will have mutated the
+      # local_schemas_cache, adding all the new schemas it found.
+      load_schema!($1, nil, local_schemas_cache)
 
-      # Re-resolve the original schema now that the dependency has been resolved.
-      @schemas.delete(fullname)
-      load_schema!(fullname)
+      # Attempt to re-parse the original schema now that the dependency
+      # has been resolved and use the now-updated local_schemas_cache to
+      # pick up where we left off.
+      local_schemas_cache.delete(fullname)
+      load_schema!(fullname, nil, local_schemas_cache)
     else
       raise
     end

data/lib/avro_turf/version.rb

@@ -1,3 +1,3 @@
 class AvroTurf
-  VERSION = "0.11.0"
+  VERSION = "1.2.0"
 end

data/spec/messaging_spec.rb

@@ -297,4 +297,106 @@ describe AvroTurf::Messaging do
       end
     end
   end
+
+  context "validating" do
+    subject(:encode){ avro.encode(message, schema_name: "person", validate: true) }
+
+    context "for correct message" do
+      it { expect { encode }.not_to raise_error }
+    end
+
+    context "when message has wrong type" do
+      let(:message) { { "full_name" => 123 } }
+
+      it { expect { encode }.to raise_error(Avro::SchemaValidator::ValidationError, /\.full_name expected type string, got int/) }
+    end
+
+    context "when message contains extra fields (typo in key)" do
+      let(:message) { { "fulll_name" => "John Doe" } }
+
+      it { expect { encode }.to raise_error(Avro::SchemaValidator::ValidationError, /extra field 'fulll_name'/) }
+    end
+  end
+
+  context 'fetching and registering schema' do
+    let(:schema_store) { AvroTurf::SchemaStore.new(path: "spec/schemas") }
+
+    let(:registry) { AvroTurf::ConfluentSchemaRegistry.new(registry_url, logger: logger) }
+
+    let(:avro) do
+      AvroTurf::Messaging.new(
+        registry: registry,
+        schema_store: schema_store,
+        logger: logger
+      )
+    end
+
+    let(:schema_id) { 234 }
+
+    context 'using fetch_schema' do
+      subject { avro.fetch_schema(subject: subj, version: version) }
+
+      let(:subj) { 'subject' }
+
+      let(:version) { 'version' }
+
+      let(:response) { {'id' => schema_id, 'schema' => schema_json} }
+
+      before do
+        allow(registry).to receive(:subject_version).with(subj, version).and_return(response)
+      end
+
+      it 'gets schema from registry' do
+        expect(subject).to eq([schema, schema_id])
+      end
+    end
+
+    context 'using fetch_schema_by_id' do
+      subject { avro.fetch_schema_by_id(schema_id) }
+
+      before do
+        allow(registry).to receive(:fetch).with(schema_id).and_return(schema_json)
+      end
+
+      it 'gets schema from registry' do
+        expect(subject).to eq([schema, schema_id])
+      end
+    end
+
+    context 'using register_schema' do
+      let(:schema_name) { 'schema_name' }
+
+      let(:namespace) { 'namespace' }
+
+      before do
+        allow(schema_store).to receive(:find).with(schema_name, namespace).and_return(schema)
+      end
+
+      context 'when subject is not set' do
+        subject { avro.register_schema(schema_name: schema_name, namespace: namespace) }
+
+        before do
+          allow(registry).to receive(:register).with(schema.fullname, schema).and_return(schema_id)
+        end
+
+        it 'registers schema in registry' do
+          expect(subject).to eq([schema, schema_id])
+        end
+      end
+
+      context 'when subject is set' do
+        subject { avro.register_schema(schema_name: schema_name, namespace: namespace, subject: subj) }
+
+        let(:subj) { 'subject' }
+
+        before do
+          allow(registry).to receive(:register).with(subj, schema).and_return(schema_id)
+        end
+
+        it 'registers schema in registry' do
+          expect(subject).to eq([schema, schema_id])
+        end
+      end
+    end
+  end
 end

data/spec/schema_store_spec.rb

@@ -198,6 +198,104 @@ describe AvroTurf::SchemaStore do
       expect(schema.fullname).to eq "person"
     end
 
+    # This test would fail under avro_turf <= v0.11.0
+    it "does NOT cache *nested* schemas in memory" do
+      FileUtils.mkdir_p("spec/schemas/test")
+
+      define_schema "test/person.avsc", <<-AVSC
+        {
+          "name": "person",
+          "namespace": "test",
+          "type": "record",
+          "fields": [
+            {
+              "name": "address",
+              "type": {
+                "name": "address",
+                "type": "record",
+                "fields": [
+                  { "name": "addr1", "type": "string" },
+                  { "name": "addr2", "type": "string" },
+                  { "name": "city", "type": "string" },
+                  { "name": "zip", "type": "string" }
+                ]
+              }
+            }
+          ]
+        }
+      AVSC
+
+      schema = store.find('person', 'test')
+      expect(schema.fullname).to eq "test.person"
+
+      expect { store.find('address', 'test') }.
+        to raise_error(AvroTurf::SchemaNotFoundError)
+    end
+
+    # This test would fail under avro_turf <= v0.11.0
+    it "allows two different avsc files to define nested sub-schemas with the same fullname" do
+      FileUtils.mkdir_p("spec/schemas/test")
+
+      define_schema "test/person.avsc", <<-AVSC
+        {
+          "name": "person",
+          "namespace": "test",
+          "type": "record",
+          "fields": [
+            {
+              "name": "location",
+              "type": {
+                "name": "location",
+                "type": "record",
+                "fields": [
+                  { "name": "city", "type": "string" },
+                  { "name": "zipcode", "type": "string" }
+                ]
+              }
+            }
+          ]
+        }
+      AVSC
+
+      define_schema "test/company.avsc", <<-AVSC
+        {
+          "name": "company",
+          "namespace": "test",
+          "type": "record",
+          "fields": [
+            {
+              "name": "headquarters",
+              "type": {
+                "name": "location",
+                "type": "record",
+                "fields": [
+                  { "name": "city", "type": "string" },
+                  { "name": "postcode", "type": "string" }
+                ]
+              }
+            }
+          ]
+        }
+      AVSC
+
+      company = nil
+      person = store.find('person', 'test')
+
+      # This should *NOT* raise the error:
+      # #<Avro::SchemaParseError: The name "test.location" is already in use.>
+      expect { company = store.find('company', 'test') }.not_to raise_error
+
+      person_location_field = person.fields_hash['location']
+      expect(person_location_field.type.name).to eq('location')
+      expect(person_location_field.type.fields_hash).to include('zipcode')
+      expect(person_location_field.type.fields_hash).not_to include('postcode')
+
+      company_headquarters_field = company.fields_hash['headquarters']
+      expect(company_headquarters_field.type.name).to eq('location')
+      expect(company_headquarters_field.type.fields_hash).to include('postcode')
+      expect(company_headquarters_field.type.fields_hash).not_to include('zipcode')
+    end
+
     it "is thread safe" do
       define_schema "address.avsc", <<-AVSC
         {

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: avro_turf
 version: !ruby/object:Gem::Version
-  version: 0.11.0
+  version: 1.2.0
 platform: ruby
 authors:
 - Daniel Schierbeck
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-10-25 00:00:00.000000000 Z
+date: 2020-08-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: avro
@@ -19,7 +19,7 @@ dependencies:
         version: 1.7.7
     - - "<"
       - !ruby/object:Gem::Version
-        version: '1.10'
+        version: '1.11'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
@@ -29,21 +29,21 @@ dependencies:
         version: 1.7.7
     - - "<"
       - !ruby/object:Gem::Version
-        version: '1.10'
+        version: '1.11'
 - !ruby/object:Gem::Dependency
   name: excon
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.45'
+        version: '0.71'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.45'
+        version: '0.71'
 - !ruby/object:Gem::Dependency
   name: bundler
   requirement: !ruby/object:Gem::Requirement
@@ -165,6 +165,7 @@ extra_rdoc_files: []
 files:
 - ".circleci/config.yml"
 - ".github/workflows/ruby.yml"
+- ".github/workflows/stale.yml"
 - ".gitignore"
 - ".rspec"
 - CHANGELOG.md