aranha-parsers 0.21.0 → 0.22.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6020b51d20ef2c677664058978015a69c4ed80c15c7c169f300acaa5b5335a22
-  data.tar.gz: 44bfaaa2b28b7cacd935074b9f48bfaf9cf62e9d952f2744e8bb76fe4045a72f
+  metadata.gz: d2ddf8d96b3eb46a3690802194d7bac810d77955dc2f79e6c0329c80ef326592
+  data.tar.gz: 7caeceb58462533b58c9976ed6ab942f995a8756d4db00ff50f61fd730f98735
 SHA512:
-  metadata.gz: f1a926d3f7b051f17ee9df3c4642fcf9cf7707c9c9c70600d9a58be63bce9f80fdc35e589c41899365a854e0c506ddcf038cb96776761c2fb0916f7acd8e84da
-  data.tar.gz: c27820771a1fd288597218fc7af862c43a4eed9eb3d9c87454cc4da84814b8e270782507d16a697bd42ffc85c859b1a672978fbe02401ec4978fb81eb22cbb78
+  metadata.gz: 394353da9e1e5859005fcb65104629535162e9a89d2acf70c515f71358e2fca7b1291b6210549f4f1f9d64577aaef6a947670f8ae1903af4f16af1ade346c4ca
+  data.tar.gz: 0bba5dff7e92e219a978f4708ccaa083a208671c0a810ad204da7397760e633645228aa2ddb718d33e508ca76ca8cb3da6f7c3f1b043a5d6d241c6d3cf2b8e37

data/lib/aranha/parsers/base.rb

@@ -54,7 +54,7 @@ module Aranha
       def log_content(content, suffix = '')
         path = log_file(suffix)
 
-        File.open(path, 'wb') { |file| file.write(content) } if path
+        File.binwrite(path, content) if path
 
         content
       end
@@ -70,7 +70,7 @@ module Aranha
 
       def log_parsers_dir
         return ENV[LOG_DIR_ENVVAR] if ENV[LOG_DIR_ENVVAR]
-        return ::Rails.root.join('log', 'parsers') if rails_root_exist?
+        return ::Rails.root.join('log/parsers') if rails_root_exist?
 
         nil
       end

data/lib/aranha/parsers/html/node/base.rb

@@ -12,11 +12,11 @@ module Aranha
           end
 
           def parse(node)
-            fields.map do |f|
+            fields.to_h do |f|
               [f[0], parse_field(node, f[2], f[1])]
             rescue StandardError => e
               raise StandardError, "#{e.message}\nFields: #{f}"
-            end.to_h
+            end
           end
 
           private

data/lib/aranha/parsers/html/node/default/numeric_support.rb

@@ -102,7 +102,7 @@ module Aranha
                 m[0].delete(delimiter).to_f
               elsif required
                 raise "decimal [Separator=\"#{separator}, Delimiter=\"#{delimiter}\"] value not " \
-                  "found in \"#{s}\""
+                      "found in \"#{s}\""
               end
             end
 

data/lib/aranha/parsers/html/node/default/string_support.rb

@@ -13,7 +13,7 @@ module Aranha
               s = string_value(node, xpath)
               return '' unless s
 
-              m = /\"([^\"]+)\"/.match(s)
+              m = /"([^\"]+)"/.match(s)
               return m[1] if m
 
               ''
@@ -35,7 +35,7 @@ module Aranha
               found ? sanitize_string(found.text) : ''
             end
 
-            def string_recursive_value(node, xpath, required = true)
+            def string_recursive_value(node, xpath, required = true) # rubocop:disable Style/OptionalBooleanParameter
               root = node_value(node, xpath)
               if root.blank?
                 return nil unless required
@@ -65,7 +65,7 @@ module Aranha
               s = ''
               node.children.each do |child|
                 child_s = string_recursive(child)
-                s += ' ' + child_s if child_s.present?
+                s += " #{child_s}" if child_s.present?
               end
               sanitize_string(s)
             end

data/lib/aranha/parsers/patches/ofx_parser.rb

@@ -34,5 +34,5 @@ end
 
 require 'eac_ruby_utils/patch'
 
-::EacRubyUtils.patch(::OfxParser::OfxParser, ::Aranha::Parsers::Patches::OfxParser::OfxParser)
-::EacRubyUtils.patch(::OfxParser::Transaction, ::Aranha::Parsers::Patches::OfxParser::Transaction)
+EacRubyUtils.patch(OfxParser::OfxParser, Aranha::Parsers::Patches::OfxParser::OfxParser)
+EacRubyUtils.patch(OfxParser::Transaction, Aranha::Parsers::Patches::OfxParser::Transaction)

data/lib/aranha/parsers/source_address/file.rb

@@ -10,7 +10,7 @@ module Aranha
 
         class << self
           def valid_source?(source)
-            source.to_s.start_with?(SCHEME + '/', '/')
+            source.to_s.start_with?("#{SCHEME}/", '/')
           end
         end
 
@@ -23,7 +23,7 @@ module Aranha
         end
 
         def content
-          ::File.open(source, &:read)
+          ::File.read(source)
         end
       end
     end

data/lib/aranha/parsers/source_address/hash_http_base/headers.rb

@@ -0,0 +1,43 @@
+# frozen_string_literal: true
+
+require 'eac_ruby_utils'
+
+module Aranha
+  module Parsers
+    class SourceAddress
+      class HashHttpBase
+        class Headers
+          acts_as_immutable
+          immutable_accessor :value, type: :hash
+
+          # @!method to_h()
+          # @return [Hash]
+          delegate :to_h, to: :values
+
+          # @param other_values [Array, Hash]
+          # @return [Aranha::Parsers::SourceAddress::HashHttpBase::Headers]
+          def merge(other_values)
+            if other_values.is_a?(::Hash)
+              values(values.merge(other_values))
+            elsif other_values.is_a?(::Enumerable)
+              merge(other_values.to_h { |v| merge_array_item_to_h(v) })
+            else
+              raise ::ArgumentError,
+                    "\"other_values\"=\"#{other_values}\" should be a Array or a Hash"
+            end
+          end
+
+          private
+
+          # @param item [Array]
+          # @return [Array
+          def merge_array_item_to_h(item)
+            2.times.inject(Array(item)) do |a, e|
+              e >= a.count ? a + [nil] : a
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/aranha/parsers/source_address/hash_http_base.rb

@@ -26,6 +26,8 @@ module Aranha
         DEFAULT_FOLLOW_REDIRECT = true
         DEFAULT_HEADERS = {}.freeze
         DEFAULT_PARAMS = {}.freeze
+        USER_AGENT_KEY = 'user-agent'
+        USER_AGENT_VALUE = 'aranha-parsers'
 
         enable_simple_cache
 
@@ -73,12 +75,26 @@ module Aranha
 
         # @return [EacEnvs::Http::Request]
         def http_request
-          r = ::EacEnvs::Http::Request.new.verb(self.class.http_method).url(url)
-          r = headers.if_present(r) { |v| r.headers(v) }
+          r = initial_http_request
+          r = headers.if_present(r) { |v| r.headers(initial_headers.merge(v).to_h) }
           r = body.if_present(r) { |v| r.body_data(v) }
           r = r.follow_redirect(true) if follow_redirect?
           r
         end
+
+        # @return [Hash]
+        def initial_headers
+          ::Aranha::Parsers::SourceAddress::HashHttpBase::Headers.new
+            .value(USER_AGENT_KEY, USER_AGENT_VALUE)
+        end
+
+        # @return [EacEnvs::Http::Request]
+        def initial_http_request
+          ::EacEnvs::Http::Request.new.verb(self.class.http_method).url(url)
+            .headers(initial_headers.to_h)
+        end
+
+        require_sub __FILE__
       end
     end
   end

data/lib/aranha/parsers/source_address/http_get.rb

@@ -40,6 +40,7 @@ module Aranha
 
         def content
           request = ::EacEnvs::Http::Request.new.url(url).retry(true).follow_redirect(true)
+                      .header('user-agent', self.class.name)
           request.response.body_str
         rescue ::EacEnvs::Http::Error => e
           raise ::Aranha::Parsers::SourceAddress::FetchContentError, e.message, request

data/lib/aranha/parsers/source_address.rb

@@ -44,7 +44,7 @@ module Aranha
       end
 
       def serialize
-        sub.serialize.strip + "\n"
+        "#{sub.serialize.strip}\n"
       end
 
       private

data/lib/aranha/parsers/version.rb

@@ -2,6 +2,6 @@
 
 module Aranha
   module Parsers
-    VERSION = '0.21.0'
+    VERSION = '0.22.0'
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: aranha-parsers
 version: !ruby/object:Gem::Version
-  version: 0.21.0
+  version: 0.22.0
 platform: ruby
 authors:
 - Esquilo Azul Company
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2023-06-21 00:00:00.000000000 Z
+date: 2023-11-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport
@@ -33,7 +33,7 @@ dependencies:
         version: '2.8'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 2.8.4
+        version: 2.8.5
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
@@ -43,7 +43,7 @@ dependencies:
         version: '2.8'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 2.8.4
+        version: 2.8.5
 - !ruby/object:Gem::Dependency
   name: eac_envs-http
   requirement: !ruby/object:Gem::Requirement
@@ -51,6 +51,9 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '0.4'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 0.4.1
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
@@ -58,26 +61,29 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '0.4'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 0.4.1
 - !ruby/object:Gem::Dependency
   name: eac_ruby_utils
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.117'
+        version: '0.119'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.117.1
+        version: 0.119.2
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.117'
+        version: '0.119'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.117.1
+        version: 0.119.2
 - !ruby/object:Gem::Dependency
   name: nokogiri
   requirement: !ruby/object:Gem::Requirement
@@ -87,7 +93,7 @@ dependencies:
         version: '1.15'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.15.2
+        version: 1.15.4
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
@@ -97,7 +103,7 @@ dependencies:
         version: '1.15'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.15.2
+        version: 1.15.4
 - !ruby/object:Gem::Dependency
   name: ofx-parser
   requirement: !ruby/object:Gem::Requirement
@@ -118,14 +124,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.5.1
+        version: '0.9'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.5.1
+        version: '0.9'
 description: 
 email: 
 executables: []
@@ -150,19 +156,14 @@ files:
 - lib/aranha/parsers/json/base.rb
 - lib/aranha/parsers/patches.rb
 - lib/aranha/parsers/patches/ofx_parser.rb
-- lib/aranha/parsers/rspec.rb
-- lib/aranha/parsers/rspec/setup.rb
-- lib/aranha/parsers/rspec/shared_examples/source_target_fixtures.rb
-- lib/aranha/parsers/rspec/source_target_fixtures_controller.rb
 - lib/aranha/parsers/source_address.rb
 - lib/aranha/parsers/source_address/fetch_content_error.rb
 - lib/aranha/parsers/source_address/file.rb
 - lib/aranha/parsers/source_address/hash_http_base.rb
+- lib/aranha/parsers/source_address/hash_http_base/headers.rb
 - lib/aranha/parsers/source_address/hash_http_get.rb
 - lib/aranha/parsers/source_address/hash_http_post.rb
 - lib/aranha/parsers/source_address/http_get.rb
-- lib/aranha/parsers/source_target_fixtures.rb
-- lib/aranha/parsers/source_target_fixtures/source_target_file.rb
 - lib/aranha/parsers/version.rb
 homepage: 
 licenses: []
@@ -175,7 +176,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '0'
+      version: '2.7'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="

data/lib/aranha/parsers/rspec/setup.rb

@@ -1,15 +0,0 @@
-# frozen_string_literal: true
-
-require 'eac_ruby_utils/core_ext'
-
-module Aranha
-  module Parsers
-    module Rspec
-      module Setup
-        def self.extended(_setup_obj)
-          require 'aranha/parsers/rspec/shared_examples/source_target_fixtures'
-        end
-      end
-    end
-  end
-end

data/lib/aranha/parsers/rspec/shared_examples/source_target_fixtures.rb

@@ -1,74 +0,0 @@
-# frozen_string_literal: true
-
-require 'aranha/parsers/rspec/source_target_fixtures_controller'
-require 'aranha/parsers/source_target_fixtures'
-require 'yaml'
-
-RSpec.shared_examples 'source_target_fixtures' do |spec_file| # rubocop:disable Metrics/BlockLength
-  fixtures_controller = ::Aranha::Parsers::Rspec::SourceTargetFixturesController
-                          .new(self, spec_file)
-
-  let(:fixtures_controller) { fixtures_controller }
-  let(:spec_file) { spec_file }
-
-  it 'fixtures directory should exist' do
-    expect(::File.directory?(fixtures_controller.fixtures_dir)).to be true
-  end
-
-  context 'with fixtures directory' do
-    it 'has at least one file' do
-      expect(source_target_fixtures.source_target_files.count).to be > 0 # rubocop:disable Style/NumericPredicate
-    end
-
-    fixtures_controller.source_target_fixtures.source_target_files.each do |st|
-      context "when source file is \"#{::File.basename(st.source)}\"" do
-        if fixtures_controller.write_target_fixtures?
-          it 'writes target data' do
-            sd = sort_results(source_data(st.source))
-            basename = ::Aranha::Parsers::SourceTargetFixtures.source_target_basename(st.source)
-            target_file = File.expand_path("../#{basename}.target#{target_file_extname}", st.source)
-            File.write(target_file, target_content(sd))
-          end
-        else
-          it 'parses data' do
-            assert_source_target_complete(st)
-            sd = source_data(st.source)
-            td = target_data(st.target)
-            expect(sort_results(sd)).to eq(sort_results(td))
-          end
-        end
-      end
-    end
-  end
-
-  delegate :source_target_fixtures, to: :fixtures_controller
-
-  def assert_source_target_complete(source_target)
-    expect(source_target.source).to(be_truthy, "Source not found (Target: #{source_target.target})")
-    expect(source_target.target).to(be_truthy, "Target not found (Source: #{source_target.source})")
-  end
-
-  def source_data(source_file)
-    instance = described_class.new(source_file)
-    return instance.data if instance.respond_to?(:data)
-
-    raise "#{instance} has no \"data\" method. You need to implement \"#{instance}.data\" or " \
-      "\"#{self}.source_data(source_file)\""
-  end
-
-  def sort_results(results)
-    results
-  end
-
-  def target_data(target_file)
-    ::YAML.load_file(target_file)
-  end
-
-  def target_content(data)
-    data.to_yaml
-  end
-
-  def target_file_extname
-    '.yaml'
-  end
-end

data/lib/aranha/parsers/rspec/source_target_fixtures_controller.rb

@@ -1,39 +0,0 @@
-# frozen_string_literal: true
-
-require 'eac_ruby_utils/core_ext'
-
-module Aranha
-  module Parsers
-    module Rspec
-      class SourceTargetFixturesController
-        enable_simple_cache
-        common_constructor :example, :spec_file
-
-        def default_fixtures_dir
-          ::File.join(
-            ::File.dirname(spec_file),
-            ::File.basename(spec_file, '.*') + '_files'
-          )
-        end
-
-        def fixtures_dir
-          if example.respond_to?(:fixtures_dir)
-            example.fixtures_dir
-          else
-            default_fixtures_dir
-          end
-        end
-
-        def write_target_fixtures?
-          ENV['WRITE_TARGET_FIXTURES'].to_bool
-        end
-
-        private
-
-        def source_target_fixtures_uncached
-          ::Aranha::Parsers::SourceTargetFixtures.new(fixtures_dir)
-        end
-      end
-    end
-  end
-end

data/lib/aranha/parsers/rspec.rb

@@ -1,11 +0,0 @@
-# frozen_string_literal: true
-
-require 'eac_ruby_utils/core_ext'
-
-module Aranha
-  module Parsers
-    module Rspec
-      require_sub __FILE__
-    end
-  end
-end

data/lib/aranha/parsers/source_target_fixtures/source_target_file.rb

@@ -1,19 +0,0 @@
-# frozen_string_literal: true
-
-module Aranha
-  module Parsers
-    class SourceTargetFixtures
-      class SourceTargetFile
-        common_constructor :owner, :basename
-
-        def source
-          owner.source_file(basename)
-        end
-
-        def target
-          owner.target_file(basename)
-        end
-      end
-    end
-  end
-end

data/lib/aranha/parsers/source_target_fixtures.rb

@@ -1,78 +0,0 @@
-# frozen_string_literal: true
-
-require 'active_support/core_ext/object'
-require 'active_support/core_ext/string'
-
-module Aranha
-  module Parsers
-    # Lists pairs of source/target files in a directory.
-    class SourceTargetFixtures
-      require_sub __FILE__
-
-      class << self
-        def source_target_basename(file)
-          m = /^(.+)\.(?:source|target)(?:\..+)?$/.match(File.basename(file))
-          m ? m[1] : nil
-        end
-      end
-
-      attr_reader :fixtures_directory
-
-      def initialize(fixtures_directory)
-        @fixtures_directory = fixtures_directory
-      end
-
-      def source_target_files
-        sources_targets_basenames.map do |basename|
-          ::Aranha::Parsers::SourceTargetFixtures::SourceTargetFile.new(self, basename)
-        end
-      end
-
-      def source_files
-        r = []
-        source_target_files.each do |st|
-          r << st.source if st.source
-        end
-        r
-      end
-
-      def target_files
-        r = []
-        source_target_files.each do |st|
-          r << st.target if st.target
-        end
-        r
-      end
-
-      def target_file(basename)
-        fixture_file(basename, 'target')
-      end
-
-      def source_file(basename)
-        fixture_file(basename, 'source')
-      end
-
-      private
-
-      def fixture_file(basename, suffix)
-        prefix = "#{basename}.#{suffix}"
-        Dir.foreach(fixtures_directory) do |item|
-          next if %w[. ..].include?(item)
-          return File.expand_path(item, fixtures_directory) if item.starts_with?(prefix)
-        end
-        nil
-      end
-
-      def sources_targets_basenames
-        basenames = Set.new
-        Dir.foreach(fixtures_directory) do |item|
-          next if %w[. ..].include?(item)
-
-          b = self.class.source_target_basename(item)
-          basenames << b if b.present?
-        end
-        basenames
-      end
-    end
-  end
-end