aranha-parsers 0.8.5 → 0.9.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9fd66650df384bfd84cc234f0ed901e426f0000a7f11bd8cdcec959d5690efdb
-  data.tar.gz: a0b80baee8f763633753c9fefacd5a0ab528a8047a8f14a876a08f0cd443c6a1
+  metadata.gz: cb3631929ccd8b029ab5b097554f96f5eb0cc27dcf8767d3a8a97f050ee60b80
+  data.tar.gz: f30f306452ac72997887550051954d290f7e6c321b694236fd9f49bc6fff5d4a
 SHA512:
-  metadata.gz: f9bd1e8d032eb4f782526e45b549aecdf6951a9c28ba55296293128d4a655a14f8cafda545560d794c5f39a4798c65b96c39e2e5dcea324057a85c5dec492d01
-  data.tar.gz: 577bd16ed893fa9df121de6be1844aefc1f178790871aa89caa9073633e83682059f326de48d8561a7928857d26c08f57bbc655d0205b2a7ef7126dd922e2398
+  metadata.gz: 35ebd10ceebadd29ade03ef0c5a75ed024e8d4377b19295861d8ef4c4e3bb4e091df5ff9f74398f523c98e0e3ee65eac1352bedc39eae033000cdc8561c5b749
+  data.tar.gz: 68a7664bef2ffd163cf110f14779feb92305b90cfb6b23e8d254c4cf2ca6871233d64032704790dd30397383cc67fd5a5b5dc9e67a1d866f1365524707b5ebb5

data/lib/aranha/parsers/base.rb

@@ -11,7 +11,9 @@ module Aranha
       class << self
         def from_content(content)
           ::EacRubyUtils::Fs::Temp.on_file do |path|
-            path.write(content)
+            ::File.open(path.to_s, 'w:UTF-8') do |f|
+              f.write content.force_encoding('UTF-8')
+            end
             r = new(path.to_path)
             r.content
             r

data/lib/aranha/parsers/json/base.rb

@@ -0,0 +1,20 @@
+# frozen_string_literal: true
+
+require 'aranha/parsers/base'
+require 'json'
+
+module Aranha
+  module Parsers
+    module Json
+      class Base < ::Aranha::Parsers::Base
+        def data
+          default_data
+        end
+
+        def default_data
+          ::JSON.parse(content)
+        end
+      end
+    end
+  end
+end

data/lib/aranha/parsers/json.rb

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

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

@@ -32,6 +32,11 @@ module Aranha
           source
         end
 
+        def final_url
+          content unless @final_url
+          @final_url
+        end
+
         def content
           c = ::Curl::Easy.new(url)
           c.follow_location = true
@@ -53,6 +58,7 @@ module Aranha
             raise(::Aranha::Parsers::SourceAddress::FetchContentError,
                   "Curl perform failed (URL: #{url})")
           end
+          @final_url = curl.url
         rescue Curl::Err::CurlError => e
           raise ::Aranha::Parsers::SourceAddress::FetchContentError, "CURL error: #{e.class.name}"
         end

data/lib/aranha/parsers/version.rb

@@ -2,6 +2,6 @@
 
 module Aranha
   module Parsers
-    VERSION = '0.8.5'
+    VERSION = '0.9.0'
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: aranha-parsers
 version: !ruby/object:Gem::Version
-  version: 0.8.5
+  version: 0.9.0
 platform: ruby
 authors:
 - Esquilo Azul Company
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-09-21 00:00:00.000000000 Z
+date: 2021-10-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport
@@ -157,6 +157,8 @@ files:
 - lib/aranha/parsers/html/node/base.rb
 - lib/aranha/parsers/html/node/default.rb
 - lib/aranha/parsers/invalid_state_exception.rb
+- lib/aranha/parsers/json.rb
+- lib/aranha/parsers/json/base.rb
 - lib/aranha/parsers/patches.rb
 - lib/aranha/parsers/patches/ofx_parser.rb
 - lib/aranha/parsers/rspec.rb