opener-chained-daemon 3.0.5 → 3.1.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 3d0b1b563afb30110fd2592c79c196622df4388aa370b98356e951cef818db73
-  data.tar.gz: bd223a5042e3ccf8fa04b736cb6d8403f20173f45d7bbd306a96fc254c5a77a8
+  metadata.gz: 2f35c2981bde4e7e2b83deb6b9a412cdd5aa61ccc6c3c89bc454d728a7c54ee9
+  data.tar.gz: e12cf2b8dd8df0483fa1bb304cf32b6185aa110bf468bcdb991cd3053aef48ed
 SHA512:
-  metadata.gz: 0ab381cdb194d8d189478f775749d2e18c9ab62105ae731a86217f588d5e7164569845ea7bb919b7d5f7761209b5e53ae47959a6facc2706bee461f96081b00a
-  data.tar.gz: 9ca55c537ce234fc74748d116a328f82da37fb7815ac6cb6bf0af9172e0839eb677e7fe30e84370fc8230e472dd5cc5d211e41899e3ba63a5acd0e1191b8efc3
+  metadata.gz: 56185c5656db86a96b1171d932e4cd746d3ce4e4ec9022647889076e07ddddac580cb5770d9bcb676d2d140182bf47a99d08ffd55db7cd15192752205e567a3e
+  data.tar.gz: 7c13d54407fb02a78b96e6495cddea4a338798cf1a86066623c831ba562a6e785f8a9eb4fec764e263508b451e9b1469de4bb58526d35f385afe61b45b7d4bc5

data/lib/opener/chained_daemon.rb

@@ -3,6 +3,9 @@ require 'oga'
 require 'hashie'
 require 'google/cloud/translate'
 
+require 'rexml/document'
+require 'rexml/formatters/pretty'
+
 require 'opener/daemons'
 
 require_relative 'chained_daemon/languages_cache'

data/lib/opener/chained_daemon/chained_daemon.rb

@@ -47,6 +47,7 @@ module Opener
         output = xml.to_s
       end
 
+      output = pretty_print output if params[:cache_keys][:environment] == 'staging'
       output
 
     rescue Core::UnsupportedLanguageError
@@ -54,6 +55,17 @@ module Opener
       output
     end
 
+    def pretty_print(output)
+      doc = REXML::Document.new output
+      doc.context[:attribute_quote] = :quote
+      out = ""
+      formatter = REXML::Formatters::Pretty.new
+      formatter.compact = true
+      formatter.write(doc, out)
+
+      out.strip
+    end
+
     def translate xml, params
       raw = xml.at :raw
       case translate_service params

data/lib/opener/chained_daemon/cli.rb

@@ -47,7 +47,14 @@ Example:
           run do |opts, args|
             daemon = ChainedDaemon.new args: args
             input  = STDIN.tty? ? nil : STDIN.read
-            params = JSON.parse ENV['PARAMS'] if ENV['PARAMS']
+
+            params = if ENV['PARAMS']
+                       JSON.parse ENV['PARAMS']
+                     else
+                       {}
+                     end
+            # Set environment as staging from console for testing purposes
+            params[:cache_keys] = { environment: 'staging', merged: true }
 
             puts daemon.run input, params || {}
           end

data/lib/opener/chained_daemon/languages_cache.rb

@@ -23,7 +23,7 @@ module Opener
       end
 
       def cache_update
-        puts "loading supported languages from url #{@url}"
+        puts "loading supported languages from url #{@url}" if ENV['DEBUG']
 
         languages     = JSON.parse http.get(@url).body
         @cache        = languages['data'].map { |l| l['code'] }

data/lib/opener/chained_daemon/version.rb

@@ -1,7 +1,7 @@
 module Opener
   class ChainedDaemon
 
-    VERSION = '3.0.5'
+    VERSION = '3.1.2'
 
   end
 end

data/lib/opener/stanza/tokenizer_pos.rb

@@ -8,6 +8,9 @@ module Opener
       BASE_URL        = ENV['STANZA_SERVER']
       LANGUAGES_CACHE = Opener::ChainedDaemon::LanguagesCache.new
 
+      RTL_LANGUAGES   = [ "ar", "ara", "arc", "ae", "ave", "egy", "he", "heb", "nqo", "pal", "phn", "sam",
+                          "syc", "syr", "fa", "per", "fas", "ku", "kur", "ur", "urd" ]
+
       POS             = {
         'DET'   => 'D',
         'ADJ'   => 'G',
@@ -35,7 +38,8 @@ module Opener
 
         kaf      = KAF::Document.from_xml input
 
-        unless LANGUAGES_CACHE.get.include? kaf.language
+        prod     = params[:cache_keys][:environment] != 'staging'
+        if prod and !LANGUAGES_CACHE.get.include?(kaf.language)
           raise Core::UnsupportedLanguageError.new kaf.language
         end
 
@@ -45,16 +49,27 @@ module Opener
         tokens   = JSON.parse response.body
 
         w_index = 0
+
+        tokens.map{ |t| t.reverse! } if RTL_LANGUAGES.include? kaf.language
         tokens.each_with_index do |sentence, s_index|
+          miscs = {}
           sentence.each_with_index do |word|
             w_index += 1
-            misc   = word['misc']
+            # save misc for later usase in a MWT case
+            if word['id'].is_a? Array
+              word['id'].each { |id| miscs[id] = word['misc'] }
+              next
+            end
+            misc = word['misc'] || miscs[word['id']]
+
             offset = misc.match(/start_char=(\d+)|/)[1].to_i
             length = misc.match(/end_char=(\d+)/)[1].to_i - offset
 
             u_pos  = word['upos']
             pos    = POS[u_pos]
-            raise "Didn't find a map for #{u_pos}" if pos.nil?
+            Rollbar.scoped({ input: input, params: params, tokens: tokens }) do
+              raise "Didn't find a map for #{u_pos}"
+            end if pos.nil?
             type   = if POS_OPEN.include? pos then 'open' else 'close' end
 
             params = {

data/opener-chained-daemon.gemspec

@@ -29,14 +29,14 @@ Gem::Specification.new do |spec|
   spec.add_dependency 'roda'
   spec.add_dependency 'rack-timeout'
   spec.add_dependency 'faraday'
-  spec.add_dependency 'opener-daemons', '~> 2.7.1'
+  spec.add_dependency 'opener-daemons', '~> 2.7.2'
   spec.add_dependency 'opener-callback-handler', '~> 1.0'
 
   spec.add_dependency 'opener-language-identifier', '>= 4.4.0'
   spec.add_dependency 'opener-tokenizer',  '>= 2.2.0'
   spec.add_dependency 'opener-pos-tagger', '>= 3.2.0'
   spec.add_dependency 'opener-polarity-tagger', '>= 3.2.7'
-  spec.add_dependency 'opener-property-tagger', '>= 3.3.4'
+  spec.add_dependency 'opener-property-tagger', '>= 3.3.5'
   spec.add_dependency 'opener-opinion-detector-basic', '>= 3.2.3'
 
   spec.add_development_dependency 'bundler', '~> 1.3'

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: opener-chained-daemon
 version: !ruby/object:Gem::Version
-  version: 3.0.5
+  version: 3.1.2
 platform: ruby
 authors:
 - development@olery.com
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2020-11-09 00:00:00.000000000 Z
+date: 2020-11-16 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport
@@ -114,14 +114,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.7.1
+        version: 2.7.2
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.7.1
+        version: 2.7.2
 - !ruby/object:Gem::Dependency
   name: opener-callback-handler
   requirement: !ruby/object:Gem::Requirement
@@ -198,14 +198,14 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 3.3.4
+        version: 3.3.5
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 3.3.4
+        version: 3.3.5
 - !ruby/object:Gem::Dependency
   name: opener-opinion-detector-basic
   requirement: !ruby/object:Gem::Requirement