wcc-text-analysis 0.0.1 → 0.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 53076bbd0f7310bd02cab7548cd71edb8417ad7bd320c4828c0f80dc72cdc0ad
-  data.tar.gz: e161b846a746e14d0912898d8028f5c954f2be6eeda844a16e202744962bfeb1
+  metadata.gz: 66a18a3f32298d0124f6e045f4e83db791fd2d67502c327be659fed2ab82ffad
+  data.tar.gz: fee8ce474ebaa139fc0ea4fb8fa706cdb5646fe2b9b3eda768ed8ef23994d137
 SHA512:
-  metadata.gz: 1f07d19805a3723020d8d978c53c5c921958122d76f51a09e39b7501dca28df2239b49eacef6f2be154eacffd9bcc588ffdd7fd2c72dbd16b328e1305f556b4f
-  data.tar.gz: f685000411857bcb532f3e0bf46acfa1201ceb4bc53e470d3e011554489e676437e3048dcf798245f943bf471ec3d23b9dd057de544ba1728232194f3a5be24d
+  metadata.gz: e69ce56fe1ffec9d057e9694ff0b45cda625e894b42abca76c94c8b14992c94dbdcca4045be5452c50f496877cd2842ee9489027fde95a55d7a418057f488e63
+  data.tar.gz: 88730022fd77784a9ae04c34ff7ddc32dff8c75414065ed53e078e1471a2af980f495b5587c62753252d156ad1da12cb5f1569d4e073e798300673083ef781a2

data/.gitignore

@@ -1 +1,2 @@
-Gemfile.lock
+Gemfile.lock
+pkg/*

data/lib/wcc/text_analysis.rb

@@ -8,11 +8,22 @@ module WCC
 
     attr_reader :normalized, :stripped
 
-    def self.extract_terms(file)
-      File.read("db/#{file}.txt").split("\n")
+    def self.extract_terms(db_file)
+      File.read(
+        File.join(File.dirname(__FILE__), '../../db', db_file),
+      ).split("\n")
     end
 
-    def initialize(string, stop_words: STOPWORDS)
+    # Stopwords from http://www.ranks.nl/stopwords
+    def self.default_stopwords
+      @default_stopwords ||= extract_terms('stop_words.txt')
+    end
+
+    def self.default_exclusions
+      @default_exclusions ||= extract_terms('transcript_exclusions.txt')
+    end
+
+    def initialize(string, stop_words: self.class.default_stopwords)
       @original = string
       @stop_words = stop_words
     end
@@ -61,11 +72,7 @@ module WCC
     end
 
     def remove_ignored_tokens(string)
-      string - (@stop_words + EXCLUSIONS)
+      string - (@stop_words + self.class.default_exclusions)
     end
-
-    # Stopwords from http://www.ranks.nl/stopwords
-    STOPWORDS  = extract_terms("stop_words").freeze
-    EXCLUSIONS = extract_terms("transcript_exclusions").freeze
   end
 end

data/lib/wcc/text_analysis/version.rb

@@ -1,5 +1,5 @@
 module WCC
   class TextAnalysis
-    VERSION = '0.0.1'.freeze
+    VERSION = '0.0.2'.freeze
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: wcc-text-analysis
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.0.2
 platform: ruby
 authors:
 - Watermark Dev