textmood 0.0.4 → 0.0.5

data/README.md

@@ -112,7 +112,10 @@ OPTIONAL options:
                                      (default 1). Note that this only makes sense if the
                                      sentiment file has tokens of similar N-gram length
 
-    -n, --normalize                  Return 1 (positive), -1 (negative) or 0 (neutral)
+    -n, --normalize-output           Return 1 (positive), -1 (negative) or 0 (neutral)
+                                     instead of the actual score. See also --min and --max.
+
+        --normalize-score            Return 1 (positive), -1 (negative) or 0 (neutral)
                                      instead of the actual score. See also --min and --max.
 
         --min-threshold FLOAT        Scores lower than this are considered negative when
@@ -149,8 +152,8 @@ and contain one colon-separated line per token, like so:
 0.875: well-situated
 0.6: well suited
 ```
-The score is to the left of the first ':', and everything to the right is the
-(potentially multi-word) token.
+The score, which must be between -1.0 and 1.0, is to the left of the first ':', 
+and everything to the right is the (potentially multi-word) token.
 
 ## Contribute
 Including baseline word/N-gram scores for many different languages is one 

data/bin/textmood

@@ -56,35 +56,44 @@ opts_parser = OptionParser.new do |opts|
   end
   opts.separator ""
   opts.separator "OPTIONAL options:"
-  opts.on("--start-ngram INTEGER", "The lowest word N-gram number to split the text into",
-                                   "(default 1). Note that this only makes sense if the",
-                                   "sentiment file has tokens of similar N-gram length") do |start_ngram|
-    options[:start_ngram] = start_ngram.to_i
+  opts.on("-o", "--normalize-output", "Return 1 (positive), -1 (negative) or 0 (neutral)",
+                                      "instead of the actual score. See also --min and --max.") do |n|
+    options[:normalize_output] = true
   end
   opts.separator ""
-  opts.on("--end-ngram INTEGER", "The highest word N-gram number to to split the text into",
-                                 "(default 1). Note that this only makes sense if the",
-                                 "sentiment file has tokens of similar N-gram length") do |end_ngram|
-    options[:end_ngram] = end_ngram.to_i
-  end
-  opts.separator ""
-  opts.on("-n", "--normalize", "Return 1 (positive), -1 (negative) or 0 (neutral)",
-                               "instead of the actual score. See also --min and --max.") do |n|
-    options[:normalize] = true
+  opts.on("-s", "--normalize-score", "Tries to normalize the score to an integer between +/- 100",
+                                     "according to the number of tokens that were scored, making",
+                                     "it more feasible to compare scores between texts of different",
+                                     "length") do |ns|
+    options[:normalize_score] = true
   end
   opts.separator ""
-  opts.on("--min-threshold FLOAT", "Scores lower than this are considered negative when",
-                                   "using --normalize (default -0.5)") do |min|
+  opts.on("-i", "--min-threshold FLOAT", "Scores lower than this are considered negative when",
+                                   "using --normalize-output (default 0.5). Note that the",
+                                   "threshold is compared to the normalized score, if applicable") do |min|
     options[:min_threshold] = min.to_f
   end
   opts.separator ""
-  opts.on("--max-threshold FLOAT", "Scores higher than this are considered positive when",
-                                   "using --normalize (default 0.5)") do |max|
+  opts.on("-x", "--max-threshold FLOAT", "Scores higher than this are considered positive when",
+                                   "using --normalize-output (default 0.5). Note that the",
+                                   "threshold is compared to the normalized score, if applicable") do |max|
     options[:max_threshold] = max.to_f
   end
   opts.separator ""
-  opts.on("-s", "--skip-symbols", "Do not include symbols file (emoticons etc.).",
-                                  "Only applies when using -l/--language.") do |s|
+  opts.on("-b", "--start-ngram INTEGER", "The lowest word N-gram number to split the text into",
+                                   "(default 1). Note that this only makes sense if the",
+                                   "sentiment file has tokens of similar N-gram length") do |start_ngram|
+    options[:start_ngram] = start_ngram.to_i
+  end
+  opts.separator ""
+  opts.on("-e", "--end-ngram INTEGER", "The highest word N-gram number to to split the text into",
+                                 "(default 1). Note that this only makes sense if the",
+                                 "sentiment file has tokens of similar N-gram length") do |end_ngram|
+    options[:end_ngram] = end_ngram.to_i
+  end
+  opts.separator ""
+  opts.on("-k", "--skip-symbols", "Do not include symbols file (emoticons etc.). Only applies",
+                            "when using -l/--language.") do |s|
     options[:include_symbols] = false
   end
   opts.separator ""

data/lang/no_NB.txt

@@ -8,21 +8,14 @@
 0.87: godt fast
 0.87: upretensiøs
 0.87: undervurdert
-0.87: top-hulls
 0.87: terapeutisk
-0.87: stirrer
-0.87: utlagte masser
 0.87: selvrespekt
-0.87: self-respektfull
-0.87: selvrespekt
-0.87: self-hensyn
 0.87: helsebringende
 0.87: rosenrød i kinnene
 0.87: utstråle
 0.87: utstråling
 0.87: velstand
 0.87: prinsipiell
-0.87: ut-og-ytre
 0.87: moraliserende
 0.87: mesterstykke
 0.87: mesterlig
@@ -64,7 +57,6 @@
 0.75: fagmessig
 0.75: snedig
 0.75: helhet
-0.75: godt gjennomtenkt av
 0.75: velproporsjonert
 0.75: godt bevart
 0.75: godt favoriserte
@@ -78,8 +70,6 @@
 0.75: infisert
 0.75: urettferdig
 0.75: kle
-0.75: revet ut av villfarelse
-0.75: riv ut av villfarelse
 0.75: pålitelighet
 0.75: øverste nivå
 0.75: brødrister
@@ -171,7 +161,6 @@
 0.75: halvhjertede
 0.75: skyldfri
 0.75: pen
-0.75: get-at-stand
 0.75: famlende
 0.75: influensa
 0.75: finere
@@ -252,7 +241,6 @@
 0.68: hedret
 0.68: treffende
 0.68: konstruktiv
-0.68: kommer-at-stand
 0.68: munter
 0.66: produktiv
 0.66: gledelig
@@ -262,11 +250,8 @@
 0.64: verdig
 0.63: god
 0.62: ettergivende
-0.62: lengtet-for
 0.62: verdig
-0.62: ord-splitting
 0.62: lurer
-0.62: ønsket-for
 0.62: klokt
 0.62: velvære
 0.62: velprøvd
@@ -292,19 +277,16 @@
 0.62: ubetenkelig
 0.62: forståelig
 0.62: kraftens
-0.62: Ullr
 0.62: sannferdig
 0.62: tillitsfullt
 0.62: klarert
 0.62: overskride
 0.62: ro
-0.62: ro
 0.62: trening
 0.62: sporbar
 0.62: sporbar
 0.62: totaliteten
 0.62: topper
-0.62: fløtekaramell-nosed
 0.62: forsagt
 0.62: tre kvarter
 0.62: vitnesbyrd
@@ -317,22 +299,18 @@
 0.62: vellykket
 0.62: stilistisk
 0.62: iherdig
-0.62: Stoppable
 0.62: vekst
 0.62: angitt
 0.62: sfærisk
 0.62: tale-endowed
-0.62: Sparer
+0.62: sparer
 0.62: ønsket
 0.62: soignée
-0.62: SOIGNE
 0.62: røykfritt
 0.62: knusende
 0.62: treghet
-0.62: klapse-up
 0.62: synde
 0.62: forenklede
-0.62: SID
 0.62: sjokksikker
 0.62: Shivaree
 0.62: velskapt
@@ -373,7 +351,6 @@
 0.62: oppladbart
 0.62: betryggende
 0.62: rimelig
-0.62: fornyet stadfestelse
 0.62: hev
 0.62: kvantifiserbare
 0.62: målrettet
@@ -411,7 +388,6 @@
 0.62: overtrekk
 0.62: overkompensere
 0.62: rangeres foran
-0.62: ut-og-ut
 0.62: ortodoks
 0.62: rikest
 0.62: overdådig
@@ -430,7 +406,6 @@
 0.62: danser
 0.62: naboskap
 0.62: naturalisering
-0.62: naturalisering
 0.62: narsissisme
 0.62: naivt
 0.62: foranderlig
@@ -674,6 +649,8 @@
 0.62: ablativ
 0.62: underdanig
 0.62: abbed
+0.62: lykke
+0.62: lykkelig
 0.58: håndgripelig
 0.58: barmhjertig
 0.58: verdighet
@@ -1046,7 +1023,7 @@
 0.50: edel-mindedness
 0.50: knuslete
 0.50: pent
-0.50: ny-laget
+0.50: nylaget
 0.50: nøytralisert
 0.50: nøytralisert
 0.50: nevrobiologiske
@@ -4250,6 +4227,7 @@
 0.20: ladet
 0.20: godartet
 0.20: våken
+0.10: lykkes
 -0.20: stamme
 -0.20: variasjon
 -0.20: varians
@@ -4515,7 +4493,6 @@
 -0.25: sukker-belagt
 -0.25: kvelende
 -0.25: kveles
--0.25: lykkes
 -0.25: underjordiske
 -0.25: underjordiske
 -0.25: innsynkning
@@ -8993,7 +8970,6 @@
 -0.75: hode
 -0.75: Hayseed
 -0.75: Haredi
--0.75: lykke
 -0.75: Hamming
 -0.75: skinke-handed
 -0.75: skinke-fisted

data/lib/textmood.rb

@@ -7,6 +7,8 @@ else
   Encoding.default_internal = Encoding::UTF_8
 end
 
+NORMALIZE_TO = 100
+
 class TextMood
 
   def initialize(options = {})
@@ -38,13 +40,21 @@ class TextMood
   def score_text(text)
     sentiment_total = 0.0
 
+    scores_added = 0
     (@options[:start_ngram]..@options[:end_ngram]).each do |i|
       ngrams(i, text.to_s).each do |token|
-        sentiment_total += score_token(token)
+        score = score_token(token)
+        unless score.nil?
+          sentiment_total += score
+          scores_added += 1
+        end
       end
     end
     
-    if @options[:normalize]
+    if @options[:normalize_score]
+      sentiment_total = normalize_score(sentiment_total, scores_added)
+    end
+    if @options[:normalize_output]
       if sentiment_total > @options[:max_threshold]
         1
       elsif sentiment_total < @options[:min_threshold]
@@ -76,7 +86,7 @@ class TextMood
       sentiment_value
     else
       puts "#{used_token}: nil" if @options[:debug]
-      0.0
+      nil
     end
   end
 
@@ -104,4 +114,9 @@ class TextMood
     sentiment_values
   end
 
+  def normalize_score(score, count)
+    factor = NORMALIZE_TO / count
+    (score * factor).to_i
+  end
+
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: textmood
 version: !ruby/object:Gem::Version
-  version: 0.0.4
+  version: 0.0.5
   prerelease: 
 platform: ruby
 authors: