Implemented the word-based dialect detection

src/poalses/jonathan/dialect/dialect_nlp.clj

@@ -47,11 +47,11 @@
 
 ;; Word sets that will show a sentence as being of that dialect
 
-(def australian-words #{})
+(def australian-words #{"incorrect" "why"})
 
-(def scottish-words #{})
+(def scottish-words #{"hence"})
 
-(def american-words #{})
+(def american-words #{"like"})
 
 ;; Predicate sets to check a sentence and see if it grammatically matches a dialect
 
@@ -65,7 +65,29 @@
 ;; Take a sentence and figure out its dialect
 
 (defn detect-sentence-dialect [sentence]
-  (if (some bad-words (dl/text (dl/tokens sentence))) :bad :good))
+  (let [tokens (dl/tokens sentence)
+        dialects1 (when (some australian-words (dl/text tokens)) [:australian])
+        dialects2 (when (some scottish-words (dl/text tokens)) [:scottish])
+        dialects3 (when (some american-words (dl/text tokens)) [:american])
+        dialects (remove nil? (flatten (conj dialects1 dialects2 dialects3)))]
+    (if (empty? dialects) [:standard] dialects)))
+
+
+
+;; Another failed attempt
+;(defn detect-sentence-dialect [sentence]
+;  (let [dialects []
+;        tokens   (dl/tokens sentence)]
+;    (when (some australian-words (dl/text (dl/tokens tokens)))
+;      (let [dialects (conj dialects :australian)]
+;        (when (some scottish-words (dl/text (dl/tokens tokens)))
+;          (let [dialects (conj dialects :scottish)]
+;            (when (some american-words (dl/text (dl/tokens tokens)))
+;              (let [ dialects (conj dialects :american)]
+;                (if (empty? dialects) (conj dialects :standard))
+;                dialects))))))))
+
+
 
 ;; Take a text sample and separate it into its sentences, then for each sentence find its dialects, and return the most common dialect
 ;; A sentence can have an indeterminate number of dialects associated with it, as detect-sentence-dialects can return a collection,
@@ -149,6 +171,15 @@
 
   (def rats (datafy (dl/dependency-graph (nth sentences-one 1))))
 
+
+  (first (last (sort-by val (frequencies (flatten (map detect-sentence-dialect (dl/sentences (nlp test-sentence-one))))))))
+  (first (last (sort-by val (frequencies (flatten (map detect-sentence-dialect (dl/sentences (nlp test-sentence-two))))))))
+  (first (last (sort-by val (frequencies (flatten (map detect-sentence-dialect (dl/sentences (nlp test-sentence-three))))))))
+  (first (last (sort-by val (frequencies (flatten (map detect-sentence-dialect (dl/sentences (nlp test-sentence-four))))))))
+  (first (last (sort-by val (frequencies (flatten (map detect-sentence-dialect (dl/sentences (nlp test-sentence-five))))))))
+  (first (last (sort-by val (frequencies (flatten (map detect-sentence-dialect (dl/sentences (nlp test-sentence-six))))))))
+
+
   (last (vals rats))
   (.getTarget (first (last (vals rats))))
   (bean (first (last (vals rats))))