Added examples of different usages of the word "like".

src/poalses/jonathan/dialect/dialect_nlp.clj

@@ -19,6 +19,15 @@
 (def test-sentence-three
   "This is an example of incorrect language usage. I had an idea about using garlic butter with chicken, hence why the food was burnt.")
 
+(def test-sentence-four
+  "This is an example of incorrect language usage. I had an idea about using garlic butter with chicken, hence why the food was burnt. This one has an extra sentence to tip the balance to good.")
+
+(def test-sentence-five
+  "This is an example of Californian slang. So, like, I had this great idea, where we put like, garlic butter with chicken, it's so rad.")
+
+(def test-sentence-six
+  "This is all about like. I like lychees, they're delicious. Sometimes life is like watching paint dry, basically very boring")
+
 (def nlp
   (dl/->pipeline {:annotators ["truecase"
                                "quote"
@@ -42,10 +51,7 @@
 ;; Take a text sample and separate it into it's sentences, then for each sentence find its dialect, and return the most common dialect
 
 (defn detect-sample-dialect [sample]
-  (first (last (sort-by val (frequencies (map detect-sentence-dialect sample))))))
-
-
-
+  (first (last (sort-by val (frequencies (map detect-sentence-dialect (dl/sentences (nlp sample))))))))
 
 (def annotated-example
   (delay (nlp example)))
@@ -65,11 +71,29 @@
 (def sentences-three
   (dl/sentences @annotated-example-three))
 
+(def annotated-example-four
+  (delay (nlp test-sentence-four)))
+
+(def sentences-four
+  (dl/sentences @annotated-example-four))
+
+(def annotated-example-five
+  (delay (nlp test-sentence-five)))
+
+(def sentences-five
+  (dl/sentences @annotated-example-five))
+
+(def annotated-example-six
+  (delay (nlp test-sentence-six)))
+
+(def sentences-six
+  (dl/sentences @annotated-example-six))
+
 (def annotated-example-two
   (delay (nlp test-sentence-two)))
 
 (def sentences-two
-  (delay (dl/sentences @annotated-example-two)))
+  (dl/sentences @annotated-example-two))
 
 (defn show-dependencies []
   (clojure.pprint/pprint (dl/dependency-graph (nth sentences-one 1))))
@@ -102,6 +126,49 @@
 
   (def rats (datafy (dl/dependency-graph (nth sentences-one 1))))
 
+  (last (vals rats))
+  (.getTarget (first (last (vals rats))))
+  (bean (first (last (vals rats))))
+  (bean (:relation (bean (first (last (vals rats))))))
+
+  (datafy (dl/dependency-graph (nth sentences-one 1)))
+  (datafy (dl/dependency-graph (nth sentences-five 1)))
+  (datafy (dl/dependency-graph (nth sentences-six 0)))
+  (datafy (dl/dependency-graph (nth sentences-six 1)))
+  (datafy (dl/dependency-graph (nth sentences-six 2)))
+
+  (bean (:relation (bean (nth (nth (vals (datafy (dl/dependency-graph (nth sentences-five 1)))) 3) 2))))
+  (.tag (:target (bean (nth (nth (vals (datafy (dl/dependency-graph (nth sentences-five 1)))) 3) 2))))
+  (.word (:target (bean (nth (nth (vals (datafy (dl/dependency-graph (nth sentences-five 1)))) 3) 2))))
+
+
+  (.word (nth (keys (datafy (dl/dependency-graph (nth sentences-one 1)))) 9))
+  (.after (nth (keys (datafy (dl/dependency-graph (nth sentences-one 1)))) 9))
+  (.word (nth (keys (datafy (dl/dependency-graph (nth sentences-one 1)))) 10))
+  (.word (nth (keys (datafy (dl/dependency-graph (nth sentences-one 1)))) 7))
+  (.after (nth (keys (datafy (dl/dependency-graph (nth sentences-one 1)))) 7))
+
+  (.word (nth (keys (datafy (dl/dependency-graph (nth sentences-two 1)))) 9))
+  (.after (nth (keys (datafy (dl/dependency-graph (nth sentences-two 1)))) 9))
+  (.word (nth (keys (datafy (dl/dependency-graph (nth sentences-two 1)))) 10))
+  (.tag (nth (keys (datafy (dl/dependency-graph (nth sentences-two 1)))) 10))
+  (.tag (nth (keys (datafy (dl/dependency-graph (nth sentences-two 1)))) 9))
+  (.tag (nth (keys (datafy (dl/dependency-graph (nth sentences-two 1)))) 10))
+  (.tag (nth (keys (datafy (dl/dependency-graph (nth sentences-two 1)))) 9))
+
+
+  ;; So two main ways, word use and grammar. Word use is simple enough,
+  ;;just check if a word in the sentence matches the expected word
+
+  ;; For example, this returns true if the word "why" is in the given sentence
+  (some #(= "why" %) (dl/text (dl/tokens sentence)))
+
+  ;; Then we have grammar, which is a lot harder to check. For instance, a sentence using "like"
+  ;;as an interjection should return true when passed to the following code.
+
+
+  (some #(when (even? %) %) '(1 2 3 4))
+
   (def something [5647 5858 76 938 62626])
 
   (reduce + something)
@@ -122,7 +189,13 @@
   (detect-sample-dialect sentences-one)
   (detect-sample-dialect @sentences-two)
   (detect-sample-dialect sentences-three)
+  (detect-sample-dialect sentences-four)
+
 
+  (detect-sample-dialect test-sentence-one)
+  (detect-sample-dialect test-sentence-two)
+  (detect-sample-dialect test-sentence-three)
+  (detect-sample-dialect test-sentence-four)
 
   (first (last (sort-by val (frequencies (map detect-sentence-dialect sentences-one)))))
   (first (last (sort-by val (frequencies (map detect-sentence-dialect @sentences-two)))))