changed the word predicate to be functions instead of a var containing an...

changed the word predicate to be functions instead of a var containing an anonymous function, started main function

src/poalses/jonathan/dialect/dialect_nlp.clj

@@ -53,9 +53,17 @@
 
 (def american-words #{"like"})
 
-(def australian-word-predicate #(some australian-words (dl/text (dl/tokens %))))
-(def scottish-word-predicate #(some scottish-words (dl/text (dl/tokens %))))
-(def american-word-predicate #(some american-words (dl/text (dl/tokens %))))
+(defn australian-word-predicate
+  [sentence]
+  (some australian-words (dl/text (dl/tokens sentence))))
+
+(defn scottish-word-predicate
+  [sentence]
+  (some scottish-words (dl/text (dl/tokens sentence))))
+
+(defn american-word-predicate
+  [sentence]
+  (some american-words (dl/text (dl/tokens sentence))))
 
 (defn fake-test
   [fake]
@@ -63,11 +71,14 @@
 
 ;; Predicate vectors to check a sentence and see if it grammatically matches a dialect
 
-(def australian-predicates [fake-test australian-word-predicate])
+(def australian-predicates
+  [fake-test australian-word-predicate])
 
-(def scottish-predicates [fake-test scottish-word-predicate])
+(def scottish-predicates
+  [fake-test scottish-word-predicate])
 
-(def american-predicates [fake-test american-word-predicate])
+(def american-predicates
+  [fake-test american-word-predicate])
 
 
 ;; Take a sentence and figure out its dialect
@@ -83,7 +94,6 @@
 
 (cond)
 
-
 ;; 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,
 ;;when no dialect can be detected it defaults to standard. IE if there's a sample with 3 sentences, one reads as scottish,
@@ -138,6 +148,15 @@
 (defn show-dependencies []
   (clojure.pprint/pprint (dl/dependency-graph (nth sentences-one 1))))
 
+;; slurp a file containing a map of all the samples and their expected dialect
+;; detect each sample, collecting the results, and comparing against the expected
+;; output the results into both a file, as a map of the samples and if it was successful or not,
+;; and onto the command line as a fraction and percentage, with a breakdown per dialect
+
+(defn detect-dialects
+  [filename])
+  
+
 
 (comment
   ;; Test every annotator in the pipeline
@@ -342,6 +361,8 @@
   (slurp "test.txt")
   (read-string (slurp "test2.txt"))
 
+  (clojure.edn/read-string)
+
   ;; Don't use eval with arbitrary input
   (def horror (eval (read-string (slurp "test2.txt"))))