Include rhyme in best-of calc

src/com/owoga/corpus/markov.clj

@@ -1032,47 +1032,51 @@
               banned-words
               (into #{} (->> existing-lines
                              (map (comp last last))))
-              seed (if existing-lines
+
-                     (->> existing-lines
+              line
-                          rand-nth
+              (take-until
-                          reverse
+               (best-of-20)
-                          (map first)
+               (fn []
-                          (apply concat)
+                 (let [seed (if existing-lines
-                          (#(rhymes
+                              (->> existing-lines
-                             rhyme-trie
+                                   rand-nth
-                             %
+                                   reverse
-                             (fn [choices]
+                                   (map first)
-                               (->> choices
+                                   (apply concat)
-                                    (map (fn [[phones wordset]]
+                                   (#(rhymes
-                                           [phones
+                                      rhyme-trie
-                                            (set/difference
+                                      %
-                                             wordset
+                                      (fn [choices]
-                                             banned-words)]))
+                                        (->> choices
-                                    (remove (comp empty? second))))))
+                                             (map (fn [[phones wordset]]
-                          rand-nth
+                                                    [phones
-                          ((fn [[phones wordset]]
+                                                     (set/difference
-                             (let [word (rand-nth (vec wordset))]
+                                                      wordset
-                               [(rand-nth (phonetics/get-phones word))
+                                                      banned-words)]))
-                                word])))
+                                             (remove (comp empty? second))))))
-                          vector)
+                                   rand-nth
-                     (->> (get-next-markov
+                                   ((fn [[phones wordset]]
-                           markov-trie
+                                      (let [word (rand-nth (vec wordset))]
-                           [eos eos eos]
+                                        [(rand-nth (phonetics/get-phones word))
-                           (fn [children]
+                                         word])))
-                             (remove
+                                   vector)
-                              #(#{eos bos} (.key %)) children)))
+                              (->> (get-next-markov
-                          database
+                                    markov-trie
-                          (#(vector (rand-nth (phonetics/get-phones %)) %))
+                                    [eos eos eos]
-                          vector))
+                                    (fn [children]
-              line (take-until
+                                      (remove
-                    (best-of-20)
+                                       #(#{eos bos} (.key %)) children)))
-                    #(tightly-generate-n-syllable-sentence-v2
+                                   database
-                      database
+                                   (#(vector (rand-nth (phonetics/get-phones %)) %))
-                      markov-trie
+                                   vector))
-                      4
+                       line (tightly-generate-n-syllable-sentence-v2
-                      syllable-count
+                             database
-                      (make-markov-filter [eos bos])
+                             markov-trie
-                      seed))]
+                             4
+                             syllable-count
+                             (make-markov-filter [eos bos])
+                             seed)]
+                   line)))]
           (recur (rest scheme)
                  (update result (first scheme) (fnil conj []) line)))))))
 
@@ -1081,6 +1085,19 @@
     (rhyme-from-scheme-v2
      scheme database markov-tight-trie rhymetrie))
 
+  (phonetics/get-phones "unleashed")
+  (rhymes
+   rhymetrie
+   ["IY" "SH" "T"]
+   (fn [choices]
+     (->> choices
+          (map (fn [[phones wordset]]
+                 [phones
+                  (set/difference
+                   wordset
+                   #{"unleashed"})]))
+          (remove (comp empty? second)))))
+
   )
 
 (comment

src/com/owoga/prhyme/nlp/core.clj

@@ -167,6 +167,12 @@
        (#(parse-top-n % 100)))
   )
 
+(comment
+  (likely-sentence?
+   "have s y como tu cama")
+
+  )
+
 (defn valid-sentence?
   "Tokenizes and parses the phrase using OpenNLP models from
   http://opennlp.sourceforge.net/models-1.5/