Add documentation link, comment code

1 files changed, 15 insertions, 10 deletions

diff --git a/bible/encode.clj b/bible/encode.clj
index 084db0a..ac6195d 100644
--- a/bible/encode.clj
+++ b/bible/encode.clj
@@ -109,6 +109,7 @@
 
 ;;; Build the canonical encodings
 
+;; From: https://en.wikipedia.org/wiki/Canonical_Huffman_code
 ;; Each of the existing codes are replaced with a new one of the same length, using the following algorithm:
 
 ;;     The first symbol in the list gets assigned a codeword which is the same length as the symbol's original codeword but all zeros. This will often be a single zero ('0').
@@ -147,23 +148,27 @@
                (bit-shift-left 
                  next-base-code 
                  (- prev-len
-                    (- 63 (Long/numberOfLeadingZeros next-base-code))))
+                    (- 63 (Long/numberOfLeadingZeros next-base-code)))) ;Not 100% confident about this
                (inc prev-len)))]
        (recur (conj codes next-codeword) (rest symbols)))
      codes)))
 
-;(0001 next 00011)
-;(0000 next 1<<<<)
-;
-
 (def canonical-encodings
   (build-canonical-encodings sorted-huffman-tree-codewords))
 
 (assert
   (= (map #(.length %1) sorted-huffman-tree-codewords)
-     (map #(.length %1) sorted-huffman-tree-codewords)))
+     (map #(.length %1) sorted-huffman-tree-codewords))
+  "Some of the codes changed length when canonicalizing")
+
+(assert
+  (= (count canonical-encodings)
+     (count (set (map #(Long/toBinaryString (.code %1)) canonical-encodings))))
+  "There appears to be duplicate canonical encodings")
 
-(take 10 (sort-by (juxt (comp count val) key) huffman-tree-syms))
-(take 10 sorted-huffman-tree-codewords)
-(take 10 canonical-encodings)
-(take 10 (map #(Long/toBinaryString (.code %1)) canonical-encodings)) ;; The results of this *seem* wrong.
+(comment
+  "Some basic stuff"
+  (take 10 (sort-by (juxt (comp count val) key) huffman-tree-syms))
+  (take 10 sorted-huffman-tree-codewords)
+  (take 10 canonical-encodings)
+  (take 10 (map #(Long/toBinaryString (.code %1)) canonical-encodings)))