search: fuzzy matching via edit distance + 18 tests

editDist as an O(m*n) row-based Levenshtein DP (naive recursion is exponential
and times out under load); fuzzyTerms/fuzzyDocs/fuzzyRankTfIdf expand a term to
indexed terms within a max edit distance. 166/166.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

lib/search/fuzzy.sx

@@ -0,0 +1,12 @@
+;; search fuzzy matching — Haskell source fragment. Depends on index + rank.
+;; Levenshtein edit distance (O(m*n) row-based DP — the naive recursive version is
+;; exponential and far too slow under load) expands a query term to all indexed
+;; terms within a max distance, then unions / ranks their docs.
+;;   editDist        :: String -> String -> Int
+;;   fuzzyTerms      :: Int -> String -> Index -> [Term]   (sorted)
+;;   fuzzyDocs       :: Int -> String -> Index -> [DocId]  (sorted union)
+;;   fuzzyRankTfIdf  :: Int -> String -> Index -> [DocId]
+
+(define
+  search/fuzzy-src
+  "edMin3 a b c = min a (min b c)\nedCost x y = if x == y then 0 else 1\nedUpto i n = if i > n then [] else i : edUpto (i + 1) n\nedLast [x] = x\nedLast (x:xs) = edLast xs\nedNrow x [] prev left = []\nedNrow x (y:ys) prev left = let v = edMin3 (head (tail prev) + 1) (left + 1) (head prev + edCost x y) in v : edNrow x ys (tail prev) v\nedRow x ys prev = let f = head prev + 1 in f : edNrow x ys prev f\nedRows [] ys prev = prev\nedRows (x:xs) ys prev = edRows xs ys (edRow x ys prev)\neditDist xs ys = edLast (edRows xs ys (edUpto 0 (length ys)))\nqWithinDist maxd term t = editDist term t <= maxd\nfuzzyTerms maxd term idx = filter (qWithinDist maxd term) (allTerms idx)\nfuzzyDocs maxd term idx = foldl (candStep idx) [] (fuzzyTerms maxd term idx)\nfuzzyRankTfIdf maxd term idx = rankTfIdf (fuzzyTerms maxd term idx) idx\n")