search: did-you-mean spelling suggestion + 9 tests

suggest/suggestN rank indexed terms by edit distance to a (misspelled) query
term, alphabetical tiebreak. 234/234.

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

lib/search/api.sx

@@ -8,7 +8,7 @@
 ;; paginate, pageTfIdf, pageBm25, resultCount, editDist, fuzzyTerms, fuzzyDocs,
 ;; fuzzyRankTfIdf, highlight, snippet, stem, stemText, stemTokens, indexStemmed,
 ;; nearDocs, expandTerm, synDocs, synRankTfIdf, queryTerms, searchRankTfIdf,
-;; searchRankBm25.
+;; searchRankBm25, suggestN, suggest.
 
 (define
   search/src
@@ -39,4 +39,6 @@
     "\n"
     search/syn-src
     "\n"
-    search/rankq-src))
+    search/rankq-src
+    "\n"
+    search/suggest-src))

lib/search/conformance.conf

@@ -32,6 +32,7 @@ PRELOADS=(
   lib/search/near.sx
   lib/search/syn.sx
   lib/search/rankq.sx
+  lib/search/suggest.sx
   lib/search/api.sx
   lib/search/testlib.sx
 )
@@ -50,4 +51,5 @@ SUITES=(
   "near:lib/search/tests/near.sx"
   "syn:lib/search/tests/syn.sx"
   "rankq:lib/search/tests/rankq.sx"
+  "suggest:lib/search/tests/suggest.sx"
 )

lib/search/scoreboard.json

@@ -1,8 +1,8 @@
 {
   "lang": "search",
-  "total_passed": 225,
+  "total_passed": 234,
   "total_failed": 0,
-  "total": 225,
+  "total": 234,
   "suites": [
     {"name":"index","passed":18,"failed":0,"total":18},
     {"name":"boolean","passed":28,"failed":0,"total":28},
@@ -16,7 +16,8 @@
     {"name":"stem","passed":18,"failed":0,"total":18},
     {"name":"near","passed":9,"failed":0,"total":9},
     {"name":"syn","passed":9,"failed":0,"total":9},
-    {"name":"rankq","passed":11,"failed":0,"total":11}
+    {"name":"rankq","passed":11,"failed":0,"total":11},
+    {"name":"suggest","passed":9,"failed":0,"total":9}
   ],
-  "generated": "2026-06-06T23:58:05+00:00"
+  "generated": "2026-06-07T00:44:05+00:00"
 }

lib/search/scoreboard.md

@@ -1,6 +1,6 @@
 # search scoreboard
 
-**225 / 225 passing** (0 failure(s)).
+**234 / 234 passing** (0 failure(s)).
 
 | Suite | Passed | Total | Status |
 |-------|--------|-------|--------|
@@ -17,3 +17,4 @@
 | near | 9 | 9 | ok |
 | syn | 9 | 9 | ok |
 | rankq | 11 | 11 | ok |
+| suggest | 9 | 9 | ok |

lib/search/suggest.sx

@@ -0,0 +1,9 @@
+;; search did-you-mean / spelling suggestion — Haskell source fragment.
+;; Depends on fuzzy (editDist) + index (allTerms). Ranks indexed terms by edit
+;; distance to a (possibly misspelled) query term; ties broken alphabetically.
+;;   suggestN :: Int -> String -> Index -> [Term]
+;;   suggest  :: String -> Index -> Term   ("" if the index has no terms)
+
+(define
+  search/suggest-src
+  "sgMk term t = (editDist term t, t)\nsgPairs term idx = map (sgMk term) (allTerms idx)\nsgCmp p1 p2 = if fst p1 < fst p2 then LT else if fst p1 > fst p2 then GT else compare (snd p1) (snd p2)\nsuggestN n term idx = take n (map snd (sortBy sgCmp (sgPairs term idx)))\nsgHead [] = \"\"\nsgHead (x:xs) = x\nsuggest term idx = sgHead (suggestN 1 term idx)\n")

lib/search/tests/suggest.sx

@@ -0,0 +1,42 @@
+;; Extension — did-you-mean / spelling suggestion.
+;; Corpus terms (sorted): ample apple apply banana orange
+
+(define
+  suggest-setup
+  "idx = indexDoc 1 \"apple apply ample banana orange\" emptyIndex\n")
+
+(define
+  suggest-cases
+  (list
+    (list "suggest exact term" "[suggest \"apple\" idx]" (list "apple"))
+    (list
+      "suggest misspelled banana"
+      "[suggest \"bananna\" idx]"
+      (list "banana"))
+    (list
+      "suggest missing letter orange"
+      "[suggest \"orang\" idx]"
+      (list "orange"))
+    (list "suggest closest apply" "[suggest \"aply\" idx]" (list "apply"))
+    (list "suggestN 1 banana" "suggestN 1 \"bananna\" idx" (list "banana"))
+    (list
+      "suggestN 2 ties alpha"
+      "suggestN 2 \"aple\" idx"
+      (list "ample" "apple"))
+    (list "suggest empty term shortest" "[suggest \"\" idx]" (list "ample"))
+    (list "suggest empty index" "[suggest \"apple\" emptyIndex]" (list ""))
+    (list "suggestN empty index" "suggestN 1 \"apple\" emptyIndex" (list))))
+
+(define
+  suggest-results
+  (search-batch
+    suggest-setup
+    (map (fn (c) (nth c 1)) suggest-cases)))
+
+(map-indexed
+  (fn
+    (i c)
+    (hk-test (nth c 0) (nth suggest-results i) (nth c 2)))
+  suggest-cases)
+
+{:fail hk-test-fail :pass hk-test-pass :fails hk-test-fails}

plans/search-on-sx.md

@@ -122,9 +122,14 @@ lib/search/index.sx                     lib/search/eval.sx
 - [x] synonym / query expansion — `expandTerm`, `synDocs`, `synRankTfIdf` — 9 tests
 - [x] boolean-filtered ranked search — `queryTerms`, `searchRankTfIdf`,
   `searchRankBm25` (filter by boolean query, rank survivors by relevance) — 11 tests
+- [x] did-you-mean / spelling suggestion — `suggest`, `suggestN` (closest indexed
+  terms by edit distance, alphabetical tiebreak) — 9 tests
 
 ## Progress log
 
+- **Extension: did-you-mean / spelling suggestion (234/234 total).** `suggest`/`suggestN`
+  rank indexed terms by edit distance to a (misspelled) query term, alphabetical
+  tiebreak. 9 tests.
 - **Extension: boolean-filtered ranked search (225/225 total).** `searchRankTfIdf`/
   `searchRankBm25` parse a boolean query, filter docs via evalQuery, then rank the
   survivors by relevance over the query's leaf terms (`queryTerms`) — the real-world