Sync word-count tests (#433)

Author: BNAndras

exercises/practice/word-count/.meta/config.json

@@ -4,6 +4,7 @@
   ],
   "contributors": [
     "arguello",
+    "BNAndras",
     "cousinitt",
     "PurityControl",
     "robertpostill",

exercises/practice/word-count/.meta/example.rkt

@@ -3,22 +3,8 @@
 (provide word-count)
 
 (define (word-count sentence)
-  (let ([normalized-sentence (normalize-input sentence)])
-    (for/fold ([result (hash)])
-              ([word (in-list (string-split normalized-sentence))])
-      (hash-set result word (add1 (hash-ref result word 0))))))
-
-(define (normalize-input s)
-  (handle-single-quoted-words
-   (list->string
-    (map
-     (λ (c)
-       (cond [(char<=? #\a c #\z) c]
-             [(char<=? #\0 c #\9) c]
-             [(char=? #\' c) c]
-             [else #\space]))
-     (string->list
-      (string-downcase s))))))
-
-(define (handle-single-quoted-words s)
-  (regexp-replace* #px"'\\b(\\w+)\\b'" s " \\1 "))
+  (let* ([lowered (string-downcase sentence)]
+         [words (regexp-match* #px"[a-z0-9]+(?:'[a-z0-9]+)?" lowered)])
+    (for/fold ([counts (hash)])
+              ([word (in-list words)])
+      (hash-set counts word (add1 (hash-ref counts word 0))))))

exercises/practice/word-count/.meta/tests.toml

@@ -35,6 +35,11 @@ description = "normalize case"
 
 [4185a902-bdb0-4074-864c-f416e42a0f19]
 description = "with apostrophes"
+include = false
+
+[4ff6c7d7-fcfc-43ef-b8e7-34ff1837a2d3]
+description = "with apostrophes"
+reimplements = "4185a902-bdb0-4074-864c-f416e42a0f19"
 
 [be72af2b-8afe-4337-b151-b297202e4a7b]
 description = "with quotations"
@@ -47,3 +52,6 @@ description = "multiple spaces not detected as a word"
 
 [50176e8a-fe8e-4f4c-b6b6-aa9cf8f20360]
 description = "alternating word separators not detected as a word"
+
+[6d00f1db-901c-4bec-9829-d20eb3044557]
+description = "quotation for word with apostrophe"

exercises/practice/word-count/word-count-test.rkt

@@ -1,68 +1,71 @@
 #lang racket
 
-; Tests adapted from `problem-specifications/canonical-data.json v1.4.0
 (require "word-count.rkt")
 
 (module+ test
   (require rackunit rackunit/text-ui)
 
-  (define-binary-check (hashes-equal? actual expected)
-    (equal? (sort (hash->list actual) string<? #:key car)
-            (sort (hash->list expected) string<? #:key car)))
-
   (define suite
     (test-suite
-     "For each word in the input, count the number of times it appears in the entire sentence."
-
-     (hashes-equal? (word-count "word")
-                    '#hash(("word" . 1))
-                    "count one word")
-
-     (hashes-equal? (word-count "one of each")
-                    '#hash(("one" . 1) ("of" . 1) ("each" . 1))
-                    "count one of each word")
-
-     (hashes-equal? (word-count "one fish two fish red fish blue fish")
-                    '#hash(("one" . 1) ("fish" . 4) ("two" . 1) ("red" . 1) ("blue" . 1))
-                    "multiple occurrences of a word")
-
-     (hashes-equal? (word-count "one,two,three")
-                    '#hash(("one" . 1) ("two" . 1) ("three" . 1))
-                    "handles cramped lists")
-
-     (hashes-equal? (word-count "one,\ntwo,\nthree")
-                    '#hash(("one" . 1) ("two" . 1) ("three" . 1))
-                    "handles expanded lists")
-
-     (hashes-equal? (word-count "car : carpet as java : javascript!!&@$%^&")
-                    '#hash(("car" . 1) ("carpet" . 1) ("as" . 1) ("java" . 1) ("javascript" . 1))
-                    "ignore punctuation")
-
-     (hashes-equal? (word-count "testing, 1, 2, testing")
-                    '#hash(("testing" . 2) ("1" . 1) ("2" . 1))
-                    "include numbers")
-
-     (hashes-equal? (word-count "go Go GO Stop stop")
-                    '#hash(("go" . 3) ("stop" . 2))
-                    "normalize case")
-
-     (hashes-equal? (word-count "Joe can't tell between 'large' and large.")
-                    '#hash(("joe" . 1) ("can't" . 1) ("tell" . 1) ("between" . 1) ("large" . 2)
-                                       ("and" . 1))
-                    "with quotations")
-
-     (hashes-equal? (word-count "Joe can't tell between app, apple and a.")
-                    '#hash(("joe" . 1) ("can't" . 1) ("tell" . 1) ("between" . 1) ("app" . 1)
-                                       ("apple" . 1) ("and" . 1) ("a" . 1))
-                    "substrings from the beginning")
-
-
-     (hashes-equal? (word-count " multiple   whitespaces")
-                    '#hash(("multiple" . 1) ("whitespaces" . 1))
-                    "multiple spaces not detected as a word")
-
-     (hashes-equal? (word-count ",\n,one,\n ,two \n 'three'")
-                    '#hash(("one" . 1) ("two" . 1) ("three" . 1))
-                    "alternating word separators not detected as a word")))
+     "word-count tests"
+
+     (test-equal? "count one word"
+                  (word-count "word")
+                  '#hash(("word" . 1)))
+
+     (test-equal? "count one of each word"
+                  (word-count "one of each")
+                  '#hash(("one" . 1) ("of" . 1) ("each" . 1)))
+
+     (test-equal? "multiple occurrences of a word"
+                  (word-count "one fish two fish red fish blue fish")
+                  '#hash(("one" . 1) ("fish" . 4) ("two" . 1) ("red" . 1) ("blue" . 1)))
+
+     (test-equal? "handles cramped lists"
+                  (word-count "one,two,three")
+                  '#hash(("one" . 1) ("two" . 1) ("three" . 1)))
+
+     (test-equal? "handles expanded lists"
+                  (word-count "one,\ntwo,\nthree")
+                  '#hash(("one" . 1) ("two" . 1) ("three" . 1)))
+
+     (test-equal? "ignore punctuation"
+                  (word-count "car : carpet as java : javascript!!&@$%^&")
+                  '#hash(("car" . 1) ("carpet" . 1) ("as" . 1) ("java" . 1) ("javascript" . 1)))
+
+     (test-equal? "include numbers"
+                  (word-count "testing, 1, 2, testing")
+                  '#hash(("testing" . 2) ("1" . 1) ("2" . 1)))
+
+     (test-equal? "normalize case"
+                  (word-count "go Go GO Stop stop")
+                  '#hash(("go" . 3) ("stop" . 2)))
+
+     (test-equal? "with apostrophes"
+                  (word-count "'First: don't laugh. Then: don't cry. You're getting it.'")
+                  '#hash(("first" . 1) ("don't" . 2) ("laugh" . 1) ("then" . 1) ("cry" . 1)
+                                       ("you're" . 1) ("getting" . 1) ("it" . 1)))
+
+     (test-equal? "with quotations"
+                  (word-count "Joe can't tell between 'large' and large.")
+                  '#hash(("joe" . 1) ("can't" . 1) ("tell" . 1) ("between" . 1) ("large" . 2)
+                                     ("and" . 1)))
+
+     (test-equal? "substrings from the beginning"
+                  (word-count "Joe can't tell between app, apple and a.")
+                  '#hash(("joe" . 1) ("can't" . 1) ("tell" . 1) ("between" . 1) ("app" . 1)
+                                     ("apple" . 1) ("and" . 1) ("a" . 1)))
+
+     (test-equal? "multiple spaces not detected as a word"
+                  (word-count " multiple   whitespaces")
+                  '#hash(("multiple" . 1) ("whitespaces" . 1)))
+
+     (test-equal? "alternating word separators not detected as a word"
+                  (word-count ",\n,one,\n ,two \n 'three'")
+                  '#hash(("one" . 1) ("two" . 1) ("three" . 1)))
+
+     (test-equal? "quotation for word with apostrophe"
+                  (word-count "can, can't, 'can't'")
+                  '#hash(("can" . 1) ("can't" . 2)))))
 
   (run-tests suite))

exercises/practice/word-count/word-count.rkt

@@ -2,5 +2,5 @@
 
 (provide word-count)
 
-(define (word-count string)
+(define (word-count sentence)
   (error "Not implemented yet"))