strgen

A Clojure library with a test.check generator that generates strings from regular expressions. Use miner.strgen/string-generator with test.check and clojure.spec.

There is also a case-insensitive variant called miner.strgen/case-insensitive-string-generator which generates strings with mixed case even though they were not explicitly represented as such in the regex. This feature is somewhat analogous to setting the CASE_INSENSITIVE flag with Java regular expressions.

The clojure.spec library is new in Clojure 1.9.

Version

[com.velisco/strgen "0.2.5"]

Usage

(require '[miner.strgen :as sg])
(require '[clojure.spec.alpha :as s])
(require '[clojure.test.check.generators :as gen])

(gen/sample (sg/string-generator #"[A-Z]{2,4}"))

;;=> ("ZOHX" "ZOXZ" "INX" "JO" "MRMZ" "TO" "PEHM" "YNK" "FJ" "JWH")

(s/def ::foobar (let [re #"fo+(bar)?"] 
                   (s/spec (s/and string? #(re-matches re %))
                           :gen  #(sg/string-generator re))))
						   
(s/exercise ::foobar)

;;=> (["fo" "fo"] ["fobar" "fobar"] ["fo" "fo"] ["foo" "foo"] ["foooo" "foooo"] 
;;    ["fooo" "fooo"] ["fo" "fo"] ["foobar" "foobar"] ["fooooobar" "fooooobar"] 
;;    ["fooooobar" "fooooobar"])


(gen/sample (sg/case-insensitive-string-generator #"fo+ba[rz]"))

;;=> ("FobAz" "FOBAR" "fOBaZ" "FOOBAR" "fooBaZ" "FoOOoObar" "foooBAr" "foooObAZ" "fOOOoObAz"
;;    "fooooooooobar")

Limitations

Not all Java regular expressions are supported. The basics work, including: x* y? z+. [abc] [a-z] [^a] \n (a|b) \w \W \d \D \s \S x{N} x{N,} x{M,N}. ^x$ ignores the leading ^ and trailing $ as they generate no characters. (Capture groups) and \1 style back references are not supported. Character groups like [:alnum:] are not supported. All the other fancy flags, quantifiers, character classes, etc. are unsupported. In summary, if I couldn't use the regex feature without looking it up, I didn't support it. If people ask for something, I might work on it.

When generating X-or-more items for regular expressions such as #"x*" or #"y+", the generator limits the number of items to a reasonably small count. You can control this with an optional second arg or-more-limit (an integer, default 9) when calling string-generator.

Related Projects

If you need better support for Java regular expressions when generating Strings, you should consider using the test.chuck library which provides the string-from-regex generator.

License