feat: regex sampler

Vendors a tiny dependency (total dependency tree <10 KB
uncompressed).

Other than removing unused code from the dependency, the only change
made is to change the following line:

https://github.com/fent/randexp.js/blob/2b35ea607883fa8cafa63bd3bdb5c12d695f873a/lib/randexp.js#L109

to a separately customizable callback.

package.json

@@ -88,6 +88,7 @@
   },
   "dependencies": {
     "@types/json-schema": "^7.0.7",
-    "json-pointer": "0.6.2"
+    "json-pointer": "0.6.2",
+    "randexp": "^0.5.3"
   }
 }

src/samplers/regex.js

@@ -0,0 +1,63 @@
+const RandExp = require('../vendor/randexp');
+
+const RANDOM_SEARCH_MAX_ITERS = 250;
+
+/**
+ * Generates a string that matches a given regex pattern and,
+ * if possible, falls within the given length constraints.
+ *
+ * For some advanced regexes we may fail to meet the length constraints exactly.
+ * (e.g. knapsacks or chicken mcnugget problems)
+ *
+ * @param {string} pattern
+ * @param {number} minLength
+ * @param {number} maxLength
+ * @return {string}
+ */
+export function sampleRegex(pattern, minLength = 0, maxLength = Infinity) {
+  const randexp = new RandExp(pattern);
+
+  RandExp.prototype.randIntRepetition = (from) => from;
+  const initialSample = randexp.gen();
+  if (initialSample.length >= minLength && initialSample <= maxLength) {
+    return initialSample;
+  }
+
+  // Add on average this many characters to each repetition
+  let seed = 1;
+  // Adjust the seed by this much when result is too low or too high.
+  let learningRate = 2;
+
+  RandExp.prototype.randIntRepetition = (from, to) => {
+    const roll = seed * Math.random() * 2;
+    const remainder = roll % 1;
+    const intRoll = Math.floor(roll) + (Math.random() < remainder ? 1 : 0);
+    return from + (intRoll % (to - from + 1));
+  };
+
+  let distance = Infinity; // how much the best generation is too low or high by
+  let best;
+  for (let i = 0; i < RANDOM_SEARCH_MAX_ITERS; i++) {
+    const sample = randexp.gen();
+
+    if (sample.length < minLength) {
+      seed *= learningRate;
+      if (minLength - sample.length < distance) {
+        distance = minLength - sample.length;
+        best = sample;
+      }
+    } else if (sample.length > maxLength) {
+      seed /= learningRate;
+      if (sample.length - maxLength < distance) {
+        distance = sample.length - maxLength;
+        best = sample;
+      }
+    } else {
+      return sample;
+    }
+
+    learningRate -= 1 / RANDOM_SEARCH_MAX_ITERS;
+  }
+
+  return best;
+}

src/samplers/string.js

@@ -1,9 +1,18 @@
 'use strict';
 
 import { ensureMinLength, toRFCDateTime, uuid } from '../utils';
+import { sampleRegex } from './regex';
 
 const passwordSymbols = 'qwerty!@#$%^123456';
 
+function truncateString(str, min, max) {
+  let res = ensureMinLength(str, min);
+  if (max && res.length > max) {
+    res = res.substring(0, max);
+  }
+  return res;
+}
+
 function emailSample() {
   return '[email protected]';
 }
@@ -42,12 +51,10 @@ function timeSample(min, max) {
   return commonDateTimeSample({ min, max, omitTime: false, omitDate: true }).slice(1);
 }
 
-function defaultSample(min, max) {
-  let res = ensureMinLength('string', min);
-  if (max && res.length > max) {
-    res = res.substring(0, max);
-  }
-  return res;
+function defaultSample(min, max, _propertyName, pattern) {
+  return pattern
+    ? sampleRegex(pattern, min, max)
+    : truncateString('string', min, max)
 }
 
 function ipv4Sample() {
@@ -96,8 +103,10 @@ function relativeJsonPointerSample() {
   return '1/relative/json/pointer';
 }
 
-function regexSample() {
-  return '/regex/';
+function regexSample(min, max, _propertyName, pattern) {
+  return pattern
+    ? sampleRegex(pattern, min, max)
+    : truncateString('/regex/', min, max)
 }
 
 const stringFormats = {
@@ -127,5 +136,10 @@ export function sampleString(schema, options, spec, context) {
   let format = schema.format || 'default';
   let sampler = stringFormats[format] || defaultSample;
   let propertyName = context && context.propertyName;
-  return sampler(schema.minLength | 0, schema.maxLength, propertyName);
+  return sampler(
+    schema.minLength || 0,
+    schema.maxLength,
+    propertyName,
+    schema.pattern,
+  );
 }