From 542d6b0119f18d6adcaf667a697a1e53d416fc17 Mon Sep 17 00:00:00 2001
From: Dustin Popp <dpopp07@gmail.com>
Date: Fri, 24 Jan 2025 11:22:29 -0600
Subject: [PATCH] fix(ibm-use-date-based-format): tighten heuristic for
 flagging date-time values (#717)

The regular expressions used to determine if a value should be considered a
date-time value are too loose and allow false positives to sneak in. It's a
tricky problem, but this commit makes an attempt to tighten them by dismissing
the possibility of a date-time value if any letters are found that are not
relevant to a date-time value.

Signed-off-by: Dustin Popp <dpopp07@gmail.com>
---
 .../ruleset/src/utils/date-based-utils.js     | 40 ++++++++++++++++---
 .../test/utils/date-based-utils.test.js       |  6 ++-
 2 files changed, 39 insertions(+), 7 deletions(-)

diff --git a/packages/ruleset/src/utils/date-based-utils.js b/packages/ruleset/src/utils/date-based-utils.js
index 3ac62615..68dd7891 100644
--- a/packages/ruleset/src/utils/date-based-utils.js
+++ b/packages/ruleset/src/utils/date-based-utils.js
@@ -1,5 +1,5 @@
 /**
- * Copyright 2024 IBM Corporation.
+ * Copyright 2024-2025 IBM Corporation.
  * SPDX-License-Identifier: Apache2.0
  */
 
@@ -77,12 +77,40 @@ function isDateBasedName(name) {
  * @returns a boolean value indicating that the value seems to be date-based
  */
 function isDateBasedValue(value) {
-  const regularExpressions = [
-    // Includes abbreviated month name.
-    /^\b(Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)\b/,
+  // The full and abbreviated values for months will be used in the generic
+  // string check below, as well as in the group of date-based expressions in
+  // the primary check below that.
+  const months =
+    /\b(Jan(uary)?|Feb(ruary)?|Mar(ch)?|Apr(il)?|May|Jun(e)?|Jul(y)?|Aug(ust)?|Sep(tember)?|Oct(ober)?|Nov(ember)?|Dec(ember)?)\b/g;
+
+  // In an effort to avoid false positives with strings that may contain
+  // date-time values, but are not themselves date-time values, check for
+  // the presence of any lowercase letters that are not included in the
+  // names of months or days.
+  if (typeof value === 'string') {
+    // "Day" strings are not enough on their own to determine date-time values,
+    // but we need to remove them from any values before we check for any
+    // letters that are not relevant to a date-time value.
+    const days =
+      /\b(Mon(day)?|Tue(sday)?|Wed(nesday)?|Thu(rsday)?|Fri(day)?|Sat(urday)?|Sun(day)?)\b/g;
+
+    // Only lowercase letters are checked for because 1) there are a number of
+    // valid date-time uses of uppercase letters like T, Z, GMT, UTC, etc. but
+    // not lowercase letters and 2) lowercase letters are more likely to
+    // indicate a general string over a formatted value.
+    const hasNonDateLetters = !!value
+      .replaceAll(months, '')
+      .replaceAll(days, '')
+      .match(/[a-z]/);
+
+    if (hasNonDateLetters) {
+      return false;
+    }
+  }
 
-    // Includes full month name.
-    /^\b(January|February|March|April|May|June|July|August|September|October|November|December)\b/,
+  const regularExpressions = [
+    // Includes full or abbreviated month name.
+    months,
 
     // Includes date in the format YYYY(./-)MM(./-)DD(T).
     /\b\d{4}[./-](0?[1-9]|1[012])[./-]([012]?[1-9]|3[01])(\b|T)/,
diff --git a/packages/ruleset/test/utils/date-based-utils.test.js b/packages/ruleset/test/utils/date-based-utils.test.js
index b3d50275..8f6a6113 100644
--- a/packages/ruleset/test/utils/date-based-utils.test.js
+++ b/packages/ruleset/test/utils/date-based-utils.test.js
@@ -1,5 +1,5 @@
 /**
- * Copyright 2024 IBM Corporation.
+ * Copyright 2024-2025 IBM Corporation.
  * SPDX-License-Identifier: Apache2.0
  */
 
@@ -26,7 +26,11 @@ describe('Date-based utility functions', () => {
       expect(isDateBasedValue('This certificate is good until June 2032')).toBe(
         false
       );
+      expect(
+        isDateBasedValue('0000-0000-0000-0000/abc/2022/2/22/12345678_data.json')
+      ).toBe(false);
       expect(isDateBasedValue('Octopus')).toBe(false);
+      expect(isDateBasedValue('Januaryuary 31')).toBe(false);
       expect(isDateBasedValue('12345678')).toBe(false);
       expect(isDateBasedValue('0001-01-2000')).toBe(false);
       expect(isDateBasedValue('10.1.24.1')).toBe(false);