@askeleven/sms-guard 0.1.0

package/src/data/gsm.js

@@ -0,0 +1,53 @@
+/**
+ * The GSM 03.38 default alphabet (3GPP TS 23.038), in code order, without the escape
+ * at 0x1B. Non-ASCII characters are written as escapes so no lookalike can creep in.
+ * Note that it is not ASCII: 0x24 is the currency sign, and the backtick is missing.
+ */
+export const GSM_BASIC = new Set(
+  [
+    // 0x00
+    '@\u00A3$\u00A5\u00E8\u00E9\u00F9\u00EC\u00F2\u00C7\n\u00D8\u00F8\r\u00C5\u00E5',
+    // 0x10, escape at 0x1B left out
+    '\u0394_\u03A6\u0393\u039B\u03A9\u03A0\u03A8\u03A3\u0398\u039E\u00C6\u00E6\u00DF\u00C9',
+    // 0x20
+    ' !"#\u00A4%&\'()*+,-./0123456789:;<=>?',
+    // 0x40
+    '\u00A1ABCDEFGHIJKLMNOPQRSTUVWXYZ\u00C4\u00D6\u00D1\u00DC\u00A7',
+    // 0x60
+    '\u00BFabcdefghijklmnopqrstuvwxyz\u00E4\u00F6\u00F1\u00FC\u00E0',
+  ].join(''),
+)
+
+/**
+ * The extension table: each is sent as the escape plus one septet, so it costs two.
+ */
+export const GSM_EXTENSION = new Set(['\f', '^', '{', '}', '\\', '[', '~', ']', '|', '\u20AC'])
+
+/**
+ * Typographic characters with a GSM-7 equivalent that reads the same. Only these are
+ * ever offered as replacements. Accented letters are real text and are left alone.
+ *
+ * @type {Map<string, string>}
+ */
+export const GSM_REPLACEMENTS = new Map([
+  // Single quotes and primes
+  ['\u2018', "'"], ['\u2019', "'"], ['\u201A', "'"], ['\u201B', "'"], ['\u2032', "'"],
+  ['\u2039', "'"], ['\u203A', "'"],
+  // Double quotes, guillemets
+  ['\u201C', '"'], ['\u201D', '"'], ['\u201E', '"'], ['\u201F', '"'], ['\u2033', '"'],
+  ['\u00AB', '"'], ['\u00BB', '"'],
+  // Dashes and the minus sign
+  ['\u2010', '-'], ['\u2011', '-'], ['\u2012', '-'], ['\u2013', '-'], ['\u2014', '-'],
+  ['\u2015', '-'], ['\u2212', '-'],
+  // Ellipsis
+  ['\u2026', '...'],
+  // Bullets
+  ['\u2022', '-'], ['\u2023', '-'], ['\u2043', '-'],
+  // Tab, no-break and other spaces
+  ['\t', ' '], ['\u00A0', ' '], ['\u2000', ' '], ['\u2001', ' '], ['\u2002', ' '],
+  ['\u2003', ' '], ['\u2004', ' '], ['\u2005', ' '], ['\u2006', ' '], ['\u2007', ' '],
+  ['\u2008', ' '], ['\u2009', ' '], ['\u200A', ' '], ['\u202F', ' '], ['\u205F', ' '],
+  ['\u3000', ' '],
+  // Invisible: zero-width space, word joiner, byte order mark
+  ['\u200B', ''], ['\u2060', ''], ['\uFEFF', ''],
+])

package/src/data/nanp.js

@@ -0,0 +1,271 @@
+/**
+ * Geographic NANP area codes for the US, its territories, Canada, and the Caribbean
+ * countries on +1, mapped to IANA time zones.
+ *
+ * An area code whose territory crosses a time zone line lists every zone it covers,
+ * and sms-guard reports it as ambiguous instead of picking one. Zones follow the
+ * official boundaries, not local habit (a town that unofficially keeps its
+ * neighbour's time is not listed separately). Where we were not sure an area code
+ * stays inside one zone, it is listed as ambiguous: a question to a person is cheaper
+ * than a text at the wrong hour.
+ *
+ * Non-geographic codes (toll-free 8XX, 500, 600, 700, 900) are left out on purpose:
+ * they say nothing about where the phone is.
+ *
+ * Each row is [region, zone or zones, area codes].
+ *
+ * @type {Array<[string, string | string[], string[]]>}
+ */
+const ROWS = [
+  // United States
+  ['Alabama', 'America/Chicago', ['205', '251', '256', '334', '659', '938']],
+  // The Aleutians west of 169.5W keep Hawaii-Aleutian time.
+  ['Alaska', ['America/Anchorage', 'America/Adak'], ['907']],
+  ['Arizona', 'America/Phoenix', ['480', '520', '602', '623']],
+  // Northern Arizona includes the Navajo Nation, which observes daylight saving time.
+  ['Arizona', ['America/Phoenix', 'America/Denver'], ['928']],
+  ['Arkansas', 'America/Chicago', ['327', '479', '501', '870']],
+  [
+    'California',
+    'America/Los_Angeles',
+    [
+      '209', '213', '279', '310', '323', '341', '350', '369', '408', '415', '424',
+      '442', '510', '530', '559', '562', '619', '626', '628', '650', '657', '661',
+      '669', '707', '714', '747', '760', '805', '818', '820', '831', '840', '858',
+      '909', '916', '925', '949', '951',
+    ],
+  ],
+  ['Colorado', 'America/Denver', ['303', '719', '720', '970', '983']],
+  ['Connecticut', 'America/New_York', ['203', '475', '860', '959']],
+  ['Delaware', 'America/New_York', ['302']],
+  ['District of Columbia', 'America/New_York', ['202', '771']],
+  [
+    'Florida',
+    'America/New_York',
+    [
+      '239', '305', '321', '324', '352', '386', '407', '561', '645', '656', '689',
+      '727', '754', '772', '786', '813', '863', '904', '941', '954',
+    ],
+  ],
+  // The panhandle west of the Apalachicola River is Central.
+  ['Florida', ['America/New_York', 'America/Chicago'], ['448', '850']],
+  [
+    'Georgia',
+    'America/New_York',
+    ['229', '404', '470', '478', '678', '706', '762', '770', '912'],
+  ],
+  ['Hawaii', 'Pacific/Honolulu', ['808']],
+  // The panhandle is Pacific, the south is Mountain.
+  ['Idaho', ['America/Boise', 'America/Los_Angeles'], ['208', '986']],
+  [
+    'Illinois',
+    'America/Chicago',
+    [
+      '217', '224', '309', '312', '331', '447', '464', '618', '630', '708', '730',
+      '773', '779', '815', '847', '872',
+    ],
+  ],
+  ['Indiana', 'America/Indiana/Indianapolis', ['260', '317', '463', '765']],
+  // Northwest and southwest Indiana are Central; the rest is Eastern.
+  [
+    'Indiana',
+    ['America/Chicago', 'America/Indiana/Indianapolis'],
+    ['219', '574', '812', '930'],
+  ],
+  ['Iowa', 'America/Chicago', ['319', '515', '563', '641', '712']],
+  ['Kansas', 'America/Chicago', ['316', '913']],
+  // Four western counties are Mountain.
+  ['Kansas', ['America/Chicago', 'America/Denver'], ['620', '785']],
+  ['Kentucky', 'America/Kentucky/Louisville', ['502', '859']],
+  ['Kentucky', ['America/Kentucky/Louisville', 'America/Chicago'], ['270', '364', '606']],
+  ['Louisiana', 'America/Chicago', ['225', '318', '337', '504', '985']],
+  ['Maine', 'America/New_York', ['207']],
+  ['Maryland', 'America/New_York', ['227', '240', '301', '410', '443', '667']],
+  [
+    'Massachusetts',
+    'America/New_York',
+    ['339', '351', '413', '508', '617', '774', '781', '857', '978'],
+  ],
+  [
+    'Michigan',
+    'America/Detroit',
+    [
+      '231', '248', '269', '313', '517', '586', '616', '679', '734', '810', '947',
+      '989',
+    ],
+  ],
+  // Four Upper Peninsula counties on the Wisconsin border are Central.
+  ['Michigan', ['America/Detroit', 'America/Menominee'], ['906']],
+  ['Minnesota', 'America/Chicago', ['218', '320', '507', '612', '651', '763', '952']],
+  ['Mississippi', 'America/Chicago', ['228', '601', '662', '769']],
+  [
+    'Missouri',
+    'America/Chicago',
+    ['314', '417', '557', '573', '636', '660', '816', '975'],
+  ],
+  ['Montana', 'America/Denver', ['406']],
+  ['Nebraska', 'America/Chicago', ['402', '531']],
+  ['Nebraska', ['America/Chicago', 'America/Denver'], ['308']],
+  ['Nevada', 'America/Los_Angeles', ['702', '725']],
+  // West Wendover is officially Mountain.
+  ['Nevada', ['America/Los_Angeles', 'America/Denver'], ['775']],
+  ['New Hampshire', 'America/New_York', ['603']],
+  [
+    'New Jersey',
+    'America/New_York',
+    ['201', '551', '609', '640', '732', '848', '856', '862', '908', '973'],
+  ],
+  ['New Mexico', 'America/Denver', ['505', '575']],
+  [
+    'New York',
+    'America/New_York',
+    [
+      '212', '315', '329', '332', '347', '363', '516', '518', '585', '607', '624',
+      '631', '646', '680', '716', '718', '838', '845', '914', '917', '929', '934',
+    ],
+  ],
+  [
+    'North Carolina',
+    'America/New_York',
+    ['252', '336', '472', '704', '743', '828', '910', '919', '980', '984'],
+  ],
+  // The southwest is Mountain.
+  ['North Dakota', ['America/Chicago', 'America/Denver'], ['701']],
+  [
+    'Ohio',
+    'America/New_York',
+    [
+      '216', '220', '234', '283', '326', '330', '380', '419', '436', '440', '513',
+      '567', '614', '740', '937',
+    ],
+  ],
+  ['Oklahoma', 'America/Chicago', ['405', '539', '572', '580', '918']],
+  ['Oregon', 'America/Los_Angeles', ['503', '971']],
+  // Most of Malheur County is Mountain.
+  ['Oregon', ['America/Los_Angeles', 'America/Boise'], ['458', '541']],
+  [
+    'Pennsylvania',
+    'America/New_York',
+    [
+      '215', '223', '267', '272', '412', '445', '484', '570', '582', '610', '717',
+      '724', '814', '835', '878',
+    ],
+  ],
+  ['Rhode Island', 'America/New_York', ['401']],
+  ['South Carolina', 'America/New_York', ['803', '821', '839', '843', '854', '864']],
+  // East river is Central, west river is Mountain.
+  ['South Dakota', ['America/Chicago', 'America/Denver'], ['605']],
+  ['Tennessee', 'America/Chicago', ['615', '629', '731', '901']],
+  ['Tennessee', 'America/New_York', ['865']],
+  // Both cover counties close to the Eastern/Central line.
+  ['Tennessee', ['America/New_York', 'America/Chicago'], ['423', '931']],
+  [
+    'Texas',
+    'America/Chicago',
+    [
+      '210', '214', '254', '281', '325', '346', '361', '409', '430', '432', '469',
+      '512', '682', '713', '726', '737', '806', '817', '830', '832', '903', '936',
+      '940', '945', '956', '972', '979',
+    ],
+  ],
+  // El Paso and Hudspeth counties.
+  ['Texas', 'America/Denver', ['915']],
+  ['Utah', 'America/Denver', ['385', '435', '801']],
+  ['Vermont', 'America/New_York', ['802']],
+  [
+    'Virginia',
+    'America/New_York',
+    ['276', '434', '540', '571', '686', '703', '757', '804', '826', '948'],
+  ],
+  ['Washington', 'America/Los_Angeles', ['206', '253', '360', '425', '509', '564']],
+  ['West Virginia', 'America/New_York', ['304', '681']],
+  [
+    'Wisconsin',
+    'America/Chicago',
+    ['262', '274', '353', '414', '534', '608', '715', '920'],
+  ],
+  ['Wyoming', 'America/Denver', ['307']],
+
+  // US territories
+  ['Puerto Rico', 'America/Puerto_Rico', ['787', '939']],
+  ['US Virgin Islands', 'America/St_Thomas', ['340']],
+  ['Guam', 'Pacific/Guam', ['671']],
+  ['Northern Mariana Islands', 'Pacific/Saipan', ['670']],
+  ['American Samoa', 'Pacific/Pago_Pago', ['684']],
+
+  // Canada
+  ['Alberta', 'America/Edmonton', ['368', '403', '587', '780', '825']],
+  // 604 is the Lower Mainland only. The province-wide overlays also cover the East
+  // Kootenays (Mountain) and the Peace River region (Mountain, no daylight saving).
+  ['British Columbia', 'America/Vancouver', ['604']],
+  [
+    'British Columbia',
+    ['America/Vancouver', 'America/Edmonton', 'America/Dawson_Creek'],
+    ['236', '250', '672', '778'],
+  ],
+  ['Manitoba', 'America/Winnipeg', ['204', '431', '584']],
+  ['New Brunswick', 'America/Moncton', ['428', '506']],
+  // Labrador is mostly on Atlantic time.
+  ['Newfoundland and Labrador', ['America/St_Johns', 'America/Goose_Bay'], ['709', '879']],
+  ['Nova Scotia and Prince Edward Island', 'America/Halifax', ['782', '902']],
+  [
+    'Ontario',
+    'America/Toronto',
+    [
+      '226', '249', '289', '343', '365', '382', '416', '437', '519', '548', '613',
+      '647', '683', '705', '742', '753', '905', '942',
+    ],
+  ],
+  // Northwestern Ontario west of 90W is Central; Atikokan keeps Eastern Standard.
+  ['Ontario', ['America/Toronto', 'America/Winnipeg', 'America/Atikokan'], ['807']],
+  [
+    'Quebec',
+    'America/Toronto',
+    ['263', '354', '438', '450', '468', '514', '579', '819', '873'],
+  ],
+  // Eastern Quebec includes the Magdalen Islands and the Lower North Shore, which
+  // are on Atlantic time.
+  ['Quebec', ['America/Toronto', 'America/Halifax', 'America/Blanc-Sablon'], ['367', '418', '581']],
+  // Lloydminster, on the Alberta border, keeps Alberta time.
+  ['Saskatchewan', ['America/Regina', 'America/Edmonton'], ['306', '474', '639']],
+  [
+    'Yukon, Northwest Territories and Nunavut',
+    ['America/Whitehorse', 'America/Edmonton', 'America/Winnipeg', 'America/Toronto'],
+    ['867'],
+  ],
+
+  // Caribbean and Atlantic countries on +1
+  ['Anguilla', 'America/Anguilla', ['264']],
+  ['Antigua and Barbuda', 'America/Antigua', ['268']],
+  ['Bahamas', 'America/Nassau', ['242']],
+  ['Barbados', 'America/Barbados', ['246']],
+  ['Bermuda', 'Atlantic/Bermuda', ['441']],
+  ['British Virgin Islands', 'America/Tortola', ['284']],
+  ['Cayman Islands', 'America/Cayman', ['345']],
+  ['Dominica', 'America/Dominica', ['767']],
+  ['Dominican Republic', 'America/Santo_Domingo', ['809', '829', '849']],
+  ['Grenada', 'America/Grenada', ['473']],
+  ['Jamaica', 'America/Jamaica', ['658', '876']],
+  ['Montserrat', 'America/Montserrat', ['664']],
+  ['Saint Kitts and Nevis', 'America/St_Kitts', ['869']],
+  ['Saint Lucia', 'America/St_Lucia', ['758']],
+  ['Saint Vincent and the Grenadines', 'America/St_Vincent', ['784']],
+  ['Sint Maarten', 'America/Lower_Princes', ['721']],
+  ['Trinidad and Tobago', 'America/Port_of_Spain', ['868']],
+  ['Turks and Caicos Islands', 'America/Grand_Turk', ['649']],
+]
+
+/**
+ * @typedef {object} ZoneEntry
+ * @property {string} region State, province, territory or country.
+ * @property {string[]} zones IANA zones. More than one means ambiguous.
+ */
+
+/** @type {Map<string, ZoneEntry>} */
+export const NANP_AREA_CODES = new Map()
+for (const [region, zone, codes] of ROWS) {
+  for (const code of codes) {
+    if (NANP_AREA_CODES.has(code)) throw new Error(`Area code ${code} is listed twice.`)
+    NANP_AREA_CODES.set(code, { region, zones: Array.isArray(zone) ? zone : [zone] })
+  }
+}

package/src/data/shorteners.js

@@ -0,0 +1,10 @@
+/**
+ * Public URL shorteners. Carriers filter links on shared shortener domains, because
+ * anyone can hide any destination behind one. Matched exactly, and on subdomains.
+ */
+export const SHORTENER_DOMAINS = new Set([
+  'bit.ly', 'bitly.com', 'buff.ly', 'cutt.ly', 'goo.gl', 'is.gd', 'lnkd.in', 'ow.ly',
+  'rb.gy', 'rebrand.ly', 'shorturl.at', 't.co', 't.ly', 'tiny.cc', 'tinyurl.com',
+  'v.gd', 'x.co', 'bl.ink', 'short.io', 'shorte.st', 'soo.gd', 'clck.ru', 'qrco.de',
+  's.id', 'urlz.fr', 'tr.ee',
+])