Add checks for profanity to profile updates (#12445)

* fix(profile): detect attempt to use banned words as display name. refactor profanity detection method.

* fix(profile): detect attempt to use banned words in blurb. further refactor profanity detection. inform the user their chat privileges have been revoked.

* refactor: add function to normalize Unicode strings and remove diacritics

* fix: improve regEx to prevent false partial matches e.g. 'hello' being recognised as banned words. porting fix from #12309

* fix(profile): refactor of profanity detection for #12445

* fix(profile): add test for swear words in new profile. fix existing tests

* fix(profile): show different error message for attempted slur use in username by new users.

* fix(profile): remove incorrect slur test

* fix(profile): fix slurs not caught at start of end of strings connect by punctuation

* tests(profile): fix tests for profanity checking

* remove exclusive test

* 11865 - update text for slur warnings

* 11865 - remove unused string from locale files

* 11865 - improve naming of banned word usage locale string

* 11865 - improve logic so that differentiated warnings are shown depending on whether a slur or other profanity has been used in a display name

* 11865 - construct slur regexes outside the validation function in which they are used

* 11865 - fix tests
This commit is contained in:
Carlton McFarlane
2021-04-30 22:47:39 +02:00
committed by GitHub
parent 4d7304ab8d
commit a53355872b
9 changed files with 82 additions and 39 deletions

View File

@@ -91,7 +91,15 @@ describe('PUT /user', () => {
})).to.eventually.be.rejected.and.eql({
code: 400,
error: 'BadRequest',
message: t('displaynameIssueSlur'),
message: t('bannedSlurUsedInProfile'),
});
await expect(user.put('/user', {
'profile.name': 'TESTPLACEHOLDERSWEARWORDHERE',
})).to.eventually.be.rejected.and.eql({
code: 400,
error: 'BadRequest',
message: t('bannedWordUsedInProfile'),
});
await expect(user.put('/user', {

View File

@@ -139,7 +139,7 @@ describe('PUT /user/auth/update-username', async () => {
})).to.eventually.be.rejected.and.eql({
code: 400,
error: 'BadRequest',
message: [t('usernameIssueLength'), t('usernameIssueSlur')].join(' '),
message: [t('usernameIssueLength'), t('bannedSlurUsedInProfile')].join(' '),
});
});
@@ -149,21 +149,14 @@ describe('PUT /user/auth/update-username', async () => {
})).to.eventually.be.rejected.and.eql({
code: 400,
error: 'BadRequest',
message: [t('usernameIssueLength'), t('usernameIssueSlur')].join(' '),
message: [t('usernameIssueLength'), t('bannedSlurUsedInProfile')].join(' '),
});
await expect(user.put(ENDPOINT, {
username: 'something_TESTPLACEHOLDERSLURWORDHERE',
})).to.eventually.be.rejected.and.eql({
code: 400,
error: 'BadRequest',
message: [t('usernameIssueLength'), t('usernameIssueSlur')].join(' '),
});
await expect(user.put(ENDPOINT, {
username: 'somethingTESTPLACEHOLDERSLURWORDHEREotherword',
})).to.eventually.be.rejected.and.eql({
code: 400,
error: 'BadRequest',
message: [t('usernameIssueLength'), t('usernameIssueSlur')].join(' '),
message: [t('usernameIssueLength'), t('bannedSlurUsedInProfile')].join(' '),
});
});

View File

@@ -33,19 +33,22 @@ describe('POST /user/auth/verify-display-name', async () => {
it('errors if display name is a slur', async () => {
await expect(user.post(ENDPOINT, {
displayName: 'TESTPLACEHOLDERSLURWORDHERE',
})).to.eventually.eql({ isUsable: false, issues: [t('displaynameIssueSlur')] });
})).to.eventually.eql({ isUsable: false, issues: [t('bannedSlurUsedInProfile')] });
});
it('errors if display name contains a slur', async () => {
await expect(user.post(ENDPOINT, {
displayName: 'TESTPLACEHOLDERSLURWORDHERE_otherword',
})).to.eventually.eql({ isUsable: false, issues: [t('displaynameIssueLength'), t('displaynameIssueSlur')] });
})).to.eventually.eql({
isUsable: false,
issues: [t('displaynameIssueLength'), t('bannedSlurUsedInProfile')],
});
await expect(user.post(ENDPOINT, {
displayName: 'something_TESTPLACEHOLDERSLURWORDHERE',
})).to.eventually.eql({ isUsable: false, issues: [t('displaynameIssueLength'), t('displaynameIssueSlur')] });
await expect(user.post(ENDPOINT, {
displayName: 'somethingTESTPLACEHOLDERSLURWORDHEREotherword',
})).to.eventually.eql({ isUsable: false, issues: [t('displaynameIssueLength'), t('displaynameIssueSlur')] });
})).to.eventually.eql({
isUsable: false,
issues: [t('displaynameIssueLength'), t('bannedSlurUsedInProfile')],
});
});
it('errors if display name has incorrect length', async () => {

View File

@@ -51,9 +51,6 @@ describe('POST /user/auth/verify-username', async () => {
await expect(user.post(ENDPOINT, {
username: 'something_TESTPLACEHOLDERSLURWORDHERE',
})).to.eventually.eql({ isUsable: false, issues: [t('usernameIssueLength'), t('usernameIssueSlur')] });
await expect(user.post(ENDPOINT, {
username: 'somethingTESTPLACEHOLDERSLURWORDHEREotherword',
})).to.eventually.eql({ isUsable: false, issues: [t('usernameIssueLength'), t('usernameIssueSlur')] });
});
it('errors if username is not allowed', async () => {

View File

@@ -173,11 +173,16 @@
"usernameIssueInvalidCharacters": "Usernames can only contain letters a to z, numbers 0 to 9, hyphens, or underscores.",
"currentUsername": "Current username:",
"displaynameIssueLength": "Display Names must be between 1 and 30 characters.",
"displaynameIssueSlur": "Display Names may not contain inappropriate language.",
"bannedWordUsedInProfile": "Your Display Name or About text contained inappropriate language.",
"displaynameIssueNewline": "Display Names may not contain backslashes followed by the letter N.",
"goToSettings": "Go to Settings",
"usernameVerifiedConfirmation": "Your username, <%= username %>, is confirmed!",
"usernameNotVerified": "Please confirm your username.",
"changeUsernameDisclaimer": "Your username is used for invitations, @mentions in chat, and messaging. It must be 1 to 20 characters, containing only letters a to z, numbers 0 to 9, hyphens, or underscores, and cannot include any inappropriate terms.",
"verifyUsernameVeteranPet": "One of these Veteran Pets will be waiting for you after you've finished confirming!"
"verifyUsernameVeteranPet": "One of these Veteran Pets will be waiting for you after you've finished confirming!",
"mentioning": "Mentioning",
"suggestMyUsername": "Suggest my username",
"everywhere": "Everywhere",
"onlyPrivateSpaces": "Only in private spaces",
"bannedSlurUsedInProfile": "Your Display Name or About text contained a slur, and your chat privileges have been revoked."
}

View File

@@ -213,7 +213,7 @@ api.updateUsername = {
const newUsername = req.body.username;
const issues = verifyUsername(newUsername, res);
const issues = verifyUsername(newUsername, res, false);
if (issues.length > 0) throw new BadRequest(issues.join(' '));
const { password } = req.body;

View File

@@ -1,3 +1,7 @@
export function normalizeUnicodeString (str) {
return str.normalize('NFD').replace(/[\u0300-\u036f]/g, '');
}
export function removePunctuationFromString (str) {
return str.replace(/[.,/#!@$%^&;:{}=\-_`~()]/g, ' ');
}
@@ -12,15 +16,16 @@ export function getMatchesByWordArray (str, wordsToMatch) {
// https://www.unicode.org/reports/tr15/#Canon_Compat_Equivalence
// https://unicode-table.com/en/#combining-diacritical-marks
const normalizedStr = str.normalize('NFD').replace(/[\u0300-\u036f]/g, '');
const matchedWords = [];
const wordRegexs = wordsToMatch.map(word => new RegExp(`\\b([^a-z]+)?${word}([^a-z]+)?\\b`, 'i'));
const wordRegexs = wordsToMatch.map(word => {
const normalizedWord = removePunctuationFromString(normalizeUnicodeString(word));
return new RegExp(`\\b([^a-z]+)?${normalizedWord}([^a-z]+)?\\b`, 'i');
});
for (let i = 0; i < wordRegexs.length; i += 1) {
const regEx = wordRegexs[i];
const match = normalizedStr.match(regEx);
const match = removePunctuationFromString(normalizeUnicodeString(str)).match(regEx);
if (match !== null && match[0] !== null) {
const trimmedMatch = removePunctuationFromString(match[0]).trim();
const trimmedMatch = match[0].trim();
matchedWords.push(trimmedMatch);
}
}

View File

@@ -7,7 +7,7 @@ import {
} from '../errors';
import { model as User, schema as UserSchema } from '../../models/user';
import { model as NewsPost } from '../../models/newsPost';
import { nameContainsSlur, nameContainsNewline } from './validation';
import { stringContainsProfanity, nameContainsNewline } from './validation';
export async function get (req, res, { isV3 = false }) {
const { user } = res.locals;
@@ -101,6 +101,19 @@ function checkPreferencePurchase (user, path, item) {
return _.get(user.purchased, itemPath);
}
async function checkNewInputForProfanity (user, res, newValue) {
const containsSlur = stringContainsProfanity(newValue, 'slur');
const containsBannedWord = stringContainsProfanity(newValue);
if (containsSlur || containsBannedWord) {
if (containsSlur) {
user.flags.chatRevoked = true;
await user.save();
throw new BadRequest(res.t('bannedSlurUsedInProfile'));
}
throw new BadRequest(res.t('bannedWordUsedInProfile'));
}
}
export async function update (req, res, { isV3 = false }) {
const { user } = res.locals;
@@ -110,8 +123,13 @@ export async function update (req, res, { isV3 = false }) {
const newName = req.body['profile.name'];
if (newName === null) throw new BadRequest(res.t('invalidReqParams'));
if (newName.length > 30) throw new BadRequest(res.t('displaynameIssueLength'));
if (nameContainsSlur(newName)) throw new BadRequest(res.t('displaynameIssueSlur'));
if (nameContainsNewline(newName)) throw new BadRequest(res.t('displaynameIssueNewline'));
await checkNewInputForProfanity(user, res, newName);
}
if (req.body['profile.blurb'] !== undefined) {
const newBlurb = req.body['profile.blurb'];
await checkNewInputForProfanity(user, res, newBlurb);
}
_.each(req.body, (val, key) => {

View File

@@ -1,13 +1,23 @@
import bannedSlurs from '../bannedSlurs';
import { getMatchesByWordArray } from '../stringUtils';
import bannedWords from '../bannedWords';
import {
getMatchesByWordArray,
normalizeUnicodeString,
removePunctuationFromString,
} from '../stringUtils';
import forbiddenUsernames from '../forbiddenUsernames';
const bannedSlurRegexs = bannedSlurs.map(word => new RegExp(`.*${word}.*`, 'i'));
const bannedSlurRegexes = bannedSlurs.map(word => new RegExp(`\\b([^a-z]+)?${word}([^a-z]+)?\\b`, 'i'));
const bannedWordRegexes = bannedWords.map(word => new RegExp(`\\b([^a-z]+)?${word}([^a-z]+)?\\b`, 'i'));
export function nameContainsSlur (username) {
for (let i = 0; i < bannedSlurRegexs.length; i += 1) {
const regEx = bannedSlurRegexs[i];
const match = username.match(regEx);
export function stringContainsProfanity (str, profanityType = 'bannedWord') {
const bannedRegexes = profanityType === 'slur'
? bannedSlurRegexes
: bannedWordRegexes;
for (let i = 0; i < bannedRegexes.length; i += 1) {
const regEx = bannedRegexes[i];
const match = removePunctuationFromString(normalizeUnicodeString(str)).match(regEx);
if (match !== null && match[0] !== null) {
return true;
}
@@ -33,17 +43,21 @@ function usernameContainsInvalidCharacters (username) {
export function verifyDisplayName (displayName, res) {
const issues = [];
if (displayName.length < 1 || displayName.length > 30) issues.push(res.t('displaynameIssueLength'));
if (nameContainsSlur(displayName)) issues.push(res.t('displaynameIssueSlur'));
if (stringContainsProfanity(displayName)) issues.push(res.t('bannedWordUsedInProfile'));
if (stringContainsProfanity(displayName, 'slur')) issues.push(res.t('bannedSlurUsedInProfile'));
if (nameContainsNewline(displayName)) issues.push(res.t('displaynameIssueNewline'));
return issues;
}
export function verifyUsername (username, res) {
export function verifyUsername (username, res, newUser = true) {
const slurMessage = newUser
? res.t('usernameIssueSlur')
: res.t('bannedSlurUsedInProfile');
const issues = [];
if (username.length < 1 || username.length > 20) issues.push(res.t('usernameIssueLength'));
if (usernameContainsInvalidCharacters(username)) issues.push(res.t('usernameIssueInvalidCharacters'));
if (nameContainsSlur(username)) issues.push(res.t('usernameIssueSlur'));
if (stringContainsProfanity(username, 'slur') || stringContainsProfanity(username)) issues.push(slurMessage);
if (usernameIsForbidden(username)) issues.push(res.t('usernameIssueForbidden'));
return issues;