Files
2026-10-25 08:19:44 +00:00

102 lines
3.9 KiB
TypeScript

export interface FAQItem {
question: string;
answer: string;
}
/**
* Extracts question and answer pairs from rich HTML content or structured blog data.
* Supports HTML5 <details><summary>, CSS accordion/faq classes, and question headings with paragraph answers.
*/
export function extractFaqsFromHtml(html: string): FAQItem[] {
if (!html || typeof html !== 'string') return [];
const faqs: FAQItem[] = [];
// 1. HTML5 <details><summary>Question</summary>Answer...</details>
const detailsRegex = /<details[^>]*>([\s\S]*?)<\/details>/gi;
let detailsMatch;
while ((detailsMatch = detailsRegex.exec(html)) !== null) {
const block = detailsMatch[1];
const summaryMatch = /<summary[^>]*>([\s\S]*?)<\/summary>/i.exec(block);
if (summaryMatch) {
const question = summaryMatch[1].replace(/<[^>]+>/g, '').trim();
const answerHtml = block.replace(/<summary[^>]*>[\s\S]*?<\/summary>/i, '');
const answer = answerHtml.replace(/<[^>]+>/g, ' ').replace(/\s+/g, ' ').trim();
if (question && answer && !faqs.some((f) => f.question === question)) {
faqs.push({ question, answer });
}
}
}
// 2. Class-based Accordion or FAQ items:
// e.g. <div class="accordion-item"><h3 class="accordion-title">Q</h3><div class="accordion-content">A</div></div>
// Use regex looking forward to next item or end of string
const itemSplitRegex = /<(?:div|article|section|li)[^>]*class=["'][^"']*(?:faq-item|accordion-item|faq-box|qa-item)[^"']*["'][^>]*>/gi;
const itemIndices: number[] = [];
let itemMatch;
while ((itemMatch = itemSplitRegex.exec(html)) !== null) {
itemIndices.push(itemMatch.index);
}
for (let i = 0; i < itemIndices.length; i++) {
const start = itemIndices[i];
const end = i + 1 < itemIndices.length ? itemIndices[i + 1] : html.length;
const block = html.substring(start, end);
const qMatch =
/<[^>]*class=["'][^"']*(?:question|faq-q|accordion-title|accordion-header|faq-header)[^"']*["'][^>]*>([\s\S]*?)<\/[^>]+>/i.exec(
block
);
const aMatch =
/<[^>]*class=["'][^"']*(?:answer|faq-a|accordion-content|accordion-body|faq-body)[^"']*["'][^>]*>([\s\S]*?)<\/[^>]+>/i.exec(
block
);
if (qMatch && aMatch) {
const question = qMatch[1].replace(/<[^>]+>/g, '').trim();
const answer = aMatch[1].replace(/<[^>]+>/g, ' ').replace(/\s+/g, ' ').trim();
if (question && answer && !faqs.some((f) => f.question === question)) {
faqs.push({ question, answer });
}
}
}
// 3. Headings with questions followed immediately by paragraph answers:
// e.g. <h2>پرسش: آیا این مکمل برای سگ مفید است؟</h2> <p>بله، کاملاً مفید است.</p>
const headingRegex = /<h([2-4])\b[^>]*>((?:(?!<\/?h[1-6]\b)[\s\S])*?)<\/h\1>\s*<p[^>]*>([\s\S]*?)<\/p>/gi;
let headingMatch;
while ((headingMatch = headingRegex.exec(html)) !== null) {
const rawHeading = headingMatch[2].replace(/<[^>]+>/g, '').trim();
const rawParagraph = headingMatch[3].replace(/<[^>]+>/g, ' ').replace(/\s+/g, ' ').trim();
if (
(rawHeading.endsWith('?') ||
rawHeading.endsWith('؟') ||
/^(?:پرسش|سوال|سؤال|آیا|چگونه|چرا|چه|کدام)\s*[:؛-]?/i.test(rawHeading)) &&
rawParagraph.length > 5
) {
if (!faqs.some((f) => f.question === rawHeading)) {
faqs.push({ question: rawHeading, answer: rawParagraph });
}
}
}
return faqs;
}
/**
* Builds a Schema.org FAQPage structured data entity if valid FAQs exist.
*/
export function generateFAQPageSchema(faqs: FAQItem[]) {
if (!faqs || faqs.length === 0) return null;
return {
'@context': 'https://schema.org',
'@type': 'FAQPage',
mainEntity: faqs.map((faq) => ({
'@type': 'Question',
name: faq.question,
acceptedAnswer: {
'@type': 'Answer',
text: faq.answer,
},
})),
};
}