feat(search-pagination): implement normalized weighted search and standardized pagination across blogs, products and wiki
Deploy Canina / deploy (push) Successful in 4m25s

This commit is contained in:
Canina Automation committed 2026-10-27 11:39:55 +00:00
1 parent d3a0a61d79
commit 66b9d75524
7 files changed
+469 -57

No files matched your search

@@ -0,0 +1,91 @@
import {
normalizeSearchText,
extractSearchTokens,
calculateRelevanceScore,
createPaginationMeta,
} from "./search.utils";
describe("search.utils", () => {
describe("createPaginationMeta", () => {
it("builds consistent pagination metadata", () => {
const meta = createPaginationMeta(52, 2, 10);
expect(meta).toEqual({
total: 52,
page: 2,
limit: 10,
totalPages: 6,
lastPage: 6,
hasNextPage: true,
hasPrevPage: true,
});
});
it("handles boundary zero and single page counts safely", () => {
const metaZero = createPaginationMeta(0, 1, 10);
expect(metaZero.totalPages).toBe(1);
expect(metaZero.hasNextPage).toBe(false);
expect(metaZero.hasPrevPage).toBe(false);
const metaOne = createPaginationMeta(5, 1, 10);
expect(metaOne.totalPages).toBe(1);
expect(metaOne.hasNextPage).toBe(false);
expect(metaOne.hasPrevPage).toBe(false);
});
});
describe("normalizeSearchText", () => {
it("normalizes Arabic kaf and yeh to Persian equivalents", () => {
expect(normalizeSearchText("قطره قلب كنينای ايران")).toBe("قطره قلب کنینای ایران");
});
it("strips parentheses, diacritics, and half-spaces", () => {
expect(normalizeSearchText("ال-کارنیتین (L-Carnitine) \u200c دارویی")).toBe("ال کارنیتین l carnitine دارویی");
expect(normalizeSearchText("مُکَمِّل")).toBe("مکمل");
});
it("converts Persian digits to English digits", () => {
expect(normalizeSearchText("محصول شماره ۱۲۳")).toBe("محصول شماره 123");
});
});
describe("extractSearchTokens", () => {
it("extracts meaningful tokens and removes stop words", () => {
const tokens = extractSearchTokens("بهترین قطره برای تقویت قلب سگ و گربه");
expect(tokens).toContain("قطره");
expect(tokens).toContain("تقویت");
expect(tokens).toContain("قلب");
expect(tokens).toContain("سگ");
expect(tokens).toContain("گربه");
expect(tokens).not.toContain("برای");
expect(tokens).not.toContain("و");
});
});
describe("calculateRelevanceScore", () => {
it("ranks exact title / ingredient matches far higher than general description mentions", () => {
const herzVital = {
title: "هیرتز ویتال (مکمل تقویت قلب) کنینا",
slug: "canina-herz-vital-100ml",
ingredients: [{ nameFa: "ال-کارنیتین (L-Carnitine)", nameEn: "L-Carnitine" }],
symptoms: [{ symptom: "بیماری قلبی" }],
shortDescription: "حمایت تخصصی از عضله قلب سگ و گربه",
description: "توضیحات مفصل درباره کارکرد قلب و عروق...",
};
const genericProduct = {
title: "پودر ویتامین پلاس سگ",
slug: "canina-vitamin-plus",
ingredients: [{ nameFa: "ویتامین A", nameEn: "Vitamin A" }],
symptoms: [],
shortDescription: "تقویت عمومی بدن",
description: "این مولتی ویتامین برای تمام اندام‌های بدن از جمله قلب، کبد و مفاصل مفید است.",
};
const tokens = extractSearchTokens("ال-کارنیتین قلب");
const scoreHerz = calculateRelevanceScore(herzVital, "ال-کارنیتین قلب", tokens);
const scoreGeneric = calculateRelevanceScore(genericProduct, "ال-کارنیتین قلب", tokens);
expect(scoreHerz).toBeGreaterThan(scoreGeneric * 4);
});
});
});
+215
View File
@@ -0,0 +1,215 @@
export interface PaginationMeta {
total: number;
page: number;
limit: number;
totalPages: number;
lastPage: number;
hasNextPage: boolean;
hasPrevPage: boolean;
}
export function createPaginationMeta(
total: number,
page: number,
limit: number,
): PaginationMeta {
const safeTotal = Math.max(0, Number(total) || 0);
const safeLimit = Math.max(1, Number(limit) || 10);
const safePage = Math.max(1, Number(page) || 1);
const totalPages = Math.ceil(safeTotal / safeLimit) || 1;
return {
total: safeTotal,
page: safePage,
limit: safeLimit,
totalPages,
lastPage: totalPages,
hasNextPage: safePage < totalPages,
hasPrevPage: safePage > 1,
};
}
const PERSIAN_ARABIC_DIGITS: Record<string, string> = {
"۰": "0", "۱": "1", "۲": "2", "۳": "3", "۴": "4",
"۵": "5", "۶": "6", "۷": "7", "۸": "8", "۹": "9",
"٠": "0", "١": "1", "٢": "2", "٣": "3", "٤": "4",
"٥": "5", "٦": "6", "٧": "7", "٨": "8", "٩": "9",
};
export const COMMON_STOP_WORDS = new Set([
"از", "به", "در", "با", "برای", "و", "یا", "که", "این", "آن", "ها", "های",
"تر", "ترین", "روی", "توی", "را", "تا", "بر", "اما", "اگر", "ولی", "یک",
"the", "a", "an", "and", "or", "for", "with", "in", "on", "at", "to", "of",
]);
/**
* Normalizes Persian and English text by standardizing characters,
* removing diacritics, half-spaces, punctuation, and folding case.
*/
export function normalizeSearchText(text?: string | null): string {
if (!text) return "";
let s = String(text);
// Convert Persian/Arabic digits
s = s.replace(/[۰-۹٠-٩]/g, (ch) => PERSIAN_ARABIC_DIGITS[ch] || ch);
// Normalize Arabic letters to standard Persian/English
s = s
.replace(/[يى]/g, "ی")
.replace(/[ك]/g, "ک")
.replace(/[ة]/g, "ه")
.replace(/[آأإٱ]/g, "ا")
.replace(/[ؤ]/g, "و")
.replace(/[ئ]/g, "ی");
// Remove Arabic/Persian diacritics (fatha, damma, kasra, tanwin, sukun, tashdid)
s = s.replace(/[\u064B-\u065F\u0670]/g, "");
// Replace half-space, non-breaking space, parentheses, brackets, quotes, punctuation with single space
s = s.replace(/[\u200B-\u200D\uFEFF\u00A0\u200E\u200F_~`!@#$%^&*()+=[\]{}|\\;:'",.<>/?،؛؟]/g, " ");
// Collapse hyphens surrounded by spaces or replace hyphen with space
s = s.replace(/-/g, " ");
// Lowercase & collapse whitespace
return s.toLowerCase().replace(/\s+/g, " ").trim();
}
/**
* Extracts meaningful tokens from a query, eliminating stop words
* unless the entire query only consists of stop words.
*/
export function extractSearchTokens(rawQuery?: string | null): string[] {
const normalized = normalizeSearchText(rawQuery);
if (!normalized) return [];
const rawTokens = normalized.split(" ").filter((t) => t.length > 0);
if (rawTokens.length === 0) return [];
// Filter out stop words
const meaningful = rawTokens.filter(
(t) => !COMMON_STOP_WORDS.has(t) && t.length >= 2,
);
// If filtering left nothing, return original non-empty tokens
const result = meaningful.length > 0 ? meaningful : rawTokens;
return Array.from(new Set(result));
}
export interface SearchScoreItem {
artNo?: string | null;
barcode?: string | null;
slug?: string | null;
title?: string | null;
shortDescription?: string | null;
description?: string | null;
tags?: Array<{ name?: string | null } | string> | null;
ingredients?: Array<
| { ingredient?: string | null; nameFa?: string | null; nameEn?: string | null }
| string
> | null;
symptoms?: Array<{ symptom?: string | null } | string> | null;
}
/**
* Calculates a relevance score for a product or blog item against search tokens.
*/
export function calculateRelevanceScore(
item: SearchScoreItem,
rawQuery: string,
tokens: string[],
): number {
if (!rawQuery || tokens.length === 0) return 0;
const normalizedQuery = normalizeSearchText(rawQuery);
const normalizedTitle = normalizeSearchText(item.title);
const normalizedArtNo = normalizeSearchText(item.artNo);
const normalizedBarcode = normalizeSearchText(item.barcode);
const normalizedSlug = normalizeSearchText(item.slug);
let score = 0;
// Exact matches get massive priority
if (normalizedArtNo && normalizedArtNo === normalizedQuery) score += 200;
if (normalizedBarcode && normalizedBarcode === normalizedQuery) score += 180;
if (normalizedTitle && normalizedTitle === normalizedQuery) score += 150;
if (normalizedSlug && normalizedSlug === normalizedQuery) score += 140;
// Title starts with full query
if (normalizedTitle && normalizedTitle.startsWith(normalizedQuery)) score += 80;
// Title contains exact full query phrase
if (normalizedTitle && normalizedTitle.includes(normalizedQuery)) score += 60;
// Token level matching in title (highest relevance)
const titleTokens = normalizedTitle.split(" ");
for (const token of tokens) {
if (titleTokens.includes(token)) {
score += 35;
} else if (normalizedTitle.includes(token)) {
score += 20;
}
}
// Ingredient / scientific terms matching
if (item.ingredients && Array.isArray(item.ingredients)) {
for (const ing of item.ingredients) {
let ingName = "";
if (typeof ing === "string") {
ingName = ing;
} else if (ing) {
ingName = `${ing.ingredient || ""} ${ing.nameFa || ""} ${ing.nameEn || ""}`.trim();
}
const normIng = normalizeSearchText(ingName);
if (normIng.includes(normalizedQuery)) score += 40;
for (const token of tokens) {
if (normIng.includes(token)) score += 15;
}
}
}
// Symptoms matching
if (item.symptoms && Array.isArray(item.symptoms)) {
for (const sym of item.symptoms) {
const symName = typeof sym === "string" ? sym : sym.symptom || "";
const normSym = normalizeSearchText(symName);
if (normSym.includes(normalizedQuery)) score += 30;
for (const token of tokens) {
if (normSym.includes(token)) score += 10;
}
}
}
// Tags matching
if (item.tags && Array.isArray(item.tags)) {
for (const tag of item.tags) {
const tagName = typeof tag === "string" ? tag : tag.name || "";
const normTag = normalizeSearchText(tagName);
if (normTag.includes(normalizedQuery)) score += 25;
for (const token of tokens) {
if (normTag.includes(token)) score += 8;
}
}
}
// Short description matching
const normShort = normalizeSearchText(item.shortDescription);
if (normShort) {
for (const token of tokens) {
if (normShort.includes(token)) score += 5;
}
}
// Long description (low weight, capped so it never overwhelms title/ingredients)
const normDesc = normalizeSearchText(item.description);
if (normDesc) {
let descHits = 0;
for (const token of tokens) {
if (normDesc.includes(token)) descHits += 1;
}
score += Math.min(descHits * 2, 6);
}
return score;
}