Files
rmtPocketWatcher/backend/src/scrapers/eldorado-scraper.ts
2025-12-03 18:00:10 -05:00

152 lines
5.1 KiB
TypeScript

import { Page } from 'playwright';
import { BaseScraper } from './base-scraper';
import { VendorListing } from './types';
export class EldoradoScraper extends BaseScraper {
getVendorName(): 'eldorado' {
return 'eldorado';
}
getTargetUrl(): string {
return 'https://www.eldorado.gg/star-citizen-auec/g/141-0-0';
}
async extractListings(page: Page): Promise<VendorListing[]> {
// Wait for page readiness
await page.waitForSelector('text=Star Citizen aUEC', { timeout: 15000 }).catch(() => {});
// Wait for price elements to appear
await page.waitForTimeout(3000);
const listings = await page.evaluate(() => {
const results: Array<{
amountAUEC: number;
priceUSD: number;
pricePerMillion: number;
seller?: string;
deliveryTime?: string;
}> = [];
const bodyText = document.body.innerText;
const lines = bodyText.split('\n').map(l => l.trim()).filter(l => l.length > 0);
// Track seen combinations to avoid duplicates
const seenListings = new Set<string>();
for (let i = 0; i < lines.length; i++) {
const line = lines[i];
// Look for "$/M" pattern - this is the direct price per million
// Examples: "$0.00007 / M", "$0.00018 / M", "$0.00007/M"
const pricePerMMatch = line.match(/\$\s*([\d.]+)\s*\/?\s*\/\s*M/i) || line.match(/\$\s*([\d.]+)\s*\/\s*M/i);
if (pricePerMMatch) {
const pricePerMillion = parseFloat(pricePerMMatch[1]);
// Look for "Min. qty" or "Min qty" nearby to get the quantity
let minQtyM = 10000; // Default to 10000M
for (let j = Math.max(0, i - 5); j < Math.min(lines.length, i + 5); j++) {
const qtyLine = lines[j];
// Match patterns like "Min. qty. 6000 M" or "Min qty: 16,000 M"
const qtyMatch = qtyLine.match(/Min\.?\s*qty\.?\s*:?\s*([\d,]+)\s*M/i);
if (qtyMatch) {
minQtyM = parseFloat(qtyMatch[1].replace(/,/g, ''));
break;
}
}
const amountAUEC = minQtyM * 1_000_000;
const priceUSD = pricePerMillion * minQtyM;
// Find seller name - look both backwards and forwards
// For featured seller, name appears BEFORE the price
// For other sellers, name appears in a structured list
let seller: string | undefined;
// Search backwards first (for featured seller and some list items)
for (let j = Math.max(0, i - 20); j < i; j++) {
const sellerLine = lines[j];
// Skip common non-seller text
if (
sellerLine.includes('$') ||
sellerLine.includes('Price') ||
sellerLine.includes('qty') ||
sellerLine.includes('stock') ||
sellerLine.includes('Delivery') ||
sellerLine.toLowerCase().includes('review') ||
sellerLine.includes('Rating') ||
sellerLine.includes('Offer') ||
sellerLine.includes('Details') ||
sellerLine.includes('FEATURED') ||
sellerLine.includes('Other') ||
sellerLine.includes('sellers') ||
sellerLine.includes('aUEC') ||
sellerLine === 'Star Citizen' || // Exclude the game title but not seller names
sellerLine.includes('IMF') ||
sellerLine.includes('in-game') ||
sellerLine.includes('currency') ||
sellerLine.length < 3 ||
sellerLine.length > 30
) {
continue;
}
// Match seller name patterns - alphanumeric with underscores/hyphens
// Allow some special cases like "StarCitizen"
if (/^[a-zA-Z][a-zA-Z0-9_-]*$/.test(sellerLine)) {
seller = sellerLine;
// Don't break - keep looking for a closer match
}
}
// Find delivery time
let deliveryTime: string | undefined;
for (let j = Math.max(0, i - 5); j < Math.min(lines.length, i + 5); j++) {
const deliveryLine = lines[j];
if (
deliveryLine.match(/\d+\s*min/i) ||
deliveryLine.match(/\d+\s*hour/i) ||
deliveryLine.toLowerCase().includes('instant')
) {
deliveryTime = deliveryLine;
break;
}
}
// Create unique key to avoid duplicates
const key = `${pricePerMillion}-${minQtyM}`;
if (seenListings.has(key)) continue;
seenListings.add(key);
results.push({
amountAUEC,
priceUSD,
pricePerMillion,
seller,
deliveryTime,
});
}
}
return results;
});
const scrapedAt = new Date();
const url = this.getTargetUrl();
return listings.map(listing => ({
vendor: 'eldorado' as const,
amountAUEC: listing.amountAUEC,
priceUSD: listing.priceUSD,
pricePerMillion: listing.pricePerMillion,
seller: listing.seller,
deliveryTime: listing.deliveryTime,
scrapedAt,
url,
}));
}
}