deckhearth/scripts/bulk-import-scryfall.js

218 lines
7.3 KiB
JavaScript
Raw Permalink Normal View History

/**
* Bulk import MTG cards from Scryfall's Oracle Cards bulk data file.
*
* Downloads the ~168 MB Oracle Cards JSON (one entry per unique card/oracle_id),
* then upserts every card into the `cards` table keyed on `scryfall_id`.
*
* Populates all new bulk-data columns (oracle_id, illustration_id, color_identity,
* keywords, legalities, flavor_text, artist, released_at, layout, edhrec_rank,
* reserved, reprint, finishes) in addition to the original columns.
*
* Usage:
* POSTGRES_URL=<url> node scripts/bulk-import-scryfall.js
*
* Options (env vars):
* BULK_TYPE "oracle_cards" (default) | "default_cards" | "unique_artwork"
* BATCH_SIZE rows per INSERT batch (default 100)
* DRY_RUN "true" to fetch and count without writing to DB
*/
import { neon } from '@neondatabase/serverless';
if (!process.env.POSTGRES_URL) {
console.error('POSTGRES_URL is required');
process.exit(1);
}
const sql = neon(process.env.POSTGRES_URL, { fullResults: false });
const BULK_TYPE = process.env.BULK_TYPE || 'oracle_cards';
const BATCH_SIZE = parseInt(process.env.BATCH_SIZE || '100', 10);
const DRY_RUN = process.env.DRY_RUN === 'true';
async function fetchBulkDownloadUrl(type) {
const response = await fetch(`https://api.scryfall.com/bulk-data/${type}`, {
headers: { 'User-Agent': 'DeckHearth/1.0', Accept: 'application/json' },
});
if (!response.ok) {
throw new Error(`Failed to fetch bulk-data metadata: ${response.status}`);
}
const data = await response.json();
console.log(`[bulk-import] File: ${data.name}`);
console.log(`[bulk-import] Updated: ${data.updated_at}`);
console.log(`[bulk-import] Size: ${(data.size / 1024 / 1024).toFixed(1)} MB`);
return data.download_uri;
}
function mapCard(card) {
let currentPrice = null;
let marketPrice = null;
if (card.prices) {
currentPrice = card.prices.usd ? parseFloat(card.prices.usd) : null;
marketPrice = card.prices.usd_foil ? parseFloat(card.prices.usd_foil) : null;
}
return {
scryfallId: card.id,
oracleId: card.oracle_id || null,
illustrationId: card.illustration_id || null,
name: card.name,
setName: card.set_name,
setCode: card.set,
cardNumber: card.collector_number,
rarity: card.rarity,
manaCost: card.mana_cost || null,
cmc: card.cmc != null ? Math.round(card.cmc) : null,
cardType: card.type_line || null,
colors: JSON.stringify(card.colors || []),
colorIdentity: JSON.stringify(card.color_identity || []),
oracleText: card.oracle_text || null,
power: card.power || null,
toughness: card.toughness || null,
imageUrl: card.image_uris?.normal || null,
stockImageUrl: card.image_uris?.art_crop || null,
currentPrice,
marketPrice,
keywords: JSON.stringify(card.keywords || []),
legalities: JSON.stringify(card.legalities || {}),
flavorText: card.flavor_text || null,
artist: card.artist || null,
releasedAt: card.released_at || null,
layout: card.layout || null,
edhrecRank: card.edhrec_rank || null,
reserved: card.reserved || false,
reprint: card.reprint || false,
finishes: JSON.stringify(card.finishes || []),
};
}
function buildUpsertQuery(batch) {
const columns = [
'scryfall_id', 'oracle_id', 'illustration_id', 'name', 'set_name', 'set_code',
'card_number', 'rarity', 'game', 'mana_cost', 'cmc', 'card_type', 'colors',
'color_identity', 'oracle_text', 'power', 'toughness', 'image_url', 'stock_image_url',
'current_price', 'market_price', 'keywords', 'legalities', 'flavor_text', 'artist',
'released_at', 'layout', 'edhrec_rank', 'reserved', 'reprint', 'finishes', 'verified',
];
const placeholders = [];
const values = [];
let paramIdx = 1;
for (const row of batch) {
const rowPlaceholders = [];
const rowValues = [
row.scryfallId, row.oracleId, row.illustrationId, row.name, row.setName, row.setCode,
row.cardNumber, row.rarity, 'MTG', row.manaCost, row.cmc, row.cardType, row.colors,
row.colorIdentity, row.oracleText, row.power, row.toughness, row.imageUrl, row.stockImageUrl,
row.currentPrice, row.marketPrice, row.keywords, row.legalities, row.flavorText, row.artist,
row.releasedAt, row.layout, row.edhrecRank, row.reserved, row.reprint, row.finishes, true,
];
for (let i = 0; i < rowValues.length; i++) {
rowPlaceholders.push(`$${paramIdx}`);
paramIdx += 1;
}
placeholders.push(`(${rowPlaceholders.join(', ')})`);
values.push(...rowValues);
}
const updateCols = columns
.filter((c) => c !== 'scryfall_id')
.map((c) => `${c} = EXCLUDED.${c}`)
.join(', ');
const query = `
INSERT INTO cards (${columns.join(', ')})
VALUES ${placeholders.join(',\n')}
ON CONFLICT (scryfall_id) DO UPDATE SET
${updateCols},
updated_at = CURRENT_TIMESTAMP
`;
return { query, values };
}
async function processBatch(batch, stats) {
if (DRY_RUN) {
stats.skipped += batch.length;
return;
}
const { query, values } = buildUpsertQuery(batch);
try {
await sql.query(query, values);
stats.upserted += batch.length;
} catch (error) {
console.error(`[bulk-import] Batch failed (${batch.length} rows):`, error.message);
for (const row of batch) {
try {
const { query: singleQ, values: singleV } = buildUpsertQuery([row]);
await sql.query(singleQ, singleV);
stats.upserted += 1;
} catch (singleErr) {
console.error(`[bulk-import] Failed: ${row.name} (${row.scryfallId}):`, singleErr.message);
stats.errors += 1;
}
}
}
}
async function run() {
console.log(`[bulk-import] Starting Scryfall bulk import (type=${BULK_TYPE}, batch=${BATCH_SIZE}, dryRun=${DRY_RUN})`);
const downloadUrl = await fetchBulkDownloadUrl(BULK_TYPE);
console.log(`[bulk-import] Downloading...`);
const response = await fetch(downloadUrl, {
headers: { 'User-Agent': 'DeckHearth/1.0' },
});
if (!response.ok) {
throw new Error(`Download failed: ${response.status}`);
}
const text = await response.text();
console.log(`[bulk-import] Downloaded ${(text.length / 1024 / 1024).toFixed(1)} MB, parsing...`);
const cards = JSON.parse(text);
console.log(`[bulk-import] Total cards in bulk file: ${cards.length}`);
const mtgCards = cards.filter((c) => c.lang === 'en' && !c.digital);
console.log(`[bulk-import] MTG cards (English, paper): ${mtgCards.length}`);
const stats = { upserted: 0, skipped: 0, errors: 0 };
let batch = [];
let processed = 0;
for (const card of mtgCards) {
batch.push(mapCard(card));
if (batch.length >= BATCH_SIZE) {
await processBatch(batch, stats);
processed += batch.length;
batch = [];
if (processed % 5000 === 0) {
console.log(`[bulk-import] Progress: ${processed}/${mtgCards.length} (${((processed / mtgCards.length) * 100).toFixed(1)}%)`);
}
}
}
if (batch.length > 0) {
await processBatch(batch, stats);
processed += batch.length;
}
console.log(`[bulk-import] Complete!`);
console.log(`[bulk-import] Upserted: ${stats.upserted}`);
console.log(`[bulk-import] Skipped (dry run): ${stats.skipped}`);
console.log(`[bulk-import] Errors: ${stats.errors}`);
console.log(`[bulk-import] Total processed: ${processed}`);
}
run().catch((err) => {
console.error('[bulk-import] Fatal error:', err);
process.exit(1);
});