python-batch.py

"""Vinted Smart Scraper — Python batch example.

Runs 26 EU country scrapes in parallel using the Apify client async API,
deduplicates items by `vintedId`, ranks by arbitrage spread, and writes
the top 100 to a CSV.

Setup:
    pip install apify-client pandas
    export APIFY_TOKEN="<your-token>"
    python examples/python-batch.py "vintage carhartt"

Cost: ~$0.018 × 26 starts + $0.0005 × ~5000 results ≈ $2.97 per full sweep.
"""

from __future__ import annotations

import asyncio
import csv
import os
import sys
from pathlib import Path

from apify_client import ApifyClientAsync

EU_COUNTRIES = [
    "fr", "es", "it", "de", "be", "lu", "nl", "pt", "at",
    "ie", "fi", "se", "dk", "lt", "lv", "ee", "cz", "sk",
    "pl", "hu", "ro", "bg", "hr", "si", "gr", "uk",
]


async def scrape_country(client: ApifyClientAsync, country: str, query: str) -> list[dict]:
    """Run the actor for a single country, return its items."""
    try:
        run = await client.actor("kazkn/vinted-smart-scraper").call(
            run_input={
                "searchText": query,
                "countries": [country],
                "maxResults": 200,
                "currency": "EUR",
            }
        )
        items = []
        async for item in client.dataset(run["defaultDatasetId"]).iterate_items():
            items.append(item)
        print(f"  {country}: {len(items)} items")
        return items
    except Exception as e:
        print(f"  {country}: FAILED ({e})")
        return []


async def main(query: str) -> None:
    token = os.environ.get("APIFY_TOKEN")
    if not token:
        sys.exit("Set APIFY_TOKEN env var first")

    client = ApifyClientAsync(token)

    print(f"Sweeping {len(EU_COUNTRIES)} countries for: {query}")
    print("This will take ~30-60 seconds (parallel, capped by your Apify concurrency).\n")

    all_items_per_country = await asyncio.gather(
        *(scrape_country(client, c, query) for c in EU_COUNTRIES)
    )

    # Dedup by vintedId, keep cheapest occurrence
    by_id: dict[str, dict] = {}
    for items in all_items_per_country:
        for item in items:
            vid = str(item.get("vintedId") or item.get("id"))
            if vid not in by_id or item["priceEur"] < by_id[vid]["priceEur"]:
                by_id[vid] = item

    deduped = list(by_id.values())
    print(f"\nDeduped: {sum(len(x) for x in all_items_per_country)} → {len(deduped)} items")

    # Compute spread per item: max country price / min country price - 1
    # Build a per-item country price map
    by_title: dict[str, list[dict]] = {}
    for item in deduped:
        key = (item.get("title") or "").strip().lower()
        if not key:
            continue
        by_title.setdefault(key, []).append(item)

    ranked = []
    for title, group in by_title.items():
        if len(group) < 2:
            continue
        prices = sorted(group, key=lambda i: i["priceEur"])
        cheapest, dearest = prices[0], prices[-1]
        spread = (dearest["priceEur"] - cheapest["priceEur"]) / cheapest["priceEur"]
        if spread <= 0:
            continue
        ranked.append({
            "title": title,
            "buy_country": cheapest["country"],
            "buy_price_eur": round(cheapest["priceEur"], 2),
            "sell_country": dearest["country"],
            "sell_price_eur": round(dearest["priceEur"], 2),
            "spread_pct": round(spread * 100, 1),
            "buy_url": cheapest.get("permalink"),
            "sell_url": dearest.get("permalink"),
        })

    ranked.sort(key=lambda r: r["spread_pct"], reverse=True)
    top = ranked[:100]

    out = Path(f"vinted-arbitrage-{query.replace(' ', '_')}.csv")
    with out.open("w", newline="") as f:
        writer = csv.DictWriter(f, fieldnames=top[0].keys() if top else ["title"])
        writer.writeheader()
        writer.writerows(top)

    print(f"\nTop 100 spreads written to {out}")
    if top:
        print(f"\n#1 spread:")
        for k, v in top[0].items():
            print(f"  {k}: {v}")


if __name__ == "__main__":
    query = sys.argv[1] if len(sys.argv) > 1 else "vintage carhartt"
    asyncio.run(main(query))