ISIN meaning

1. Normalize input → ISIN.
   • If user gives ticker: resolve via OpenFIGI API (free, no key for low volume) or scrape the justETF search.
   • If user gives ISIN: skip ahead.
2. ISIN → issuer. First two chars of issuer name in the ETF title tell you 90% of cases ("iShares" → BlackRock, "Xtrackers" → DWS, "Amundi", "SPDR" → State Street, "Invesco", "WisdomTree", "VanEck", "Vanguard"). Get the name from justETF's profile page (one HTTP GET, parse <h1>) or OpenFIGI's name field.
3. Issuer → product ID → holdings file. Each issuer has a stable URL pattern. The high-value ones cover ~80% of European ETF AUM (see Find your ETF markets for the issuer breakdown). The only step that needs scraping is ISIN → productId for iShares (their CSV URL uses an internal numeric ID like 319084, not the ISIN). One regex on the product page HTML handles it.
4. Parse + normalize. Each issuer's CSV has different columns. Map to a common schema: {ticker, name, isin, weight_pct, market_value, sector, country, asset_class}. Strip the disclaimer header rows iShares prepends.

resolve_isin(input)          # OpenFIGI or justETF search
  → fetch_etf_meta(isin)     # justETF page → {name, issuer}
  → issuer_adapter(issuer)   # dispatch to one of ~8 adapters
  → adapter.fetch(isin)      # returns raw CSV/XLSX bytes
  → adapter.parse(bytes)     # returns normalized rows

One adapter per issuer, behind a common interface. New issuer = new adapter file, no other changes.

1. Issuer's official CSV/XLSX (above).
2. Issuer's JSON API if discovered in DevTools (iShares and Vanguard both have one).
3. TradingView holdings page scrape (partial but better than nothing).
4. justETF top-10 (last resort, clearly mark as incomplete).