fix: collect all token amounts in portfolio_volatility instead of using last token only

[bigeez]

Bug

In onchain/analytics/analytics_tools.py, the portfolio_volatility
function iterates over all tokens to fetch price data but only uses
the last token's amount for holding_qty:

for token in tokens:
    ...
    all_price_data.append(close_prices)

holding_qty = np.array(token.amount)  # only last token!

After the loop, token refers to the last item only. This means
portfolio volatility is calculated with incorrect weights for any
wallet holding more than one token — silently producing wrong results.

Fix

Collect each token's amount inside the loop into all_holding_qty,
then build the numpy array from that list after the loop completes.

Impact

• Any user with 2+ tokens in their wallet was getting wrong volatility numbers
• Silent bug — no error thrown, just incorrect calculations