fix(technical): snap missing fib start_date to the date nearest start_date

[Bortlesboat]

Why

calculate_fib_levels in openbb_technical/helpers.py handles a start_date that isn't in the index by snapping to the nearest available date — but the lookup searches for end_date instead of start_date (a copy-paste from the end-date branch right below it):

if start_date not in data.index:
    date0 = data.index[data.index.get_indexer([end_date], method="nearest")[0]]

So any start date that isn't an exact index entry (a weekend or holiday — easy to hit since the index is trading days) silently collapses the retracement to the end date. Reproduction on a 60-business-day frame with a Saturday start_date:

min_date: 2024-03-01  max_date: 2024-03-01  span: 0.0
WARN: Start date not in data.  Using nearest: 2024-03-01

Every Fibonacci level comes back equal to the same price, and the warning reports the wrong "nearest" date, so the result looks plausible unless you inspect it.

What

One-line fix: look up start_date in the get_indexer call. With the fix, the same input snaps to Friday 2024-01-05 and the retracement spans the full range.

Added a regression test (test_calculate_fib_levels_snaps_missing_start_date_to_nearest) that pins the snapped date, the preserved end date, and a non-zero span; it fails on current develop and passes with the fix. All 10 tests in test_technical_helpers.py pass; black and ruff clean.

[deeleeramone]

Hi @Bortlesboat, thanks for the PR. Can you please check the #7489 to verify if the issue still exists there? The technical extension has been refactored, so lots of the code has changed upstream. Thanks!

[Bortlesboat]

Hi @deeleeramone, thanks for the pointer — just re-checked against current develop (post-#7489).

The bug is still there: calculate_fib_levels stayed in openbb_technical/helpers.py through the refactor, and the start-date branch still does the nearest lookup with end_date:

https://github.com/OpenBB-finance/OpenBB/blob/develop/openbb_platform/extensions/technical/openbb_technical/helpers.py#L569-L571

So any start_date that isn't an exact index entry (weekend/holiday) still collapses the retracement to the end date and every fib level comes back equal. I've updated the branch to current develop — the one-line fix and the regression test apply cleanly.