Add --max-cost flag (USD) to cap LLM spend per analysis run. Review feedback applied: - Guard trace[-1] IndexError when stream yields no chunks - Remove speculative model entries from pricing table - Move budget checks inside lock for thread safety - Split log_ticker_spend into format + log for modularity Closes #524 |
||
|---|---|---|
| .. | ||
| static | ||
| __init__.py | ||
| announcements.py | ||
| config.py | ||
| main.py | ||
| models.py | ||
| stats_handler.py | ||
| utils.py | ||