Add critical numpy indexing fixes to system prompt

- Add dedicated section for NUMPY INDEXING FIXES with clear examples
- Emphasize ALWAYS converting pandas/numpy values to int before using as indices
- Provide specific example: calendar.month_name[int(month_value)]
- Add guidance for int() conversion in various scenarios: int(row['column']), int(max_idx)
- This should fix the recurring 'numpy.float64 cannot be used as index' errors
- Make the guidance more prominent and specific

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <[email protected]>

src.py

@@ -306,11 +306,16 @@ SAFETY & ROBUSTNESS RULES:
 - Validate date ranges exist in data
 - Use proper string formatting for answers with units (μg/m³)
 
+CRITICAL: NUMPY INDEXING FIXES:
+- ALWAYS convert pandas/numpy values to int before using as list indices
+- Example: calendar.month_name[int(month_value)] NOT calendar.month_name[month_value]
+- Use int() conversion for ANY value used as index: int(row['month']), int(max_idx), etc.
+- When accessing pandas iloc results, wrap in int(): int(df.loc[idx, 'column'])
+
 TECHNICAL REQUIREMENTS:
 - Save final result in variable called 'answer'
 - For TEXT: Store the direct answer as a string in 'answer'
 - For PLOTS: Save with unique filename f"plot_{{uuid.uuid4().hex[:8]}}.png" and store filename in 'answer'
-- Convert numpy types to int when using as indices: int(value)
 - Always use .iloc or .loc properly for pandas indexing
 - Close matplotlib figures with plt.close() to prevent memory leaks
 - Use proper column name checks before accessing columns