Clarify 3 output types in system prompt
Browse filesAdd clear OUTPUT TYPES section at top of system prompt:
1. PLOTS: answer = filename (after saving plot)
2. TEXT: answer = "direct string result"
3. DATAFRAMES: answer = result_df (for rankings/lists)
Makes it crystal clear how to structure the answer variable for different question types.
π€ Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <[email protected]>
- new_system_prompt.txt +5 -0
new_system_prompt.txt
CHANGED
@@ -2,6 +2,11 @@ Generate Python code to answer the user's question about air quality data.
|
|
2 |
|
3 |
CRITICAL: Only generate Python code - no explanations, no thinking, just clean executable code.
|
4 |
|
|
|
|
|
|
|
|
|
|
|
5 |
AVAILABLE LIBRARIES:
|
6 |
- pandas, numpy (data manipulation)
|
7 |
- matplotlib, seaborn, plotly (visualization)
|
|
|
2 |
|
3 |
CRITICAL: Only generate Python code - no explanations, no thinking, just clean executable code.
|
4 |
|
5 |
+
OUTPUT TYPES (store result in 'answer' variable):
|
6 |
+
1. PLOTS: For visualization questions β save plot and store filename: answer = filename
|
7 |
+
2. TEXT: For simple questions β store direct string: answer = "The highest PM2.5 city is Delhi"
|
8 |
+
3. DATAFRAMES: For rankings/lists β store DataFrame: answer = result_df
|
9 |
+
|
10 |
AVAILABLE LIBRARIES:
|
11 |
- pandas, numpy (data manipulation)
|
12 |
- matplotlib, seaborn, plotly (visualization)
|