added cerebras

together_ai_llama_agent.py

@@ -55,6 +55,40 @@ def get_csv_info(df: pd.DataFrame) -> dict:
     return info
 
 
+# def get_csv_system_prompt(df: pd.DataFrame) -> str:
+#     """Generate system prompt for CSV analysis"""
+#     csv_info = get_csv_info(df)
+    
+#     prompt = f"""
+# You're a CSV analysis assistant. The pandas DataFrame is loaded as 'df' - use this variable.
+
+# CSV Info:
+# - Rows: {csv_info['num_rows']}, Cols: {csv_info['num_cols']}
+# - Columns: {csv_info['columns']}
+# - Sample: {csv_info['example_rows']}
+# - Dtypes: {csv_info['dtypes']}
+
+# Strict Rules:
+# 1. Never recreate 'df' - use the existing variable
+# 2. For analysis:
+#    - Include necessary imports (except pandas) and include complete code
+#    - Use df directly (e.g., print(df[...].mean()))
+# 3. For visualizations:
+#    - Adjust font sizes, rotate labels (45° if needed), truncate for readability
+#    - Figure size: (12, 6)
+#    - Descriptive titles (fontsize=14)
+#    - Colorblind-friendly palettes
+#    - Do not use any visualization library other than matplotlib or seaborn
+#    - Complete code with plt.show()
+#    - Example: plt.bar(df['x'], df['y']) \n   plt.show()
+# 4. For Lists and Dictionaries, return them as JSON
+
+# Example:
+# import json
+# print(json.dumps(df[df['col'] == 'val'].to_dict('records'), indent=2))
+# """
+#     return 
+
 def get_csv_system_prompt(df: pd.DataFrame) -> str:
     """Generate system prompt for CSV analysis"""
     csv_info = get_csv_info(df)
@@ -74,13 +108,38 @@ Strict Rules:
    - Include necessary imports (except pandas) and include complete code
    - Use df directly (e.g., print(df[...].mean()))
 3. For visualizations:
-   - Adjust font sizes, rotate labels (45° if needed), truncate for readability
-   - Figure size: (12, 6)
-   - Descriptive titles (fontsize=14)
-   - Colorblind-friendly palettes
+   - Create the most professional, publication-quality charts possible
+   - Maximize descriptive elements and detail while maintaining clarity
+   - Figure size: (14, 8) for complex charts, (12, 6) for simpler ones
+   - Use comprehensive titles (fontsize=16) and axis labels (fontsize=14)
+   - Include informative legends (fontsize=12) when appropriate
+   - Add annotations for important data points where valuable
+   - Rotate x-labels (45° if needed) with fontsize=12 for readability
+   - Use colorblind-friendly palettes (seaborn 'deep', 'muted', or 'colorblind')
+   - Add gridlines (alpha=0.3) when they improve readability
+   - Include proper margins and padding to prevent label cutoff
+   - For distributions, include kernel density estimates when appropriate
+   - For time series, use appropriate date formatting and markers
    - Do not use any visualization library other than matplotlib or seaborn
-   - Complete code with plt.show()
-   - Example: plt.bar(df['x'], df['y']) \n   plt.show()
+   - Complete code with plt.tight_layout() before plt.show()
+   - Example professional chart:
+     plt.figure(figsize=(14, 8))
+     ax = sns.barplot(x='category', y='value', data=df, palette='muted', ci=None)
+     plt.title('Detailed Analysis of Values by Category', fontsize=16, pad=20)
+     plt.xlabel('Category', fontsize=14)
+     plt.ylabel('Average Value', fontsize=14)
+     plt.xticks(rotation=45, ha='right', fontsize=12)
+     plt.yticks(fontsize=12)
+     ax.grid(True, linestyle='--', alpha=0.3)
+     for p in ax.patches:
+         ax.annotate(f'{{p.get_height():.1f}}', 
+                    (p.get_x() + p.get_width() / 2., p.get_height()), 
+                    ha='center', va='center', 
+                    xytext=(0, 10), 
+                    textcoords='offset points',
+                    fontsize=12)
+     plt.tight_layout()
+     plt.show()
 4. For Lists and Dictionaries, return them as JSON
 
 Example: