Clean up context with matplotlib style file + mandatory answer fix

CONTEXT CLEANUP:
- Create vayuchat.mplstyle file with professional styling
- Replace 21 lines of matplotlib styling with 1 line: plt.style.use('vayuchat.mplstyle')
- Dramatically reduces LLM context clutter for better code generation

SYSTEM PROMPT IMPROVEMENTS:
- Add MANDATORY answer variable section to prevent "no result saved" errors
- Move model selector to sidebar for better UX (no more scrolling to top)
- Add import requirements and plotting best practices

MAJOR CONTEXT REDUCTION:
- 20+ lines of styling boilerplate eliminated
- Cleaner, more focused code templates for LLM
- Professional consistent theming across all plots

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

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

new_system_prompt.txt

@@ -13,6 +13,10 @@ AVAILABLE LIBRARIES:
 - statsmodels, scikit-learn (analysis)
 - geopandas (geospatial analysis)
 
+IMPORT REQUIREMENTS:
+- Always import what you use: import seaborn as sns, import numpy as np
+- Standard imports are already available: pandas as pd, matplotlib.pyplot as plt
+
 ESSENTIAL RULES:
 
 DATA SAFETY:
@@ -34,8 +38,21 @@ BASIC ERROR PREVENTION:
 - For correlations: check len(data) > 20 before calculating
 - Use simple matplotlib plotting - avoid complex visualizations
 
+PLOTTING BEST PRACTICES:
+- Check data exists in each category before plotting
+- For comparisons (>, <): ensure both categories have data
+- Example: high_wind = df[df['WS'] > 3]; low_wind = df[df['WS'] <= 3]
+- If category is empty: create simple bar chart instead of box plots
+- Add data count labels: plt.text() to show sample sizes
+
 TECHNICAL REQUIREMENTS:
 - Save final result in variable called 'answer'
 - Use exact column names: 'PM2.5 (µg/m³)', 'WS (m/s)', etc.
 - Handle dates with pd.to_datetime() if needed
-- Round numerical results: round(value, 2)
+- Round numerical results: round(value, 2)
+
+MANDATORY: ALWAYS END CODE WITH ANSWER ASSIGNMENT
+- Every code block MUST end with: answer = [result]
+- If analysis fails: answer = "Unable to complete analysis with available data"
+- If plotting fails: answer = "Unable to generate visualization"
+- NEVER leave answer variable unset - this will cause system failure

src.py

@@ -252,28 +252,7 @@ import uuid
 import calendar
 import numpy as np
 # Set professional matplotlib styling
-plt.rcParams.update({{
-    'font.size': 12,
-    'figure.dpi': 400,
-    'figure.facecolor': 'white',
-    'axes.facecolor': 'white',
-    'axes.edgecolor': '#e2e8f0',
-    'axes.linewidth': 1.2,
-    'axes.labelcolor': '#374151',
-    'axes.spines.top': False,
-    'axes.spines.right': False,
-    'axes.spines.left': True,
-    'axes.spines.bottom': True,
-    'axes.grid': True,
-    'grid.color': '#f1f5f9',
-    'grid.linewidth': 0.8,
-    'grid.alpha': 0.7,
-    'xtick.color': '#6b7280',
-    'ytick.color': '#6b7280',
-    'text.color': '#374151',
-    'figure.figsize': [12, 6],
-    'axes.prop_cycle': plt.cycler('color', ['#3b82f6', '#ef4444', '#10b981', '#f59e0b', '#8b5cf6', '#06b6d4'])
-}})
+plt.style.use('vayuchat.mplstyle')
 df = pd.read_csv("AQ_met_data.csv")
 df["Timestamp"] = pd.to_datetime(df["Timestamp"])
 states_df = pd.read_csv("states_data.csv")

vayuchat.mplstyle

@@ -0,0 +1,21 @@
+# VayuChat Professional Style
+font.size: 12
+figure.dpi: 400
+figure.facecolor: white
+axes.facecolor: white
+axes.edgecolor: #e2e8f0
+axes.linewidth: 1.2
+axes.labelcolor: #374151
+axes.spines.top: False
+axes.spines.right: False
+axes.spines.left: True
+axes.spines.bottom: True
+axes.grid: True
+grid.color: #f1f5f9
+grid.linewidth: 0.8
+grid.alpha: 0.7
+xtick.color: #6b7280
+ytick.color: #6b7280
+text.color: #374151
+figure.figsize: 12, 6
+axes.prop_cycle: cycler('color', ['3b82f6', 'ef4444', '10b981', 'f59e0b', '8b5cf6', '06b6d4'])