Create state-figures-pipeline-generator.py

pipeline/state-figures-pipeline-generator.py

@@ -0,0 +1,46 @@
+import matplotlib.pyplot as plt
+import pandas as pd
+import os
+
+dicdataset = {
+    "5global":r"5% of data globally delayed",
+    "20median":r"20% of the keyes around the median are delayed",
+    "50global":r"50% of the data globally delayed",
+    "50median":r"50% of the keyes around the median are delayed",
+    "20random":r"20% of randomly chosen keyes are delayed",
+    "50random":r"50% of randomly chosen keyes are delayed",
+    "debs":r"DEBS Dataset",
+    "ordered":r"Fully Ordered"
+}
+
+# Get a list of all CSV files in the current directory
+csv_files = [file for file in os.listdir('.') if file.endswith('.csv')]
+
+# Loop through the CSV files and plot each one as a figure
+for csv_file in csv_files:
+    df = pd.read_csv(csv_file)
+    csv_file = csv_file.replace('_state_result.csv.csv', '')
+
+    # Sort the DataFrame by 'TS' column
+    df = df.sort_values(by='TS')
+    # Filter the DataFrame to keep rows with TS <= 700000
+    if csv_file != 'debs':
+        df = df[df['TS'] <= 700000]
+    print(df)
+
+    # Calculate the difference between State_Size_Flink and State_Size_Keyed
+    df['State_Size_Difference'] = df['State_Size_Keyed'] - df['State_Size_Flink']
+
+    # Create a figure for the line plot
+    plt.figure(figsize=(12, 6))
+
+    plt.plot(df['TS'], df['State_Size_Difference'], color='green', label='Above zero means Keyed state is higher, below zero means Flink state is higher')
+    # plt.grid(True)
+    plt.xlabel('Timestamps')
+    plt.ylabel('State Size Difference')
+    # plt.title(f'State Size Difference (Flink Vs. Keyed) - Dataset: {dicdataset[csv_file]}')
+    plt.legend()
+    # Show the plot
+    # plt.show()
+    # Save the figure to a file
+    plt.savefig(f'{csv_file}_graph.png', bbox_inches='tight')