create a class to automatize the charts generation for the LOD Cloud …

…evaluation

lodc_quality_evaluation/generate_charts.py

@@ -0,0 +1,134 @@
+import pandas as pd
+import os
+import seaborn as sns
+import matplotlib.pyplot as plt
+
+class GenerateCharts:
+
+    def __init__(self,evaluation_results_path,charts_output = './charts') -> None:
+        '''
+            Creates a list of CSV files that are to be parsed.
+
+            :param evaluations_results_path: path to the folder that contains the evaluation results csv files.
+            :param charts_output: path to the folder in which to place the generated graphs.
+
+        '''
+        self.analysis_results_files = []
+        self.output_file = charts_output
+        # Get all csv filename from the dir
+        for filename in os.listdir(evaluation_results_path):
+            if '.csv' in filename:
+                file_path = os.path.join(evaluation_results_path, filename)
+                self.analysis_results_files.append(file_path)    
+
+
+    def generate_boxplots_over_time(self, range='M'):
+        '''
+            Generates different boxplot, one for each metric, having the date as the x-axis and each box is measurement.
+
+            :param range: Time data selection interval, e.g., monthly(M), quarterly(Q), all (A).
+        '''
+
+        for file in self.analysis_results_files:
+            metric_analyzed = os.path.splitext(os.path.basename(file))[0]
+
+            df = pd.read_csv(file)
+
+            df['Analysis date'] = pd.to_datetime(df['Analysis date'])
+            df = df.sort_values(by='Analysis date')
+
+            if range != 'A':
+                df_filtered = df.groupby(df['Analysis date'].dt.to_period(range)).first()
+            else:
+                df_filtered = df
+
+            df_melted = df_filtered.melt(id_vars='Analysis date', value_vars=['Min', 'Q1', 'Median', 'Q3', 'Max'], 
+                var_name='Statistic', value_name='Value')
+
+            plt.figure(figsize=(10, 9))
+            sns.boxplot(x='Analysis date', y='Value', hue='Analysis date', data=df_melted)
+            plt.xticks(rotation=45)
+            plt.title(metric_analyzed)
+            plt.xlabel('Date')
+            plt.ylabel('Values')
+            plt.savefig(self.output_file + '/' + metric_analyzed)
+
+    def generate_boxplots_punctual(self,input_file):
+        '''
+            Generate a box for a specific metric for a punctual analysis.
+
+            :param input_file: filename of the file that contains the evaluation data about the metric.
+        '''
+        df = pd.read_csv(input_file)
+
+        df_melted = df.melt(id_vars='Dimension', value_vars=['Min', 'Q1', 'Median', 'Q3', 'Max'], 
+                var_name='Statistic', value_name='Value')
+
+        sns.set_context("talk", font_scale=1.7)
+        plt.figure(figsize=(55, 35))
+        sns.boxplot(x='Dimension', y='Value', hue='Dimension', data=df_melted)
+        plt.xticks(rotation=90)
+        plt.title('Quality dimension evaluation')
+        plt.xlabel('Dimension',ha='center')
+        plt.ylabel('Values')
+        plt.savefig('test')
+
+    def generate_combined_boxplot_over_time(self, time_period_range, dimensions_to_exclude):
+        """
+            Creates a boxplot where on the x-axis is time and and for each measurement, a box for each metric.
+
+            :param time_period_range: Time data selection interval, e.g., monthly(M), quarterly(Q), all (A).
+            :param dimensions_to_exclude: Array pf strings that contains the name of the metrics to exclude from the boxplot.
+        
+        """
+        dfs = []
+        for file in self.analysis_results_files:
+            df = pd.read_csv(file)
+            dimension_name = os.path.splitext(os.path.basename(file))[0]
+            dimension_name = dimension_name.split(' ')[0]
+            if dimension_name in dimensions_to_exclude:
+                continue
+
+            if dimension_name == 'Representational-Consistency':
+                dimension_name = 'Interoperability'
+            if dimension_name == 'Representational-Conciseness':
+                dimension_name = 'Rep.-Conc.'
+            if dimension_name == 'Understandability':
+                dimension_name = 'Underst.'
+
+            df["Dimension"] = dimension_name
+
+            df['Analysis date'] = pd.to_datetime(df['Analysis date'])
+            df = df.sort_values(by='Analysis date')
+            if time_period_range != 'A':
+                df_filtered = df.groupby(df['Analysis date'].dt.to_period(time_period_range)).first()
+            else:
+                df_filtered = df
+
+            dfs.append(df_filtered)
+
+        data = pd.concat(dfs)
+
+        melted_data = pd.melt(
+            data, 
+            id_vars=['Analysis date', 'Dimension'], 
+            value_vars=['Min', 'Q1', 'Median', 'Q3', 'Max'], 
+            var_name='Statistic', 
+            value_name='Value'
+        )
+
+        plt.figure(figsize=(25, 9))
+        sns.boxplot(x='Analysis date', y='Value', hue='Dimension', data=melted_data)
+
+        plt.xticks(rotation=45)
+        plt.xlabel("Analysis Date")
+        plt.ylabel("Value")
+        plt.title("Quality data for each metric")
+
+        plt.legend(bbox_to_anchor=(1.01, 1), loc='best', borderaxespad=0.)
+
+        plt.savefig(f'{self.output_file}/dimensions_over_time')
+
+charts = GenerateCharts('./evaluation_results/over_time/')
+charts.generate_combined_boxplot_over_time('M',['Accuracy'])
+