Update app.py

app.py

@@ -18,14 +18,74 @@ description = """
 Please give it 3 to 4 minutes for the model to load and Run , consider using Python code with less than 120 lines of code due to GPU constrainst
 """
 css = """.toast-wrap { display: none !important } """
-examples=[
-    ['Hello there! How are you doing?'],
-    ['Can you explain to me briefly what is Python programming language?'],
-    ['Explain the plot of Cinderella in a sentence.'],
-    ['How many hours does it take a man to eat a Helicopter?'],
-    ["Write a 100-word article on 'Benefits of Open-Source in AI research'"],
-    ]
-
+ examples=[["""  
+import pandas as pd
+import re
+import ast
+from code_bert_score import score
+import numpy as np
+def preprocess_code(source_text):
+   
+    def remove_comments_and_docstrings(source_code):
+        source_code = re.sub(r'#.*', '', source_code)
+        source_code = re.sub(r'(\'\'\'(.*?)\'\'\'|\"\"\"(.*?)\"\"\")', '', source_code, flags=re.DOTALL)
+        return source_code
+    pattern = r"```python\s+(.+?)\s+```"
+    matches = re.findall(pattern, source_text, re.DOTALL)
+    code_to_process = '\n'.join(matches) if matches else source_text
+    cleaned_code = remove_comments_and_docstrings(code_to_process)
+    return cleaned_code
+def evaluate_dataframe(df):
+   
+    results = {'P': [], 'R': [], 'F1': [], 'F3': []}
+    for index, row in df.iterrows():
+        try:
+            cands = [preprocess_code(row['generated_text'])]
+            refs = [preprocess_code(row['output'])]
+            P, R, F1, F3 = score(cands, refs, lang='python')
+            results['P'].append(P[0])
+            results['R'].append(R[0])
+            results['F1'].append(F1[0])
+            results['F3'].append(F3[0])
+        except Exception as e:
+            print(f"Error processing row {index}: {e}")
+            for key in results.keys():
+                results[key].append(None)
+    df_metrics = pd.DataFrame(results)
+    return df_metrics
+def evaluate_dataframe_multiple_runs(df, runs=3):
+    
+    all_results = []
+    for run in range(runs):
+        df_metrics = evaluate_dataframe(df)
+        all_results.append(df_metrics)
+    # Calculate mean and std deviation of metrics across runs
+    df_metrics_mean = pd.concat(all_results).groupby(level=0).mean()
+    df_metrics_std = pd.concat(all_results).groupby(level=0).std()
+    return df_metrics_mean, df_metrics_std
+""" ] ,
+                         ["""
+def analyze_sales_data(sales_records):
+    active_sales = filter(lambda record: record['status'] == 'active', sales_records)
+    sales_by_category = {}
+    for record in active_sales:
+        category = record['category']
+        total_sales = record['units_sold'] * record['price_per_unit']
+        if category not in sales_by_category:
+            sales_by_category[category] = {'total_sales': 0, 'total_units': 0}
+        sales_by_category[category]['total_sales'] += total_sales
+        sales_by_category[category]['total_units'] += record['units_sold']
+    average_sales_data = []
+    for category, data in sales_by_category.items():
+        average_sales = data['total_sales'] / data['total_units']
+        sales_by_category[category]['average_sales'] = average_sales
+        average_sales_data.append((category, average_sales))
+    average_sales_data.sort(key=lambda x: x[1], reverse=True)
+    for rank, (category, _) in enumerate(average_sales_data, start=1):
+        sales_by_category[category]['rank'] = rank
+    return sales_by_category
+"""]
+                     ]
 
 # Note: We have removed default system prompt as requested by the paper authors [Dated: 13/Oct/2023]
 # Prompting style for Llama2 without using system prompt