Update app.py

app.py

@@ -9,45 +9,64 @@ from autoviz.AutoViz_Class import AutoViz_Class
 from traceml.summary.df import DataFrameSummary
 
 def variable_table(df):
-  """
-  Analyzes a DataFrame and categorizes variables with colorful HTML formatting.
-
-  Args:
+    """
+    Analyzes a DataFrame and categorizes variables with colorful HTML formatting.
+    
+    Args:
       df (pandas.DataFrame): The DataFrame to analyze.
-
-  Returns:
+    
+    Returns:
       str: HTML code representing the analysis results with colorful highlights.
-  """
-  # Analyze variable types
-  categorical_vars = df.select_dtypes(include=['category', 'object']).columns.tolist()
-  numerical_vars = df.select_dtypes(include=['int64', 'float64']).columns.tolist()
-  text_vars = df.select_dtypes(include=['object']).difference(categorical_vars).tolist()
-
-  # Build HTML table with styles
-  table_style = 'border: 1px solid #ddd; border-collapse: collapse; text-align: left; font-size: 14px;'
-  header_style = 'background-color: #f2f2f2; padding: 5px 10px;'
-  data_style = 'padding: 5px 10px; border-bottom: 1px solid #ddd;'
-  category_color = '#90ee90'  # Light green for categorical
-  numerical_color = '#add8e6'  # Light blue for numerical
-  text_color = '#ffd9b3'  # Light yellow for text
-
-  html = f"<table style='{table_style}'>"
-  html += f"<tr><th style='{header_style}'>Variable Type</th><th style='{header_style}'>Columns</th></tr>"
-
-  # Add rows for each variable type with coloring
-  if categorical_vars:
+    """
+    # Analyze variable types
+    numerical_vars=[]
+    text_vars=[]
+    categorical_vars=[]
+    # Identify categorical, numerical, and text columns
+    # Iterate over columns
+    for col in df.columns:
+        print(col )
+        unique_values = df[col].nunique()
+        total_values = df[col].count()
+        dtype_kind = df[col].dtype.kind
+        print( 'unique value-',unique_values,' total_values ',total_values,dtype_kind)
+        # Check criteria for numerical, text, and categorical variables
+        if (dtype_kind == 'f' or dtype_kind == 'i') and unique_values >= total_values / 10:
+            numerical_vars.append(col)
+            print('Numerical')
+        elif (dtype_kind == 'O' or dtype_kind == 'b') and unique_values > total_values / 10:
+            text_vars.append(col)
+            print('Text')
+        else:
+            categorical_vars.append(col)
+            print('categorical')
+    print(categorical_vars,numerical_vars,text_vars)
+
+    # Build HTML table with styles
+    table_style = 'border: 1px solid #ddd; border-collapse: collapse; text-align: left; font-size: 14px;'
+    header_style = 'background-color: #f2f2f2; padding: 5px 10px;'
+    data_style = 'padding: 5px 10px; border-bottom: 1px solid #ddd;'
+    category_color = '#90ee90'  # Light green for categorical
+    numerical_color = '#add8e6'  # Light blue for numerical
+    text_color = '#ffd9b3'  # Light yellow for text
+    
+    html = f"<table style='{table_style}'>"
+    html += f"<tr><th style='{header_style}'>Variable Type</th><th style='{header_style}'>Columns</th></tr>"
+    
+    # Add rows for each variable type with coloring
+    if categorical_vars:
     html += f"<tr style='background-color: {category_color};'><td>Categorical</td><td style='{data_style}'>{', '.join(categorical_vars)}</td></tr>"
-  if numerical_vars:
+    if numerical_vars:
     html += f"<tr style='background-color: {numerical_color};'><td>Numerical</td><td style='{data_style}'>{', '.join(numerical_vars)}</td></tr>"
-  if text_vars:
+    if text_vars:
     html += f"<tr style='background-color: {text_color};'><td>Text</td><td style='{data_style}'>{', '.join(text_vars)}</td></tr>"
-
-  # Handle cases where no variables are found
-  if not (categorical_vars or numerical_vars or text_vars):
+    
+    # Handle cases where no variables are found
+    if not (categorical_vars or numerical_vars or text_vars):
     html += "<tr><td>No variables found!</td></tr>"
-
-  html += "</table>"
-  return html
+    
+    html += "</table>"
+    return html
 
 def generate_report(file, type):
     df = pd.read_csv(file) if file.name.endswith(".csv") else pd.read_excel(file)