optimize with new column names

.gitignore

@@ -1,6 +1,16 @@
-# Ignore all contents of these directories
-/Data/
-/attach/
-/results/
-/enarcelona/
+# 1. Broad Ignores
+/Data/*
+/attach/*
+/results/*
+/enarcelona/*
 .env
+__pycache__/
+*.pyc
+
+# 2. Ignore virtual environments COMPLETELY
+# This must come BEFORE the unignore rule
+env*/
+
+# 3. The "Unignore" rule (Whitelisting)
+# We only unignore .py files that aren't already blocked by the rules above
+!**/*.py

Data/show_plots.py

@@ -0,0 +1,570 @@
+# %% Scatter
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+
+# Load your data from TSV file
+file_path = '/home/shahin/Lab/Doktorarbeit/Barcelona/Data/join_MS_Briefe_400_with_unique_id_SHA3_explore_cleaned_results+MS_Briefe_400_with_unique_id_SHA3_explore_cleaned.tsv'
+df = pd.read_csv(file_path, sep='\t')
+
+# Replace comma with dot for numeric conversion in GT_EDSS and LLM_Results
+df['GT_EDSS'] = df['GT_EDSS'].astype(str).str.replace(',', '.')
+df['LLM_Results'] = df['LLM_Results'].astype(str).str.replace(',', '.')
+
+# Convert to float (handle invalid entries gracefully)
+df['GT_EDSS'] = pd.to_numeric(df['GT_EDSS'], errors='coerce')
+df['LLM_Results'] = pd.to_numeric(df['LLM_Results'], errors='coerce')
+
+# Drop rows where either column is NaN
+df_clean = df.dropna(subset=['GT_EDSS', 'LLM_Results'])
+
+# Create scatter plot
+plt.figure(figsize=(8, 6))
+plt.scatter(df_clean['GT_EDSS'], df_clean['LLM_Results'], alpha=0.7, color='blue')
+
+# Add labels and title
+plt.xlabel('GT_EDSS')
+plt.ylabel('LLM_Results')
+plt.title('Comparison of GT_EDSS vs LLM_Results')
+
+# Optional: Add a diagonal line for reference (perfect prediction)
+plt.plot([0, max(df_clean['GT_EDSS'])], [0, max(df_clean['GT_EDSS'])], color='red', linestyle='--', label='Perfect Prediction')
+plt.legend()
+
+# Show plot
+plt.grid(True)
+plt.tight_layout()
+plt.show()
+
+##
+
+
+# %% Bland0-altman
+
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+import statsmodels.api as sm
+
+# Load your data from TSV file
+file_path = '/home/shahin/Lab/Doktorarbeit/Barcelona/Data/join_MS_Briefe_400_with_unique_id_SHA3_explore_cleaned_results+MS_Briefe_400_with_unique_id_SHA3_explore_cleaned.tsv'
+df = pd.read_csv(file_path, sep='\t')
+
+# Replace comma with dot for numeric conversion in GT_EDSS and LLM_Results
+df['GT_EDSS'] = df['GT_EDSS'].astype(str).str.replace(',', '.')
+df['LLM_Results'] = df['LLM_Results'].astype(str).str.replace(',', '.')
+
+# Convert to float (handle invalid entries gracefully)
+df['GT_EDSS'] = pd.to_numeric(df['GT_EDSS'], errors='coerce')
+df['LLM_Results'] = pd.to_numeric(df['LLM_Results'], errors='coerce')
+
+# Drop rows where either column is NaN
+df_clean = df.dropna(subset=['GT_EDSS', 'LLM_Results'])
+
+# Create Bland-Altman plot
+f, ax = plt.subplots(1, figsize=(8, 5))
+sm.graphics.mean_diff_plot(df_clean['GT_EDSS'], df_clean['LLM_Results'], ax=ax)
+
+# Add labels and title
+ax.set_title('Bland-Altman Plot: GT_EDSS vs LLM_Results')
+ax.set_xlabel('Mean of GT_EDSS and LLM_Results')
+ax.set_ylabel('Difference between GT_EDSS and LLM_Results')
+
+# Display Bland-Altman plot
+plt.tight_layout()
+plt.show()
+
+# Print some statistics
+mean_diff = np.mean(df_clean['GT_EDSS'] - df_clean['LLM_Results'])
+std_diff = np.std(df_clean['GT_EDSS'] - df_clean['LLM_Results'])
+print(f"Mean difference: {mean_diff:.3f}")
+print(f"Standard deviation of differences: {std_diff:.3f}")
+print(f"95% Limits of Agreement: [{mean_diff - 1.96*std_diff:.3f}, {mean_diff + 1.96*std_diff:.3f}]")
+
+##
+
+
+
+# %%  Confusion matrix
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+from sklearn.metrics import confusion_matrix, classification_report
+import seaborn as sns
+
+# Load your data from TSV file
+file_path = '/home/shahin/Lab/Doktorarbeit/Barcelona/Data/Join_edssandsub.tsv'
+df = pd.read_csv(file_path, sep='\t')
+
+# Replace comma with dot for numeric conversion in GT.EDSS and result.EDSS
+df['GT.EDSS'] = df['GT.EDSS'].astype(str).str.replace(',', '.')
+df['result.EDSS'] = df['result.EDSS'].astype(str).str.replace(',', '.')
+
+# Convert to float (handle invalid entries gracefully)
+df['GT.EDSS'] = pd.to_numeric(df['GT.EDSS'], errors='coerce')
+df['result.EDSS'] = pd.to_numeric(df['result.EDSS'], errors='coerce')
+
+# Drop rows where either column is NaN
+df_clean = df.dropna(subset=['GT.EDSS', 'result.EDSS'])
+
+# For confusion matrix, we need to categorize the values
+# Let's create categories up to 10 (0-1, 1-2, 2-3, ..., 9-10)
+def categorize_edss(value):
+    if pd.isna(value):
+        return np.nan
+    elif value <= 1.0:
+        return '0-1'
+    elif value <= 2.0:
+        return '1-2'
+    elif value <= 3.0:
+        return '2-3'
+    elif value <= 4.0:
+        return '3-4'
+    elif value <= 5.0:
+        return '4-5'
+    elif value <= 6.0:
+        return '5-6'
+    elif value <= 7.0:
+        return '6-7'
+    elif value <= 8.0:
+        return '7-8'
+    elif value <= 9.0:
+        return '8-9'
+    elif value <= 10.0:
+        return '9-10'
+    else:
+        return '10+'
+
+# Create categorical versions
+df_clean['GT.EDSS_cat'] = df_clean['GT.EDSS'].apply(categorize_edss)
+df_clean['result.EDSS_cat'] = df_clean['result.EDSS'].apply(categorize_edss)
+
+# Remove any NaN categories
+df_clean = df_clean.dropna(subset=['GT.EDSS_cat', 'result.EDSS_cat'])
+
+# Create confusion matrix
+cm = confusion_matrix(df_clean['GT.EDSS_cat'], df_clean['result.EDSS_cat'],
+                     labels=['0-1', '1-2', '2-3', '3-4', '4-5', '5-6', '6-7', '7-8', '8-9', '9-10'])
+
+# Plot confusion matrix
+plt.figure(figsize=(10, 8))
+sns.heatmap(cm, annot=True, fmt='d', cmap='Blues',
+            xticklabels=['0-1', '1-2', '2-3', '3-4', '4-5', '5-6', '6-7', '7-8', '8-9', '9-10'],
+            yticklabels=['0-1', '1-2', '2-3', '3-4', '4-5', '5-6', '6-7', '7-8', '8-9', '9-10'])
+plt.title('Confusion Matrix: Ground truth EDSS vs interferred EDSS (Categorized 0-10)')
+plt.xlabel('LLM Generated EDSS')
+plt.ylabel('Ground Truth EDSS')
+plt.tight_layout()
+plt.show()
+
+# Print classification report
+print("Classification Report:")
+print(classification_report(df_clean['GT.EDSS_cat'], df_clean['result.EDSS_cat']))
+
+# Print raw counts
+print("\nConfusion Matrix (Raw Counts):")
+print(cm)
+
+##
+
+
+
+# %% Classification 
+import pandas as pd
+import matplotlib.pyplot as plt
+import seaborn as sns
+from sklearn.metrics import confusion_matrix
+import numpy as np
+
+# Load your data from TSV file
+file_path ='/home/shahin/Lab/Doktorarbeit/Barcelona/Data/Join_edssandsub.tsv'
+
+df = pd.read_csv(file_path, sep='\t')
+
+# Check data structure
+print("Data shape:", df.shape)
+print("First few rows:")
+print(df.head())
+print("\nColumn names:")
+for col in df.columns:
+    print(f"  {col}")
+
+# Function to safely convert to boolean
+def safe_bool_convert(series):
+    '''Safely convert series to boolean, handling various input formats'''
+    # Convert to string first, then to boolean
+    series_str = series.astype(str).str.strip().str.lower()
+
+    # Handle different true/false representations
+    bool_map = {
+        'true': True, '1': True, 'yes': True, 'y': True,
+        'false': False, '0': False, 'no': False, 'n': False
+    }
+
+    converted = series_str.map(bool_map)
+
+    # Handle remaining NaN values
+    converted = converted.fillna(False)  # or True, depending on your preference
+
+    return converted
+
+# Convert columns safely
+if 'result.klassifizierbar' in df.columns:
+    print("\nresult.klassifizierbar column info:")
+    print(df['result.klassifizierbar'].head(10))
+    print("Unique values:", df['result.klassifizierbar'].unique())
+
+    df['result.klassifizierbar'] = safe_bool_convert(df['result.klassifizierbar'])
+    print("After conversion:")
+    print(df['result.klassifizierbar'].value_counts())
+
+if 'GT.klassifizierbar' in df.columns:
+    print("\nGT.klassifizierbar column info:")
+    print(df['GT.klassifizierbar'].head(10))
+    print("Unique values:", df['GT.klassifizierbar'].unique())
+
+    df['GT.klassifizierbar'] = safe_bool_convert(df['GT.klassifizierbar'])
+    print("After conversion:")
+    print(df['GT.klassifizierbar'].value_counts())
+
+# Create bar chart showing only True values for klassifizierbar
+if 'result.klassifizierbar' in df.columns and 'GT.klassifizierbar' in df.columns:
+    # Get counts for True values only
+    llm_true_count = df['result.klassifizierbar'].sum()
+    gt_true_count = df['GT.klassifizierbar'].sum()
+
+    # Plot using matplotlib directly
+    fig, ax = plt.subplots(figsize=(8, 6))
+
+    x = np.arange(2)
+    width = 0.35
+
+    bars1 = ax.bar(x[0] - width/2, llm_true_count, width, label='LLM', color='skyblue', alpha=0.8)
+    bars2 = ax.bar(x[1] + width/2, gt_true_count, width, label='GT', color='lightcoral', alpha=0.8)
+
+    # Add value labels on bars
+    ax.annotate(f'{llm_true_count}',
+                xy=(x[0], llm_true_count),
+                xytext=(0, 3),
+                textcoords="offset points",
+                ha='center', va='bottom')
+
+    ax.annotate(f'{gt_true_count}',
+                xy=(x[1], gt_true_count),
+                xytext=(0, 3),
+                textcoords="offset points",
+                ha='center', va='bottom')
+
+    ax.set_xlabel('Classification Status (klassifizierbar)')
+    ax.set_ylabel('Count')
+    ax.set_title('True Values Comparison: LLM vs GT for "klassifizierbar"')
+    ax.set_xticks(x)
+    ax.set_xticklabels(['LLM', 'GT'])
+    ax.legend()
+
+    plt.tight_layout()
+    plt.show()
+
+# Create confusion matrix if both columns exist
+if 'result.klassifizierbar' in df.columns and 'GT.klassifizierbar' in df.columns:
+    try:
+        # Ensure both columns are boolean
+        llm_bool = df['result.klassifizierbar'].fillna(False).astype(bool)
+        gt_bool = df['GT.klassifizierbar'].fillna(False).astype(bool)
+
+        cm = confusion_matrix(gt_bool, llm_bool)
+
+        # Plot confusion matrix
+        fig, ax = plt.subplots(figsize=(8, 6))
+        sns.heatmap(cm, annot=True, fmt='d', cmap='Blues',
+                   xticklabels=['False ', 'True '],
+                   yticklabels=['False', 'True '],
+                   ax=ax)
+        ax.set_xlabel('LLM Predictions ')
+        ax.set_ylabel('GT Labels ')
+        ax.set_title('Confusion Matrix: LLM vs GT for "klassifizierbar"')
+
+        plt.tight_layout()
+        plt.show()
+
+        print("Confusion Matrix:")
+        print(cm)
+
+    except Exception as e:
+        print(f"Error creating confusion matrix: {e}")
+
+# Show final data info
+print("\nFinal DataFrame info:")
+print(df[['result.klassifizierbar', 'GT.klassifizierbar']].info())
+
+##
+
+
+
+
+# %% Boxplot
+import pandas as pd
+import matplotlib.pyplot as plt
+import seaborn as sns
+import numpy as np
+
+# Load your data from TSV file
+file_path = '/home/shahin/Lab/Doktorarbeit/Barcelona/Data/join_results_unique.tsv'
+df = pd.read_csv(file_path, sep='\t')
+
+# Replace comma with dot for numeric conversion in GT.EDSS and result.EDSS
+df['GT.EDSS'] = df['GT.EDSS'].astype(str).str.replace(',', '.')
+df['result.EDSS'] = df['result.EDSS'].astype(str).str.replace(',', '.')
+
+# Convert to float (handle invalid entries gracefully)
+df['GT.EDSS'] = pd.to_numeric(df['GT.EDSS'], errors='coerce')
+df['result.EDSS'] = pd.to_numeric(df['result.EDSS'], errors='coerce')
+
+# Drop rows where either column is NaN
+df_clean = df.dropna(subset=['GT.EDSS', 'result.EDSS'])
+
+# 1. DEFINE CATEGORY ORDER
+# This ensures the X-axis is numerically logical (0-1 comes before 1-2)
+category_order = ['0-1', '1-2', '2-3', '3-4', '4-5', '5-6', '6-7', '7-8', '8-9', '9-10', '10+']
+
+# Convert the column to a Categorical type with the specific order
+df_clean['GT.EDSS_cat'] = pd.Categorical(df_clean['GT.EDSS'].apply(categorize_edss), 
+                                         categories=category_order, 
+                                         ordered=True)
+
+plt.figure(figsize=(14, 8))
+
+# 2. ADD HUE FOR LEGEND
+# Assigning x to 'hue' allows Seaborn to generate a legend automatically
+box_plot = sns.boxplot(
+    data=df_clean, 
+    x='GT.EDSS_cat', 
+    y='result.EDSS',
+    hue='GT.EDSS_cat',  # Added hue
+    palette='viridis', 
+    linewidth=1.5,
+    legend=True         # Ensure legend is enabled
+)
+
+# 3. CUSTOMIZE PLOT
+plt.title('Distribution of result.EDSS by GT.EDSS Category', fontsize=18, pad=20)
+plt.xlabel('Ground Truth EDSS Category', fontsize=14)
+plt.ylabel('LLM Predicted EDSS', fontsize=14)
+
+# Move legend to the side or top
+plt.legend(title="EDSS Categories", bbox_to_anchor=(1.05, 1), loc='upper left')
+
+plt.xticks(rotation=45, ha='right', fontsize=10)
+plt.grid(True, axis='y', alpha=0.3)
+plt.tight_layout()
+
+plt.show()
+##
+
+
+# %% Postproccessing Column names
+
+import pandas as pd
+
+# Read the TSV file
+file_path = '/home/shahin/Lab/Doktorarbeit/Barcelona/Data/Join_edssandsub.tsv'
+df = pd.read_csv(file_path, sep='\t')
+
+# Create a mapping dictionary for German to English column names
+column_mapping = {
+    'EDSS':'GT.EDSS',
+    'klassifizierbar': 'GT.klassifizierbar',
+    'Sehvermögen': 'GT.VISUAL_OPTIC_FUNCTIONS',
+    'Cerebellum': 'GT.CEREBELLAR_FUNCTIONS',
+    'Hirnstamm': 'GT.BRAINSTEM_FUNCTIONS',
+    'Sensibiliät': 'GT.SENSORY_FUNCTIONS',
+    'Pyramidalmotorik': 'GT.PYRAMIDAL_FUNCTIONS',
+    'Ambulation': 'GT.AMBULATION',
+    'Cerebrale_Funktion': 'GT.CEREBRAL_FUNCTIONS',
+    'Blasen-_und_Mastdarmfunktion': 'GT.BOWEL_AND_BLADDER_FUNCTIONS'
+}
+
+# Rename columns
+df = df.rename(columns=column_mapping)
+
+# Save the modified dataframe back to TSV file
+df.to_csv(file_path, sep='\t', index=False)
+
+print("Columns have been successfully renamed!")
+print("Renamed columns:")
+for old_name, new_name in column_mapping.items():
+    if old_name in df.columns:
+        print(f"  {old_name} -> {new_name}")
+
+
+##
+
+
+
+
+# %% Styled table
+import pandas as pd
+import numpy as np
+import seaborn as sns
+import matplotlib.pyplot as plt
+import dataframe_image as dfi
+# Load data
+df = pd.read_csv("/home/shahin/Lab/Doktorarbeit/Barcelona/Data/Join_edssandsub.tsv", sep='\t')
+
+# 1. Identify all GT and result columns
+gt_columns = [col for col in df.columns if col.startswith('GT.')]
+result_columns = [col for col in df.columns if col.startswith('result.')]
+
+print("GT Columns found:", gt_columns)
+print("Result Columns found:", result_columns)
+
+# 2. Create proper mapping between GT and result columns
+# Handle various naming conventions (spaces, underscores, etc.)
+column_mapping = {}
+
+for gt_col in gt_columns:
+    base_name = gt_col.replace('GT.', '')
+
+    # Clean the base name for matching - remove spaces, underscores, etc.
+    # Try different matching approaches
+    candidates = [
+        f'result.{base_name}',  # Exact match
+        f'result.{base_name.replace(" ", "_")}',  # With underscores
+        f'result.{base_name.replace("_", " ")}',  # With spaces
+        f'result.{base_name.replace(" ", "")}',   # No spaces
+        f'result.{base_name.replace("_", "")}'    # No underscores
+    ]
+
+    # Also try case-insensitive matching
+    candidates.append(f'result.{base_name.lower()}')
+    candidates.append(f'result.{base_name.upper()}')
+
+    # Try to find matching result column
+    matched = False
+    for candidate in candidates:
+        if candidate in result_columns:
+            column_mapping[gt_col] = candidate
+            matched = True
+            break
+
+    # If no exact match found, try partial matching
+    if not matched:
+        # Try to match by removing special characters and comparing
+        base_clean = ''.join(e for e in base_name if e.isalnum() or e in ['_', ' '])
+        for result_col in result_columns:
+            result_base = result_col.replace('result.', '')
+            result_clean = ''.join(e for e in result_base if e.isalnum() or e in ['_', ' '])
+            if base_clean.lower() == result_clean.lower():
+                column_mapping[gt_col] = result_col
+                matched = True
+                break
+
+print("Column mapping:", column_mapping)
+
+# 3. Faster, vectorized computation using the corrected mapping
+data_list = []
+
+for gt_col, result_col in column_mapping.items():
+    print(f"Processing {gt_col} vs {result_col}")
+
+    # Convert to numeric, forcing errors to NaN
+    s1 = pd.to_numeric(df[gt_col], errors='coerce').astype(float)
+    s2 = pd.to_numeric(df[result_col], errors='coerce').astype(float)
+
+    # Calculate matches (abs difference <= 0.5)
+    diff = np.abs(s1 - s2)
+    matches = (diff <= 0.5).sum()
+
+    # Determine the denominator (total valid comparisons)
+    valid_count = diff.notna().sum()
+
+    if valid_count > 0:
+        percentage = (matches / valid_count) * 100
+    else:
+        percentage = 0
+
+    # Extract clean base name for display
+    base_name = gt_col.replace('GT.', '')
+
+    data_list.append({
+        'GT': base_name,
+        'Match %': round(percentage, 1)
+    })
+
+
+
+
+# 4. Prepare Data
+match_df = pd.DataFrame(data_list)
+# Clean up labels: Replace underscores with spaces and capitalize
+match_df['GT'] = match_df['GT'].str.replace('_', ' ').str.title()
+match_df = match_df.sort_values('Match %', ascending=False)
+
+# 5. Create a "Beautiful" Table using Seaborn Heatmap
+def create_luxury_table(df, output_file="edss_agreement.png"):
+    # Set the aesthetic style
+    sns.set_theme(style="white", font="sans-serif")
+
+    # Prepare data for heatmap
+    plot_data = df.set_index('GT')[['Match %']]
+
+    # Initialize the figure
+    # Height is dynamic based on number of rows
+    fig, ax = plt.subplots(figsize=(8, len(df) * 0.6))
+
+    # Create a custom diverging color map (Deep Red -> Mustard -> Emerald)
+    # This looks more professional than standard 'RdYlGn'
+    cmap = sns.diverging_palette(15, 135, s=80, l=55, as_cmap=True)
+
+    # Draw the heatmap
+    sns.heatmap(
+        plot_data,
+        annot=True,
+        fmt=".1f",
+        cmap=cmap,
+        center=85,      # Centers the color transition
+        vmin=50, vmax=100, # Range of the gradient
+        linewidths=2,
+        linecolor='white',
+        cbar=False,     # Remove color bar for a "table" look
+        annot_kws={"size": 14, "weight": "bold", "family": "sans-serif"}
+    )
+
+    # Styling the Axes (Turning the heatmap into a table)
+    ax.set_xlabel("")
+    ax.set_ylabel("")
+    ax.xaxis.tick_top() # Move "Match %" label to top
+    ax.set_xticklabels(['Agreement (%)'], fontsize=14, fontweight='bold', color='#2c3e50')
+    ax.tick_params(axis='y', labelsize=12, labelcolor='#2c3e50', length=0)
+
+    # Add a thin border around the plot
+    for _, spine in ax.spines.items():
+        spine.set_visible(True)
+        spine.set_color('#ecf0f1')
+
+    plt.title('EDSS Subcategory Consistency Analysis', fontsize=16, pad=40, fontweight='bold', color='#2c3e50')
+
+    # Add a subtle footer
+    plt.figtext(0.5, 0.0, "Tolerance: ±0.5 points",
+                wrap=True, horizontalalignment='center', fontsize=10, color='gray', style='italic')
+
+    # Save with high resolution
+    plt.tight_layout()
+    plt.savefig(output_file, dpi=300, bbox_inches='tight')
+    print(f"Beautiful table saved as {output_file}")
+
+# Execute
+create_luxury_table(match_df)
+
+
+# Run the function
+save_styled_table(match_df)
+# 6. Save as SVG
+
+plt.savefig("agreement_table.svg", format='svg', dpi=300, bbox_inches='tight')
+print("Successfully saved agreement_table.svg")
+
+# Show plot if running in a GUI environment
+plt.show()
+
+##

Data/style2.py

@@ -0,0 +1,135 @@
+import pandas as pd
+import numpy as np
+import seaborn as sns
+import matplotlib.pyplot as plt
+import dataframe_image as dfi
+# Load data
+df = pd.read_csv("/home/shahin/Lab/Doktorarbeit/Barcelona/Data/Join_edssandsub.tsv", sep='\t')
+
+# 1. Identify all GT and result columns
+gt_columns = [col for col in df.columns if col.startswith('GT.')]
+result_columns = [col for col in df.columns if col.startswith('result.')]
+
+print("GT Columns found:", gt_columns)
+print("Result Columns found:", result_columns)
+
+# 2. Create proper mapping between GT and result columns
+# Handle various naming conventions (spaces, underscores, etc.)
+column_mapping = {}
+
+for gt_col in gt_columns:
+    base_name = gt_col.replace('GT.', '')
+
+    # Clean the base name for matching - remove spaces, underscores, etc.
+    # Try different matching approaches
+    candidates = [
+        f'result.{base_name}',  # Exact match
+        f'result.{base_name.replace(" ", "_")}',  # With underscores
+        f'result.{base_name.replace("_", " ")}',  # With spaces
+        f'result.{base_name.replace(" ", "")}',   # No spaces
+        f'result.{base_name.replace("_", "")}'    # No underscores
+    ]
+
+    # Also try case-insensitive matching
+    candidates.append(f'result.{base_name.lower()}')
+    candidates.append(f'result.{base_name.upper()}')
+
+    # Try to find matching result column
+    matched = False
+    for candidate in candidates:
+        if candidate in result_columns:
+            column_mapping[gt_col] = candidate
+            matched = True
+            break
+
+    # If no exact match found, try partial matching
+    if not matched:
+        # Try to match by removing special characters and comparing
+        base_clean = ''.join(e for e in base_name if e.isalnum() or e in ['_', ' '])
+        for result_col in result_columns:
+            result_base = result_col.replace('result.', '')
+            result_clean = ''.join(e for e in result_base if e.isalnum() or e in ['_', ' '])
+            if base_clean.lower() == result_clean.lower():
+                column_mapping[gt_col] = result_col
+                matched = True
+                break
+
+print("Column mapping:", column_mapping)
+
+# 3. Faster, vectorized computation using the corrected mapping
+data_list = []
+
+for gt_col, result_col in column_mapping.items():
+    print(f"Processing {gt_col} vs {result_col}")
+
+    # Convert to numeric, forcing errors to NaN
+    s1 = pd.to_numeric(df[gt_col], errors='coerce').astype(float)
+    s2 = pd.to_numeric(df[result_col], errors='coerce').astype(float)
+
+    # Calculate matches (abs difference <= 0.5)
+    diff = np.abs(s1 - s2)
+    matches = (diff <= 0.5).sum()
+
+    # Determine the denominator (total valid comparisons)
+    valid_count = diff.notna().sum()
+
+    if valid_count > 0:
+        percentage = (matches / valid_count) * 100
+    else:
+        percentage = 0
+
+    # Extract clean base name for display
+    base_name = gt_col.replace('GT.', '')
+
+    data_list.append({
+        'GT': base_name,
+        'Match %': round(percentage, 1)
+    })
+
+
+
+# 4. Prepare Data for Plotting
+match_df = pd.DataFrame(data_list)
+match_df = match_df.sort_values('Match %', ascending=False) # Sort for better visual flow
+
+# 5. Create the Styled Gradient Table
+def style_agreement_table(df):
+    return (df.style
+        .format({'Match %': '{:.1f}%'}) # Add % sign
+        .background_gradient(cmap='RdYlGn', subset=['Match %'], vmin=50, vmax=100) # Red to Green gradient
+        .set_properties(**{
+            'text-align': 'center',
+            'font-size': '12pt',
+            'border-collapse': 'collapse',
+            'border': '1px solid #D3D3D3'
+        })
+        .set_table_styles([
+            # Style the header
+            {'selector': 'th', 'props': [
+                ('background-color', '#404040'), 
+                ('color', 'white'),
+                ('font-weight', 'bold'),
+                ('text-transform', 'uppercase'),
+                ('padding', '10px')
+            ]},
+            # Add hover effect
+            {'selector': 'tr:hover', 'props': [('background-color', '#f5f5f5')]}
+        ])
+        .set_caption("EDSS Agreement Analysis: Ground Truth vs. Results (Tolerance ±0.5)")
+    )
+
+# To display in a Jupyter Notebook:
+styled_table = style_agreement_table(match_df)
+styled_table
+
+dfi.export(styled_table, "styled_table.png")
+#styled_table.to_html("agreement_report.html")
+# 6. Save as SVG
+
+#plt.savefig("agreement_table.svg", format='svg', dpi=300, bbox_inches='tight')
+#print("Successfully saved agreement_table.svg")
+
+# Show plot if running in a GUI environment
+plt.show()
+
+

Data/styled_tables.py

@@ -0,0 +1,74 @@
+import pandas as pd
+import numpy as np
+import seaborn as sns
+
+# Sample data (replace with your actual df)
+df = pd.read_csv("/home/shahin/Lab/Doktorarbeit/Barcelona/Data/Join_edssandsub.tsv", sep='\t')
+
+# Identify GT and Result columns
+gt_columns = [col for col in df.columns if col.startswith('GT.')]
+result_columns = [col for col in df.columns if col.startswith('result.')]
+
+# Create mapping
+column_mapping = {}
+for gt_col in gt_columns:
+    base_name = gt_col.replace('GT.', '')
+    result_col = f'result.{base_name}'
+    if result_col in result_columns:
+        column_mapping[gt_col] = result_col
+
+# Function to compute match percentage for each GT-Result pair
+def compute_match_percentages(df, column_mapping):
+    percentages = []
+    for gt_col, result_col in column_mapping.items():
+        count = 0
+        total = len(df)
+
+        for _, row in df.iterrows():
+            gt_val = row[gt_col]
+            result_val = row[result_col]
+
+            # Handle NaN values
+            if pd.isna(gt_val) or pd.isna(result_val):
+                continue
+
+            # Handle non-numeric values
+            try:
+                gt_float = float(gt_val)
+                result_float = float(result_val)
+            except (ValueError, TypeError):
+                # Skip rows with non-numeric values
+                continue
+
+            # Check if values are within 0.5 tolerance
+            if abs(gt_float - result_float) <= 0.5:
+                count += 1
+
+        percentage = (count / total) * 100
+        percentages.append({
+            'GT_Column': gt_col,
+            'Result_Column': result_col,
+            'Match_Percentage': round(percentage, 1)
+        })
+
+    return pd.DataFrame(percentages)
+
+# Compute match percentages
+match_df = compute_match_percentages(df, column_mapping)
+
+# Create a pivot table for gradient display (optional but helpful)
+pivot_table = match_df.set_index(['GT_Column', 'Result_Column'])['Match_Percentage'].unstack(fill_value=0)
+
+# Apply gradient background
+cm = sns.light_palette("green", as_cmap=True)
+styled_table = pivot_table.style.background_gradient(cmap=cm, axis=None)
+
+# Display result
+print("Agreement Percentage Table (with gradient):")
+styled_table
+
+
+
+# Save the styled table to a file
+styled_table.to_html("agreement_report.html")
+print("Report saved to agreement_report.html")

app.py

@@ -214,3 +214,8 @@ if __name__ == "__main__":
     print(f"Results saved to {output_json}")
 ##
 
+
+
+# %% name
+eXXXXXXXX
+##

total_app.py

@@ -0,0 +1,149 @@
+import time
+import json
+import os
+from datetime import datetime
+import pandas as pd
+from openai import OpenAI
+from dotenv import load_dotenv
+
+# Load environment variables
+load_dotenv()
+
+# === CONFIGURATION ===
+OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
+OPENAI_BASE_URL = os.getenv("OPENAI_BASE_URL")
+MODEL_NAME = "GPT-OSS-120B"
+HEALTH_URL = f"{OPENAI_BASE_URL}/health"  # Placeholder - actual health check would need to be implemented
+CHAT_URL = f"{OPENAI_BASE_URL}/chat/completions"
+# File paths
+INPUT_CSV = "/home/shahin/Lab/Doktorarbeit/Barcelona/Data/MS_Briefe_400_with_unique_id_SHA3_explore_cleaned_unique.csv"
+EDSS_INSTRUCTIONS_PATH = "/home/shahin/Lab/Doktorarbeit/Barcelona/attach/Komplett.txt"
+#GRAMMAR_FILE = "/home/shahin/Lab/Doktorarbeit/Barcelona/attach/just_edss_schema.gbnf"
+# Initialize OpenAI client
+client = OpenAI(
+    api_key=OPENAI_API_KEY,
+    base_url=OPENAI_BASE_URL
+)
+# Read EDSS instructions from file
+with open(EDSS_INSTRUCTIONS_PATH, 'r') as f:
+    EDSS_INSTRUCTIONS = f.read().strip()
+
+# === RUN INFERENCE 2 ===
+def run_inference(patient_text, max_retries=3):
+    prompt = f'''Du bist ein medizinischer Assistent, der spezialisiert darauf ist, EDSS-Scores (Expanded Disability Status Scale) sowie alle Unterkategorien aus klinischen Berichten zu extrahieren.
+### Regeln für die Ausgabe:
+1. **Reason**: Erstelle eine prägnante Zusammenfassung (max. 400 Zeichen) der Befunde auf **DEUTSCH**, die zur Einstufung führen.
+2. **klassifizierbar**:
+   - Setze dies auf **true**, wenn ein EDSS-Wert identifiziert, berechnet oder basierend auf den klinischen Hinweisen plausibel geschätzt werden kann.
+   - Setze dies auf **false**, NUR wenn die Daten absolut unzureichend oder so widersprüchlich sind, dass keinerlei Einstufung möglich ist.
+3. **EDSS**:
+   - Dieses Feld ist **VERPFLICHTEND**, wenn "klassifizierbar" auf true steht.
+   - Es muss eine Zahl zwischen 0.0 und 10.0 sein.
+   - Versuche stets, den EDSS-Wert so präzise wie möglich zu bestimmen, auch wenn die Datenlage dünn ist (nutze verfügbare Informationen zu Gehstrecke und Funktionssystemen).
+   - Dieses Feld **DARF NICHT ERSCHEINEN**, wenn "klassifizierbar" auf false steht.
+4. **Unterkategorien**:
+   - Extrahiere alle folgenden Unterkategorien aus dem Bericht:
+   - VISUAL OPTIC FUNCTIONS (max. 6.0)
+   - BRAINSTEM FUNCTIONS (max. 6.0)
+   - PYRAMIDAL FUNCTIONS (max. 6.0)
+   - CEREBELLAR FUNCTIONS (max. 6.0)
+   - SENSORY FUNCTIONS (max. 6.0)
+   - BOWEL AND BLADDER FUNCTIONS (max. 6.0)
+   - CEREBRAL FUNCTIONS (max. 6.0)
+   - AMBULATION (max. 10.0)
+   - Jede Unterkategorie sollte eine Zahl zwischen 0.0 und der jeweiligen Obergrenze enthalten, wenn sie klassifizierbar ist
+   - Wenn eine Unterkategorie nicht klassifizierbar ist, setze den Wert auf null
+### Einschränkungen:
+- Erfinde keine Fakten, aber nutze klinische Herleitungen aus dem Bericht, um den EDSS und die Unterkategorien zu bestimmen.
+- Priorisiere die Vergabe eines EDSS-Wertes gegenüber der Markierung als nicht klassifizierbar.
+- Halte dich strikt an die JSON-Struktur.
+- Die Unterkategorien müssen immer enthalten sein, auch wenn sie null sind.
+EDSS-Bewertungsrichtlinien:
+{EDSS_INSTRUCTIONS}
+Patientenbericht:
+{patient_text}
+'''
+
+    start_time = time.time()
+    for attempt in range(max_retries + 1):
+        try:
+            response = client.chat.completions.create(
+                messages=[
+                    {
+                        "role": "system",
+                        "content": "You extract EDSS scores and all subcategories. You prioritize providing values even if data is partial, by using clinical inference."
+                    },
+                    {
+                        "role": "user",
+                        "content": prompt
+                    }
+                ],
+                model=MODEL_NAME,
+                max_tokens=2048,
+                temperature=0.0,
+                response_format={"type": "json_object"}
+            )
+            content = response.choices[0].message.content
+
+            if content is None or content.strip() == "":
+                raise ValueError("API returned empty or None response content")
+
+            parsed = json.loads(content)
+            inference_time = time.time() - start_time
+            return {
+                "success": True,
+                "result": parsed,
+                "inference_time_sec": inference_time
+            }
+
+        except Exception as e:
+            print(f"Attempt {attempt + 1} failed: {e}")
+            if attempt < max_retries:
+                time.sleep(2 ** attempt)  # Exponential backoff
+                continue
+            else:
+                print("All retries exhausted.")
+                return {
+                    "success": False,
+                    "error": str(e),
+                    "inference_time_sec": -1
+                }
+
+# === BUILD PATIENT TEXT ===
+def build_patient_text(row):
+    # Handle potential NaN or None values in the row
+    summary = str(row.get("T_Zusammenfassung", "")) if pd.notna(row.get("T_Zusammenfassung")) else ""
+    diagnoses = str(row.get("Diagnosen", "")) if pd.notna(row.get("Diagnosen")) else ""
+    clinical = str(row.get("T_KlinBef", "")) if pd.notna(row.get("T_KlinBef")) else ""
+    findings = str(row.get("T_Befunde", "")) if pd.notna(row.get("T_Befunde")) else ""
+    return "\n".join([summary, diagnoses, clinical, findings]).strip()
+if __name__ == "__main__":
+    # Read CSV file ONLY inside main block
+    df = pd.read_csv(INPUT_CSV, sep=';')
+    results = []
+    # Process each row
+    for idx, row in df.iterrows():
+        print(f"Processing row {idx + 1}/{len(df)}")
+        try:
+            patient_text = build_patient_text(row)
+            result = run_inference(patient_text)
+            # Add unique_id and MedDatum to result for tracking
+            result["unique_id"] = row.get("unique_id", f"row_{idx}")
+            result["MedDatum"] = row.get("MedDatum", None)
+            results.append(result)
+            print(json.dumps(result, indent=2, ensure_ascii=False))
+        except Exception as e:
+            print(f"Error processing row {idx}: {e}")
+            results.append({
+                "success": False,
+                "error": str(e),
+                "unique_id": row.get("unique_id", f"row_{idx}"),
+                "MedDatum": row.get("MedDatum", None)
+            })
+    # Save results to a JSON file
+    output_json = INPUT_CSV.replace(".csv", "_results_total.json")
+    with open(output_json, 'w', encoding='utf-8') as f:
+        json.dump(results, f, indent=2, ensure_ascii=False)
+    print(f"Results saved to {output_json}")
+
+