Delete latentcanon--histvis.py"

latentcanon--histvis.py/"

@@ -1,100 +0,0 @@
-from datasets import load_dataset
-import matplotlib.pyplot as plt
-import pandas as pd
-
-# Method 1: Load dataset using the custom loading script
-# This assumes you've placed histvis_loading.py in your repo
-histvis = load_dataset("./", script_files=["histvis_loading.py"])
-print(f"Dataset loaded with {len(histvis['train'])} examples")
-
-# Method 2: Load directly from Hugging Face (once the loading script is added to the repo)
-# histvis = load_dataset("latentcanon/HistVis")
-
-# Display dataset info
-print("\nDataset Features:")
-print(histvis['train'].features)
-
-# Show some basic statistics
-print("\nHistorical periods in the dataset:")
-periods = histvis['train'].unique('historical_period')
-print(periods)
-
-print("\nModels in the dataset:")
-models = histvis['train'].unique('model')
-print(models)
-
-print("\nCategories in the dataset:")
-categories = histvis['train'].unique('category')
-print(categories)
-
-# Let's see a specific example
-example = histvis['train'][0]
-print("\nExample entry:")
-for key, value in example.items():
-    if key != 'image':  # Skip printing the image data
-        print(f"{key}: {value}")
-
-# Display an image from the dataset
-plt.figure(figsize=(10, 10))
-plt.imshow(example['image'])
-plt.title(f"{example['historical_period']}: {example['universal_human_activity']}")
-plt.axis('off')
-plt.show()
-
-# Filter images by historical period and category
-filtered = histvis['train'].filter(lambda x: 
-    x['historical_period'] == '19th_century' and 
-    x['category'] == 'Art')
-print(f"\nFound {len(filtered)} images from 19th century in Art category")
-
-# Compare models for the same activity
-def compare_models_for_activity(dataset, activity, period=None):
-    """Display images from different models for the same activity"""
-    filter_fn = lambda x: x['universal_human_activity'] == activity
-    if period:
-        filter_fn = lambda x: x['universal_human_activity'] == activity and x['historical_period'] == period
-    
-    filtered = dataset.filter(filter_fn)
-    
-    if len(filtered) == 0:
-        print(f"No images found for activity: {activity}")
-        return
-    
-    # Get one example from each model
-    model_examples = {}
-    for item in filtered:
-        if item['model'] not in model_examples:
-            model_examples[item['model']] = item
-        
-        if len(model_examples) == 3:  # Assuming 3 models
-            break
-    
-    # Plot the examples
-    fig, axes = plt.subplots(1, len(model_examples), figsize=(15, 5))
-    
-    for i, (model, item) in enumerate(model_examples.items()):
-        if len(model_examples) == 1:
-            ax = axes
-        else:
-            ax = axes[i]
-        
-        ax.imshow(item['image'])
-        ax.set_title(f"Model: {model}")
-        ax.axis('off')
-    
-    period_str = f" in {period}" if period else ""
-    plt.suptitle(f"'{activity}'{period_str}")
-    plt.tight_layout()
-    plt.show()
-
-# Example usage
-compare_models_for_activity(histvis['train'], 'a person treating a wound with care', '21st_century')
-
-# Create a pandas DataFrame for easier analysis (excluding images)
-df = pd.DataFrame([
-    {k: v for k, v in example.items() if k != 'image'}
-    for example in histvis['train']
-])
-
-print("\nDataset summary:")
-print(df.describe(include='all'))