PictSure 0.1.0__tar.gz → 0.1.1__tar.gz

{pictsure-0.1.0 → pictsure-0.1.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: PictSure
-Version: 0.1.0
+Version: 0.1.1
 Summary: A package for generalized image classification using In-Context-Learning with PyTorch.
 Author-email: Cornelius Wolff <cornelius.wolff@cwi.nl>, Lukas Schiesser <lukas.schiesser@dfki.de>
 License: MIT License
@@ -58,17 +58,8 @@ PictSure is a deep learning library designed for **in-context learning** using i
 - **Easy-to-use CLI**: Manage models and weights through a simple command-line interface.
 
 ## Installation
-1. Clone this repository
 ```bash
-git clone https://git.ni.dfki.de/pictsure/pictsure-library
-```
-2. Navigate into the folder
-```bash
-cd pictsure-library
-```
-3. Install the pip package
-```bash
-pip install .
+pip install PictSure
 ```
 
 ## Quick Start
@@ -76,15 +67,10 @@ pip install .
 from PictSure import PictSure
 import torch
 
-# Initialize the model (using ViT as an example)
-model = PictSure(
-    embedding='vit',  # or 'resnet'
-    pretrained=True,  # use pretrained weights
-    device='cuda'     # or 'cpu'
-)
+DEVICE = "cpu" # or cuda, mps
 
-# you can also pull our pre-trained models from Huggingface
 model = PictSure.from_pretrained("pictsure/pictsure-vit")
+model = model.to(DEVICE)
 
 # Set your reference images and labels
 model.set_context_images(reference_images, reference_labels)
@@ -93,25 +79,6 @@ model.set_context_images(reference_images, reference_labels)
 predictions = model.predict(new_images)
 ```
 
-## Command Line Interface
-PictSure comes with a command-line interface to manage models and weights:
-
-### List Available Models
-```bash
-pictsure list-models
-```
-This command shows all available models, their status (downloaded/not downloaded), and detailed information about each model.
-
-### Remove Model Weights
-```bash
-pictsure remove <model_name> [--force]
-```
-Remove the weights of a specific model. Available models are:
-- `ViTPreAll`: ViT-based model
-- `ResPreAll`: ResNet-based model
-
-Use the `--force` or `-f` flag to skip the confirmation prompt.
-
 ## Examples
 For a complete working example, check out the Jupyter notebook in the Examples directory:
 ```bash

{pictsure-0.1.0 → pictsure-0.1.1}/PictSure.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: PictSure
-Version: 0.1.0
+Version: 0.1.1
 Summary: A package for generalized image classification using In-Context-Learning with PyTorch.
 Author-email: Cornelius Wolff <cornelius.wolff@cwi.nl>, Lukas Schiesser <lukas.schiesser@dfki.de>
 License: MIT License
@@ -58,17 +58,8 @@ PictSure is a deep learning library designed for **in-context learning** using i
 - **Easy-to-use CLI**: Manage models and weights through a simple command-line interface.
 
 ## Installation
-1. Clone this repository
 ```bash
-git clone https://git.ni.dfki.de/pictsure/pictsure-library
-```
-2. Navigate into the folder
-```bash
-cd pictsure-library
-```
-3. Install the pip package
-```bash
-pip install .
+pip install PictSure
 ```
 
 ## Quick Start
@@ -76,15 +67,10 @@ pip install .
 from PictSure import PictSure
 import torch
 
-# Initialize the model (using ViT as an example)
-model = PictSure(
-    embedding='vit',  # or 'resnet'
-    pretrained=True,  # use pretrained weights
-    device='cuda'     # or 'cpu'
-)
+DEVICE = "cpu" # or cuda, mps
 
-# you can also pull our pre-trained models from Huggingface
 model = PictSure.from_pretrained("pictsure/pictsure-vit")
+model = model.to(DEVICE)
 
 # Set your reference images and labels
 model.set_context_images(reference_images, reference_labels)
@@ -93,25 +79,6 @@ model.set_context_images(reference_images, reference_labels)
 predictions = model.predict(new_images)
 ```
 
-## Command Line Interface
-PictSure comes with a command-line interface to manage models and weights:
-
-### List Available Models
-```bash
-pictsure list-models
-```
-This command shows all available models, their status (downloaded/not downloaded), and detailed information about each model.
-
-### Remove Model Weights
-```bash
-pictsure remove <model_name> [--force]
-```
-Remove the weights of a specific model. Available models are:
-- `ViTPreAll`: ViT-based model
-- `ResPreAll`: ResNet-based model
-
-Use the `--force` or `-f` flag to skip the confirmation prompt.
-
 ## Examples
 For a complete working example, check out the Jupyter notebook in the Examples directory:
 ```bash

{pictsure-0.1.0 → pictsure-0.1.1}/PictSure.egg-info/SOURCES.txt

@@ -3,7 +3,6 @@ README.md
 pyproject.toml
 Examples/example.py
 PictSure/__init__.py
-PictSure/cli.py
 PictSure/model_PictSure.py
 PictSure/model_ViT.py
 PictSure/model_embeddings.py

{pictsure-0.1.0 → pictsure-0.1.1}/README.md

@@ -14,17 +14,8 @@ PictSure is a deep learning library designed for **in-context learning** using i
 - **Easy-to-use CLI**: Manage models and weights through a simple command-line interface.
 
 ## Installation
-1. Clone this repository
 ```bash
-git clone https://git.ni.dfki.de/pictsure/pictsure-library
-```
-2. Navigate into the folder
-```bash
-cd pictsure-library
-```
-3. Install the pip package
-```bash
-pip install .
+pip install PictSure
 ```
 
 ## Quick Start
@@ -32,15 +23,10 @@ pip install .
 from PictSure import PictSure
 import torch
 
-# Initialize the model (using ViT as an example)
-model = PictSure(
-    embedding='vit',  # or 'resnet'
-    pretrained=True,  # use pretrained weights
-    device='cuda'     # or 'cpu'
-)
+DEVICE = "cpu" # or cuda, mps
 
-# you can also pull our pre-trained models from Huggingface
 model = PictSure.from_pretrained("pictsure/pictsure-vit")
+model = model.to(DEVICE)
 
 # Set your reference images and labels
 model.set_context_images(reference_images, reference_labels)
@@ -49,25 +35,6 @@ model.set_context_images(reference_images, reference_labels)
 predictions = model.predict(new_images)
 ```
 
-## Command Line Interface
-PictSure comes with a command-line interface to manage models and weights:
-
-### List Available Models
-```bash
-pictsure list-models
-```
-This command shows all available models, their status (downloaded/not downloaded), and detailed information about each model.
-
-### Remove Model Weights
-```bash
-pictsure remove <model_name> [--force]
-```
-Remove the weights of a specific model. Available models are:
-- `ViTPreAll`: ViT-based model
-- `ResPreAll`: ResNet-based model
-
-Use the `--force` or `-f` flag to skip the confirmation prompt.
-
 ## Examples
 For a complete working example, check out the Jupyter notebook in the Examples directory:
 ```bash

{pictsure-0.1.0 → pictsure-0.1.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "PictSure"
-version = "0.1.0"
+version = "0.1.1"
 description = "A package for generalized image classification using In-Context-Learning with PyTorch."
 readme = "README.md"
 requires-python = ">=3.9"
@@ -33,4 +33,8 @@ dependencies = [
 [project.scripts]
 pictsure = "PictSure.cli:cli"
 
-[tool.setuptools.packages.find]
+[tool.setuptools.packages.find]
+
+[tool.pdm.publish]
+username = "__token__"
+password = ""

pictsure-0.1.0/PictSure/cli.py

@@ -1,84 +0,0 @@
-import os
-import click
-import shutil
-from .config import PRETRAINED
-import pkg_resources
-
-@click.group()
-def cli():
-    """PictSure command line interface."""
-    pass
-
-@cli.command()
-def list_models():
-    """List all available and downloaded models."""
-    click.echo("Available PictSure Models:")
-    click.echo("=" * 50)
-    
-    # Get the package directory
-    package_dir = os.path.dirname(pkg_resources.resource_filename('PictSure', '__init__.py'))
-    
-    for model_name, model_info in PRETRAINED.items():
-        # Check if model is downloaded using absolute path
-        local_folder = os.path.join(package_dir, 'weights', model_info['name'])
-        weights_path = os.path.join(local_folder, 'weights.pt')
-        is_downloaded = os.path.exists(weights_path)
-        
-        # Create status indicator
-        status = "✓ Downloaded" if is_downloaded else "✗ Not downloaded"
-        status_color = "green" if is_downloaded else "red"
-        
-        # Print model information
-        click.echo(f"\nModel: {click.style(model_info['name'], bold=True)}")
-        click.echo(f"Status: {click.style(status, fg=status_color)}")
-        click.echo(f"Type: {model_info['embed_model']}")
-        click.echo(f"Resolution: {model_info['resolution']}")
-        click.echo(f"Number of classes: {model_info['num_classes']}")
-        click.echo(f"Transformer heads: {model_info['nheads']}")
-        click.echo(f"Transformer layers: {model_info['nlayer']}")
-        click.echo(f"Model size: {model_info['size']} Million Parameters")
-        click.echo(f"Path: {weights_path}")
-        click.echo("-" * 50)
-
-@cli.command()
-@click.argument('model_name', type=click.Choice([info['name'] for info in PRETRAINED.values()]))
-@click.option('--force', '-f', is_flag=True, help='Skip confirmation prompt')
-def remove(model_name, force):
-    """Remove the weights of a specific model."""
-    # Get the package directory
-    package_dir = os.path.dirname(pkg_resources.resource_filename('PictSure', '__init__.py'))
-    
-    # Find the model info by name
-    model_info = next((info for info in PRETRAINED.values() if info['name'] == model_name), None)
-    if not model_info:
-        click.echo(click.style(f"Model {model_name} not found.", fg='red'))
-        return
-    
-    # Construct paths
-    local_folder = os.path.join(package_dir, 'weights', model_info['name'])
-    weights_path = os.path.join(local_folder, 'weights.pt')
-    
-    if not os.path.exists(weights_path):
-        click.echo(click.style(f"Model {model_info['name']} is not downloaded.", fg='yellow'))
-        return
-    
-    if not force:
-        if not click.confirm(f"Are you sure you want to remove the weights for {click.style(model_info['name'], bold=True)}?"):
-            click.echo("Operation cancelled.")
-            return
-    
-    try:
-        # Remove the weights file
-        os.remove(weights_path)
-        # Try to remove the directory if it's empty
-        try:
-            os.rmdir(local_folder)
-        except OSError:
-            pass  # Directory might not be empty, which is fine
-        
-        click.echo(click.style(f"Successfully removed weights for {model_info['name']}.", fg='green'))
-    except Exception as e:
-        click.echo(click.style(f"Error removing weights: {str(e)}", fg='red'))
-
-if __name__ == '__main__':
-    cli()