app.py

import gradio as gr
from huggingface_hub import HfApi, CommitOperationDelete

def fetch_files(source_repo, hf_token):
    # Input validation
    if not hf_token.strip():
        yield gr.update(), gr.update(), "Error: Hugging Face Access Token is required.", []
        return
    if not source_repo.strip():
        yield gr.update(), gr.update(), "Error: Source repository name is required.", []
        return
    
    api = HfApi(token=hf_token.strip())
    
    try:
        # Initial status update before fetching
        yield gr.update(), gr.update(), f"Status: Fetching file list from '{source_repo}'...", []
        
        # Fetch the flat list of files from the repository
        files = api.list_repo_files(repo_id=source_repo.strip(), repo_type="model")
        files.sort()
        
        # Build a visual Linux-like tree representation for the CheckboxGroup
        choices = []
        for f in files:
            parts = f.split('/')
            if len(parts) == 1:
                display = f"📄 {f}"
            else:
                folder_path = "/".join(parts[:-1])
                display = f"📁 {folder_path}/ ├── 📄 {parts[-1]}"
            
            # Gradio allows (display_name, value) tuples for CheckboxGroup choices
            choices.append((display, f))
            
        yield (
            gr.update(visible=True), 
            gr.update(choices=choices, value=files), 
            f"Success: Fetched {len(files)} files. Uncheck the ones you do not want to copy.",
            files
        )
        
    except Exception as e:
        yield gr.update(visible=False), gr.update(), f"An error occurred: {str(e)}", []


def clone_repository(source_repo, target_repo, hf_token, selected_files, all_files):
    if not target_repo.strip():
        yield "Error: Target repository name is required."
        return
    
    api = HfApi(token=hf_token.strip())
    
    try:
        yield f"Status: Initiating server-side clone of '{source_repo}' to '{target_repo}'..."
        
        # 1. Duplicate the entire repo first (Server-side, almost instant, 0 disk space)
        api.duplicate_repo(
            from_id=source_repo.strip(),
            to_id=target_repo.strip(),
            repo_type="model"
        )
        
        # 2. Determine which files to delete based on what the user unchecked
        files_to_delete = set(all_files) - set(selected_files)
        
        if files_to_delete:
            yield f"Status: Deleting {len(files_to_delete)} unselected files from '{target_repo}'..."
            
            # Use 'path_in_repo' instead of 'path' to comply with the HF SDK
            operations = [CommitOperationDelete(path_in_repo=f) for f in files_to_delete]
            
            # Execute deletion in a single commit natively on HF servers
            api.create_commit(
                repo_id=target_repo.strip(),
                repo_type="model",
                operations=operations,
                commit_message="Clean up: Remove unselected files during selective clone"
            )
        
        yield f"Success! Selective clone completed. View your model here: https://huggingface.co/{target_repo.strip()}"
        
    except Exception as e:
        yield f"An error occurred: {str(e)}"


# Notice: The 'theme' parameter has been removed from Blocks to satisfy Gradio 6.0 specs
with gr.Blocks(title="Selective HF Model Cloner") as demo:
    gr.Markdown("# 📑 Selective Hugging Face Model Cloner")
    gr.Markdown("Fetch files from a model repository, uncheck what you don't need, and clone via high-speed server-side duplication.")
    
    with gr.Row():
        hf_token = gr.Textbox(
            label="Hugging Face Access Token", 
            placeholder="hf_...", 
            type="password",
            info="Requires 'Write' permissions."
        )
    
    with gr.Row():
        source_repo = gr.Textbox(
            label="Source Model Repository", 
            placeholder="e.g., meta-llama/Llama-3.2-1B"
        )
        fetch_btn = gr.Button("1. Fetch Files", variant="secondary")
        
    # We use a Gradio State to quietly keep track of the complete list of files fetched
    all_files_state = gr.State([])
    
    # This group remains hidden until files are successfully fetched
    with gr.Group(visible=False) as selection_group:
        gr.Markdown("### 🗂️ Repository Files")
        gr.Markdown("*All files are selected by default. Uncheck any files you do **not** want to carry over.*")
        
        file_checkboxes = gr.CheckboxGroup(label="Repository Tree", interactive=True)
        
        target_repo = gr.Textbox(
            label="Target Model Repository", 
            placeholder="e.g., your-username/my-custom-llama"
        )
        clone_btn = gr.Button("2. Clone Selected Files", variant="primary")
        
    output_logs = gr.Textbox(label="Execution Status / Logs", interactive=False)
    
    # Trigger Step 1: File Fetching
    fetch_btn.click(
        fn=fetch_files,
        inputs=[source_repo, hf_token],
        outputs=[selection_group, file_checkboxes, output_logs, all_files_state]
    )
    
    # Trigger Step 2: Cloning Process
    clone_btn.click(
        fn=clone_repository,
        inputs=[source_repo, target_repo, hf_token, file_checkboxes, all_files_state],
        outputs=output_logs
    )

if __name__ == "__main__":
    # The theme parameter is now passed into launch() per Gradio 6.0 rules
    demo.queue().launch(theme=gr.themes.Soft())