app.py

#!/usr/bin/env python3
import gradio as gr
import os
import hashlib
import json
import traceback
import zipfile
import tempfile
import shutil
from pathlib import Path
import requests
import threading
import time
from typing import List, Dict, Optional, Tuple
from dataclasses import dataclass
from concurrent.futures import ThreadPoolExecutor, as_completed

# Global variables for progress tracking
upload_progress = {"current": 0, "total": 0, "status": "", "files_processed": [], "errors": [], "final_summary": ""}
upload_lock = threading.Lock()

@dataclass
class ProcessResult:
    filename: str
    status: str  # "skipped", "uploaded", "error"
    message: str
    file_hash: Optional[str] = None

def calculate_sha256(filepath: Path) -> str:
    """Calculate SHA256 hash of a file"""
    sha256_hash = hashlib.sha256()
    with open(filepath, "rb") as f:
        # Read in 100MB chunks for better performance
        for byte_block in iter(lambda: f.read(104857600), b""):
            sha256_hash.update(byte_block)
    return sha256_hash.hexdigest()

def check_hash_exists(file_hash: str) -> bool:
    """Check if file hash already exists in datadrones.com"""
    try:
        hash_request = requests.get(
            f"https://dl.datadrones.com/api/model/sha256sum/{file_hash}",
            timeout=10
        )
        return hash_request.status_code == 200
    except Exception as e:
        print(f"Error checking hash existence: {e}")
        return False

def find_by_hash(file_hash: str) -> Optional[Dict]:
    """Find metadata by hash from Civitai and other sources"""
    # Get Civitai API key from environment variable (HuggingFace Spaces secret)
    civitai_api_key = os.getenv("CIVITAI_API_KEY")
    
    header = {
        "Content-Type": "application/json",
    }
    
    # Only add Authorization header if API key is available
    if civitai_api_key:
        header["Authorization"] = f"Bearer {civitai_api_key}"
    else:
        print("⚠️ Warning: CIVITAI_API_KEY not found in environment variables")

    print(f"Retrieving metadata by hash {file_hash}")
    
    # Try Civitai first
    try:
        response = requests.get(
            f"https://civitai.com/api/v1/model-versions/by-hash/{file_hash}", 
            headers=header,
            timeout=15
        )
        if response.status_code == 200:
            civitai_data = {"civitai": response.json()}
            return civitai_data
    except Exception as e:
        print(f"Civitai API error: {e}")

    # Try civitaiarchive as fallback
    try:
        response = requests.get(f"https://civitaiarchive.com/api/sha256/{file_hash}", timeout=15)
        if response.status_code == 200:
            civitai_data = {"civitai": response.json()}
            return civitai_data
    except Exception as e:
        print(f"CivitaiArchive API error: {e}")
    
    return None

def submit_to_datadrones(model_path: Path, metadata: Dict) -> bool:
    """Submit file to datadrones.com"""
    image_url = None
    model_versions = None
    base_model = None
    tags = None
    model_type = None
    # Start with model name if available
    description = ""
    model_name = None

    try:
        print(f"🚀 Starting upload of {model_path.name} to datadrones.com...")
        
        model_name = (metadata.get("model_name")
                      or metadata.get("civitai").get("name")
                      or metadata.get("name"))

        # Add civitai description if available
        civitai = metadata.get("civitai", {})
        is_nsfw = civitai.get("nsfw", False)

        if civitai and "modelVersions" in civitai:
            model_versions = civitai.get("modelVersions")

        # Add image if available
        if civitai and "images" in civitai and len(civitai["images"]) > 0:
            image_url = civitai["images"][0].get("url")
        if not image_url and model_versions:
            # try in model versions
            image_url = model_versions[0]["images"][0].get("url")
        if image_url:
            description += f"\n\n![Preview]({image_url})"

        if civitai and "type" in civitai:
            model_type = civitai.get("type")

        # could be version id api
        if civitai and "model" in civitai:
            model = civitai["model"]
            model_type = model.get("type")
            is_nsfw = model.get("nsfw")
            model_name = model.get("name")
            model_description = model.get("description")
            tags = model.get("tags")
            if model_description:
                description += f"\n\n{model_description}"

        if model_name:
            description = f"{model_name} \n" + description
        if civitai and "description" in civitai:
            if description:
                description += f"\n\n{civitai['description']}"

        if not description:
            description = "Possibly deleted"

        if not tags and metadata.get("tags"):
            tags = ",".join(metadata.get("tags", []))
        if not tags and civitai and "tags" in civitai:
            tags = ",".join(civitai.get("tags", []))

        if civitai and "baseModel" in civitai:
            base_model = civitai.get("baseModel")
        if civitai and "modelVersions" in civitai:
            model_versions = civitai.get("modelVersions")
            if model_versions:
                base_model = model_versions[0]["baseModel"]
                if base_model == "Hunyuan Video":
                    base_model = "HunyuanVideo"

        # Prepare form data for submission
        data = {
            "description": description,
            "base_model": base_model if base_model else "Other",
            "tags": tags if tags else "",
            "model_type": model_type if model_type else "LoRA",
            "is_nsfw": is_nsfw,
        }

        print(f"📋 Upload data for {model_path.name}:")
        print(f"   - Model name: {model_name}")
        print(f"   - Model type: {data['model_type']}")
        print(f"   - Base model: {data['base_model']}")
        print(f"   - NSFW: {data['is_nsfw']}")
        print(f"   - Tags: {data['tags']}")
        print(f"   - Image URL: {image_url}")
        print(f"   - Description length: {len(data['description'])} chars")
        print(f"   - File size: {model_path.stat().st_size / (1024*1024):.1f} MB")

        # Submit to datadrones.com, bypass cloudflare
        with open(model_path, "rb") as f:
            files = {"file": f}
            headers = {'Host': 'up.datadrones.com'}
            
            print(f"🌐 Making POST request to https://up.datadrones.com/upload for {model_path.name}...")
            
            response = requests.post("https://up.datadrones.com/upload", files=files, data=data, headers=headers, timeout=300)
            
            print(f"📡 Response for {model_path.name}:")
            print(f"   - Status code: {response.status_code}")
            if response.status_code != 200:
                print(f"   - Response text: {response.text}")

        return response.status_code == 200

    except Exception as e:
        print(f"💥 Exception during upload of {model_path.name}: {e}")
        traceback.print_exc()
        return False

def extract_model_files(uploaded_files: List) -> List[Path]:
    """Extract model files from uploaded files, handling both direct files and zip archives"""
    model_files = []
    temp_dir = Path(tempfile.mkdtemp())
    
    # Supported model file extensions
    supported_extensions = {'.safetensors', '.pt', '.bin', '.ckpt'}
    
    for file_info in uploaded_files:
        file_path = Path(file_info.name)
        
        if file_path.suffix.lower() in supported_extensions:
            # Direct model file
            dest_path = temp_dir / file_path.name
            shutil.copy2(file_path, dest_path)
            model_files.append(dest_path)
        
        elif file_path.suffix.lower() == '.zip':
            # Extract zip and find model files
            try:
                with zipfile.ZipFile(file_path, 'r') as zip_ref:
                    zip_ref.extractall(temp_dir)
                
                # Find all model files in extracted content
                for extension in supported_extensions:
                    for extracted_file in temp_dir.rglob(f"*{extension}"):
                        model_files.append(extracted_file)
                    
            except Exception as e:
                print(f"Error extracting {file_path}: {e}")
    
    return model_files

def process_single_file(model_file: Path) -> ProcessResult:
    """Process a single model file"""
    try:
        print(f"\n🔍 Processing file: {model_file.name}")
        
        # Check file size (skip if over 4GB)
        file_size = model_file.stat().st_size
        if file_size > 4 * 1024 * 1024 * 1024:  # 4GB
            print(f"⏭️ Skipping {model_file.name} - over 4GB limit")
            return ProcessResult(
                filename=model_file.name,
                status="skipped",
                message="File over 4GB size limit"
            )

        # Calculate hash
        print(f"🔢 Calculating hash for {model_file.name}...")
        file_hash = calculate_sha256(model_file)
        print(f"📝 Hash: {file_hash}")
        
        # Check if already exists in datadrones
        print(f"🔍 Checking if {file_hash} already exists on datadrones.com...")
        if check_hash_exists(file_hash):
            print(f"⏭️ File {model_file.name} already exists on datadrones.com")
            return ProcessResult(
                filename=model_file.name,
                status="skipped",
                message="Already exists in datadrones.com",
                file_hash=file_hash
            )

        # Find metadata by hash
        print(f"🔍 Looking up metadata for {file_hash}...")
        metadata = find_by_hash(file_hash)
        
        if not metadata:
            print(f"❌ No metadata found for {model_file.name}")
            return ProcessResult(
                filename=model_file.name,
                status="error",
                message="No metadata found for this file",
                file_hash=file_hash
            )
        
        print(f"✅ Found metadata for {model_file.name}")

        # Submit to datadrones
        print(f"🚀 Attempting upload of {model_file.name} to datadrones.com...")
        if submit_to_datadrones(model_file, metadata):
            print(f"✅ Successfully uploaded {model_file.name} to datadrones.com")
            return ProcessResult(
                filename=model_file.name,
                status="uploaded",
                message="Successfully uploaded to datadrones.com",
                file_hash=file_hash
            )
        else:
            print(f"❌ Failed to upload {model_file.name} to datadrones.com")
            return ProcessResult(
                filename=model_file.name,
                status="error",
                message="Failed to upload to datadrones.com",
                file_hash=file_hash
            )

    except Exception as e:
        print(f"💥 Error processing {model_file.name}: {e}")
        traceback.print_exc()
        return ProcessResult(
            filename=model_file.name,
            status="error",
            message=f"Processing error: {str(e)}"
        )

def update_progress(current: int, total: int, status: str, file_result: ProcessResult = None, final_summary: str = None):
    """Update global progress tracking"""
    with upload_lock:
        upload_progress["current"] = current
        upload_progress["total"] = total
        upload_progress["status"] = status
        
        # Store final summary when processing is complete
        if final_summary:
            upload_progress["final_summary"] = final_summary
        
        if file_result:
            upload_progress["files_processed"].append({
                "filename": file_result.filename,
                "status": file_result.status,
                "message": file_result.message,
                "hash": file_result.file_hash
            })
            
            if file_result.status == "error":
                upload_progress["errors"].append(f"{file_result.filename}: {file_result.message}")

def process_files_async(uploaded_files: List) -> str:
    """Process uploaded files asynchronously"""
    try:
        print(f"\n🎬 Starting bulk upload process...")
        
        # Reset progress
        with upload_lock:
            upload_progress.update({
                "current": 0, 
                "total": 0, 
                "status": "Extracting files...", 
                "files_processed": [], 
                "errors": [],
                "final_summary": ""
            })

        # Extract model files
        print(f"📦 Extracting model files from uploaded content...")
        model_files = extract_model_files(uploaded_files)
        total_files = len(model_files)
        
        print(f"📋 Found {total_files} model files to process")
        for i, file in enumerate(model_files, 1):
            print(f"   {i}. {file.name} ({file.stat().st_size / (1024*1024):.1f} MB)")
        
        if total_files == 0:
            print("❌ No supported model files found")
            return "No supported model files (.safetensors, .pt, .bin, '.ckpt') found in uploaded content."

        update_progress(0, total_files, "Processing files...")

        # Process files with thread pool for better performance
        print(f"🔄 Processing {total_files} files with ThreadPoolExecutor...")
        results = []
        with ThreadPoolExecutor(max_workers=3) as executor:
            future_to_file = {
                executor.submit(process_single_file, file): file 
                for file in model_files
            }
            
            for i, future in enumerate(as_completed(future_to_file), 1):
                result = future.result()
                results.append(result)
                print(f"📊 Completed {i}/{total_files}: {result.filename} -> {result.status}")
                update_progress(i, total_files, f"Processed {i}/{total_files} files", result)

        # Generate summary
        uploaded_count = sum(1 for r in results if r.status == "uploaded")
        skipped_count = sum(1 for r in results if r.status == "skipped")
        error_count = sum(1 for r in results if r.status == "error")

        summary = f"""Processing Complete!

Total files: {total_files}
✅ Uploaded: {uploaded_count}
⏭️ Skipped: {skipped_count}
❌ Errors: {error_count}"""

        # Update progress with final summary
        update_progress(total_files, total_files, "Complete", None, summary)
        
        print(f"🎉 Bulk upload completed: {uploaded_count} uploaded, {skipped_count} skipped, {error_count} errors")
        
        # Cleanup temp files
        print(f"🧹 Cleaning up temporary files...")
        for file in model_files:
            try:
                if file.exists():
                    file.unlink()
                # Also cleanup parent temp directory if empty
                parent = file.parent
                if parent.exists() and not any(parent.iterdir()):
                    parent.rmdir()
            except:
                pass

        return summary

    except Exception as e:
        error_msg = f"Processing failed: {str(e)}"
        print(f"💥 Bulk processing failed: {e}")
        traceback.print_exc()
        update_progress(0, 0, error_msg, None, error_msg)
        return error_msg

def get_progress_update():
    """Get current progress status"""
    with upload_lock:
        if upload_progress["total"] == 0:
            return "No active uploads", ""
        
        current = upload_progress["current"]
        total = upload_progress["total"]
        status = upload_progress["status"]
        
        # Show final summary if processing is complete
        if current == total and total > 0 and "final_summary" in upload_progress:
            progress_text = upload_progress["final_summary"]
        else:
            progress_text = f"Progress: {current}/{total} - {status}"
        
        # Build detailed log
        log_lines = []
        for file_info in upload_progress["files_processed"][-10:]:  # Show last 10
            status_emoji = {"uploaded": "✅", "skipped": "⏭️", "error": "❌"}.get(file_info["status"], "?")
            log_lines.append(f"{status_emoji} {file_info['filename']}: {file_info['message']}")
        
        if upload_progress["errors"]:
            log_lines.append(f"\nRecent Errors ({len(upload_progress['errors'])}):")
            log_lines.extend(upload_progress["errors"][-5:])  # Show last 5 errors
        
        detailed_log = "\n".join(log_lines)
        
        return progress_text, detailed_log

def start_upload(files):
    """Start the upload process in a separate thread"""
    if not files:
        return "No files selected", ""
    
    # Start processing in background thread
    thread = threading.Thread(target=process_files_async, args=(files,))
    thread.daemon = True
    thread.start()
    
    return "Upload started! Check progress below...", ""

# Create Gradio interface
def create_interface():
    with gr.Blocks(title="DataDrones Bulk Uploader", theme=gr.themes.Soft()) as iface:
        gr.Markdown("""
        # 🚁 DataDrones Bulk Uploader
        
        Upload multiple model files (`.safetensors`, `.pt`, `.bin`, `.ckpt`) or zip archives containing model files to datadrones.com.
        
        **Features:**
        - Supports direct model file uploads (.safetensors, .pt, .bin, .ckpt) and zip archives
        - Automatic hash checking to avoid duplicates
        - Metadata retrieval from Civitai and other sources
        - Progress tracking (Hit refresh button)
        - Concurrent processing for faster uploads
        """)
        
        with gr.Row():
            with gr.Column(scale=2):
                file_input = gr.File(
                    label="Select model files (.safetensors, .pt, .bin, .ckpt) or .zip archives",
                    file_count="multiple",
                    file_types=[".safetensors", ".pt", ".bin", ".zip", ".ckpt"]
                )
                
                upload_btn = gr.Button("🚀 Start Upload", variant="primary", size="lg")
                
            with gr.Column(scale=1):
                gr.Markdown("""
                ### Instructions:
                1. Select multiple model files (`.safetensors`, `.pt`, `.bin`, `.ckpt`) directly, or
                2. Upload `.zip` archives containing model files
                3. Click "Start Upload" to begin processing
                4. Monitor progress by clicking refresh log. Don't browse away till its complete!
                
                **Note:** Single Files over 4GB will be skipped.
                """)
        
        gr.Markdown("---")
        
        with gr.Row():
            with gr.Column():
                progress_display = gr.Textbox(
                    label="Upload Progress",
                    value="Ready to upload",
                    interactive=False
                )
                
                refresh_btn = gr.Button("🔄 Refresh Progress", size="sm")
                
                detailed_log = gr.Textbox(
                    label="Detailed Log",
                    value="",
                    lines=15,
                    interactive=False
                )
        
        # Set up event handlers
        upload_btn.click(
            fn=start_upload,
            inputs=[file_input],
            outputs=[progress_display, detailed_log]
        )
        
        # Manual refresh for progress updates
        refresh_btn.click(
            fn=get_progress_update,
            outputs=[progress_display, detailed_log]
        )
    
    return iface

if __name__ == "__main__":
    app = create_interface()
    app.queue(max_size=10)  # Enable queuing for background processing
    app.launch(
        server_name="0.0.0.0",
        server_port=7860,
        share=False
    )