# Whisper Malayalam Fine-tuning Code
# Extracted from Colab session

# Cell 1
# ============================================================
# PUSH TRAINED MODEL TO HF HUB - FIXED!
# ============================================================

from huggingface_hub import notebook_login
from transformers import WhisperForConditionalGeneration, WhisperProcessor

# 1) Login
notebook_login()

# 2) Your details
YOUR_USERNAME = "kasimali"
MODEL_NAME = "whisper-small-malayalam"

# 3) Load from LOCAL folder (not repo!)
print("📦 Loading model from local folder...")
model = WhisperForConditionalGeneration.from_pretrained("./whisper-small-ml-final", local_files_only=True)
processor = WhisperProcessor.from_pretrained("./whisper-small-ml-final", local_files_only=True)

print("✅ Model loaded!")

# 4) Push to HF Hub
print(f"📤 Pushing to {YOUR_USERNAME}/{MODEL_NAME}...")

model.push_to_hub(f"{YOUR_USERNAME}/{MODEL_NAME}", use_auth_token=True)
processor.push_to_hub(f"{YOUR_USERNAME}/{MODEL_NAME}", use_auth_token=True)

print(f"✅ DONE!")
print(f"🔗 Your model: https://huggingface.co/{YOUR_USERNAME}/{MODEL_NAME}")


# Cell 2
# ============================================================
# UPLOAD MODEL FOLDER DIRECTLY TO HF HUB
# ============================================================

from huggingface_hub import HfApi, create_repo, notebook_login
import os

# 1) Login
notebook_login()

# 2) Config
YOUR_USERNAME = "kasimali"
MODEL_NAME = "whisper-small-malayalam"
LOCAL_FOLDER = "whisper-small-ml-final"  # Without ./

# 3) Create repo on HF Hub
api = HfApi()

print(f"📦 Creating repo: {YOUR_USERNAME}/{MODEL_NAME}...")

try:
    create_repo(
        repo_id=f"{YOUR_USERNAME}/{MODEL_NAME}",
        repo_type="model",
        private=False
    )
    print("✅ Repo created!")
except Exception as e:
    print(f"⚠️ Repo might already exist: {e}")

# 4) Upload entire folder
print(f"📤 Uploading {LOCAL_FOLDER} to Hub...")

api.upload_folder(
    folder_path=LOCAL_FOLDER,
    repo_id=f"{YOUR_USERNAME}/{MODEL_NAME}",
    repo_type="model",
)

print(f"\n✅ SUCCESS!")
print(f"🔗 Your model: https://huggingface.co/{YOUR_USERNAME}/{MODEL_NAME}")


# Cell 3
# Check what folders/files exist
import os

print("📁 Files in /content:")
for item in os.listdir("/content"):
    print(f"  - {item}")

print("\n📁 Looking for model folders...")
# Check common save locations
possible_folders = [
    "whisper-small-ml",
    "whisper-small-ml-final", 
    "./whisper-small-ml",
    "./whisper-small-ml-final"
]

for folder in possible_folders:
    if os.path.exists(folder):
        print(f"✅ Found: {folder}")
        print(f"   Contents: {os.listdir(folder)[:5]}")  # Show first 5 files
    else:
        print(f"❌ Not found: {folder}")


# Cell 4
from huggingface_hub import HfApi, create_repo, notebook_login

notebook_login()

YOUR_USERNAME = "kasimali"
SPACE_NAME = "malayalam-whisper-finetuning"

# Create Space
create_repo(
    repo_id=f"{YOUR_USERNAME}/{SPACE_NAME}",
    repo_type="space",
    space_sdk="gradio",
    private=False
)

# Create app.py that displays notebook
app_code = '''import gradio as gr

# Simple viewer
def show_notebook():
    return "Upload finetunning.ipynb to this Space to share your training code!"

demo = gr.Interface(
    fn=show_notebook,
    inputs=None,
    outputs=gr.Textbox(),
    title="Malayalam Whisper Fine-tuning Notebook"
)

demo.launch()
'''

# Write and upload
with open("app.py", "w") as f:
    f.write(app_code)

api = HfApi()
api.upload_file(
    path_or_fileobj="app.py",
    path_in_repo="app.py",
    repo_id=f"{YOUR_USERNAME}/{SPACE_NAME}",
    repo_type="space"
)

# Upload your notebook
from google.colab import files
files.download("finetunning.ipynb")

api.upload_file(
    path_or_fileobj="finetunning.ipynb",
    path_in_repo="finetunning.ipynb",
    repo_id=f"{YOUR_USERNAME}/{SPACE_NAME}",
    repo_type="space"
)

print(f"✅ Space: https://huggingface.co/spaces/{YOUR_USERNAME}/{SPACE_NAME}")


# Cell 5
from huggingface_hub import HfApi, create_repo, notebook_login

# Login
notebook_login()

# Configuration
USERNAME = "kasimali"
SPACE_NAME = "whisper-malayalam-finetuning"
SPACE_ID = f"{USERNAME}/{SPACE_NAME}"

# Create Space
try:
    create_repo(repo_id=SPACE_ID, repo_type="space", space_sdk="static")
    print(f"Created Space: {SPACE_ID}")
except:
    print("Space already exists")

# Get notebook from Colab
import json
from google.colab import _message
notebook = _message.blocking_request('get_ipynb', timeout_sec=10)
notebook_content = json.dumps(notebook, indent=2)

# Save notebook locally
with open("FINETUNINNG.ipynb", "w") as f:
    f.write(notebook_content)

# Upload to Space
api = HfApi()
api.upload_file(
    path_or_fileobj="FINETUNINNG.ipynb",
    path_in_repo="FINETUNINNG.ipynb",
    repo_id=SPACE_ID,
    repo_type="space"
)

# Create README
readme = f"""---
title: Whisper Malayalam Fine-tuning
emoji: 🎤
colorFrom: blue
colorTo: green
sdk: static
---

# Whisper Malayalam Fine-tuning

This Space contains the training notebook for fine-tuning Whisper on Malayalam language.

## Files
- FINETUNINNG.ipynb: Training notebook

## Usage
Download the notebook and run it in Google Colab or Jupyter.
"""

with open("README.md", "w") as f:
    f.write(readme)

api.upload_file(
    path_or_fileobj="README.md",
    path_in_repo="README.md",
    repo_id=SPACE_ID,
    repo_type="space"
)

print(f"Done. Visit: https://huggingface.co/spaces/{SPACE_ID}")


# Cell 6
from huggingface_hub import HfApi, create_repo, notebook_login
import os
import glob

# Login once
notebook_login()

# Configuration
USERNAME = "kasimali"

# Find all notebook files
print("Searching for notebook files...")
notebook_files = []

# Search in common locations
search_paths = [
    "/content/*.ipynb",
    "/content/**/*.ipynb",
]

for pattern in search_paths:
    notebook_files.extend(glob.glob(pattern, recursive=True))

# Remove duplicates
notebook_files = list(set(notebook_files))

print(f"\nFound {len(notebook_files)} notebook(s):")
for i, nb in enumerate(notebook_files):
    print(f"{i+1}. {nb}")

if len(notebook_files) == 0:
    print("\nNo notebooks found! Make sure you have .ipynb files in /content/")
else:
    print(f"\nCreating {len(notebook_files)} separate Spaces...")
    
    api = HfApi()
    created_spaces = []
    
    for notebook_path in notebook_files:
        # Get notebook name without extension
        notebook_name = os.path.basename(notebook_path).replace(".ipynb", "")
        
        # Clean name for Space (remove special chars, lowercase)
        space_name = notebook_name.lower().replace(" ", "-").replace("_", "-")
        space_id = f"{USERNAME}/{space_name}"
        
        print(f"\n{'='*60}")
        print(f"Processing: {notebook_name}")
        print(f"Space ID: {space_id}")
        
        try:
            # Create Space
            create_repo(
                repo_id=space_id,
                repo_type="space",
                space_sdk="static",
                exist_ok=True
            )
            print(f"✅ Space created/exists")
            
            # Upload notebook
            api.upload_file(
                path_or_fileobj=notebook_path,
                path_in_repo=os.path.basename(notebook_path),
                repo_id=space_id,
                repo_type="space",
                commit_message=f"Upload {notebook_name}"
            )
            print(f"✅ Notebook uploaded")
            
            # Create README
            readme = f"""---
title: {notebook_name}
emoji: 📓
colorFrom: blue
colorTo: green
sdk: static
---

# {notebook_name}

This Space contains the notebook: **{os.path.basename(notebook_path)}**

## Usage
Download the notebook and run it in Google Colab or Jupyter.

## Files
- {os.path.basename(notebook_path)}
"""
            
            # Save and upload README
            readme_path = f"/tmp/README_{space_name}.md"
            with open(readme_path, "w") as f:
                f.write(readme)
            
            api.upload_file(
                path_or_fileobj=readme_path,
                path_in_repo="README.md",
                repo_id=space_id,
                repo_type="space",
                commit_message="Add README"
            )
            print(f"✅ README added")
            
            created_spaces.append({
                'name': notebook_name,
                'url': f"https://huggingface.co/spaces/{space_id}"
            })
            
        except Exception as e:
            print(f"❌ Error: {e}")
    
    # Summary
    print(f"\n{'='*60}")
    print(f"SUMMARY: Created {len(created_spaces)} Spaces")
    print(f"{'='*60}")
    
    for space in created_spaces:
        print(f"\n📓 {space['name']}")
        print(f"   🔗 {space['url']}")
    
    print(f"\n✅ All done!")


# Cell 7
from huggingface_hub import HfApi, create_repo, notebook_login
import json

# Login
notebook_login()

# Configuration
USERNAME = "kasimali"
SPACE_NAME = "whisper-malayalam-code"
SPACE_ID = f"{USERNAME}/{SPACE_NAME}"

# Get all code from current session
print("Extracting code from Colab session...")

# Get execution history
from IPython import get_ipython
ipython = get_ipython()

# Get all executed code
all_code = []
for i, cell in enumerate(ipython.user_ns.get('In', [])):
    if cell and cell.strip():
        all_code.append(f"# Cell {i}\n{cell}\n")

# Combine all code
full_code = "\n\n".join(all_code)

# Save as Python file
print("Creating app.py...")
app_content = f"""# Whisper Malayalam Fine-tuning Code
# Extracted from Colab session

{full_code}
"""

with open("app.py", "w") as f:
    f.write(app_content)

print(f"✅ Created app.py ({len(full_code)} characters)")

# Create Space
print(f"\nCreating Space: {SPACE_ID}...")
try:
    create_repo(repo_id=SPACE_ID, repo_type="space", space_sdk="static", exist_ok=True)
    print("✅ Space created")
except Exception as e:
    print(f"Space exists: {e}")

# Upload
api = HfApi()

print("\nUploading app.py...")
api.upload_file(
    path_or_fileobj="app.py",
    path_in_repo="app.py",
    repo_id=SPACE_ID,
    repo_type="space",
    commit_message="Upload code from Colab"
)

# Create requirements.txt
requirements = """datasets==3.1.0
transformers
accelerate
evaluate
jiwer
torch
"""

with open("requirements.txt", "w") as f:
    f.write(requirements)

api.upload_file(
    path_or_fileobj="requirements.txt",
    path_in_repo="requirements.txt",
    repo_id=SPACE_ID,
    repo_type="space",
    commit_message="Add requirements"
)

# Create README
readme = f"""---
title: Whisper Malayalam Fine-tuning Code
emoji: 🎤
colorFrom: blue
colorTo: green
sdk: static
---

# Whisper Malayalam Fine-tuning Code

This Space contains the Python code for fine-tuning Whisper on Malayalam.

## Files
- `app.py`: Main training code (extracted from Colab)
- `requirements.txt`: Python dependencies

## Usage


# Cell 8
from huggingface_hub import HfApi, create_repo, notebook_login

# Login
notebook_login()

# Configuration
USERNAME = "kasimali"
SPACE_NAME = "whisper-malayalam-code"
SPACE_ID = f"{USERNAME}/{SPACE_NAME}"

# Get all code from current session
print("Extracting code from Colab session...")

from IPython import get_ipython
ipython = get_ipython()

# Get all executed code
all_code = []
for i, cell in enumerate(ipython.user_ns.get('In', [])):
    if cell and cell.strip():
        all_code.append(f"# Cell {i}\n{cell}\n")

# Combine all code
full_code = "\n\n".join(all_code)

# Save as Python file
print("Creating app.py...")
app_content = "# Whisper Malayalam Fine-tuning Code\n"
app_content += "# Extracted from Colab session\n\n"
app_content += full_code

with open("app.py", "w") as f:
    f.write(app_content)

print(f"Created app.py ({len(full_code)} characters)")

# Create Space
print(f"\nCreating Space: {SPACE_ID}...")
try:
    create_repo(repo_id=SPACE_ID, repo_type="space", space_sdk="static", exist_ok=True)
    print("Space created")
except:
    print("Space already exists")

# Upload
api = HfApi()

print("\nUploading app.py...")
api.upload_file(
    path_or_fileobj="app.py",
    path_in_repo="app.py",
    repo_id=SPACE_ID,
    repo_type="space",
    commit_message="Upload code from Colab"
)

# Create requirements.txt
requirements = "datasets==3.1.0\ntransformers\naccelerate\nevaluate\njiwer\ntorch\n"

with open("requirements.txt", "w") as f:
    f.write(requirements)

api.upload_file(
    path_or_fileobj="requirements.txt",
    path_in_repo="requirements.txt",
    repo_id=SPACE_ID,
    repo_type="space",
    commit_message="Add requirements"
)

# Create README
readme_content = "---\n"
readme_content += "title: Whisper Malayalam Code\n"
readme_content += "emoji: 🎤\n"
readme_content += "colorFrom: blue\n"
readme_content += "colorTo: green\n"
readme_content += "sdk: static\n"
readme_content += "---\n\n"
readme_content += "# Whisper Malayalam Fine-tuning Code\n\n"
readme_content += "Python code for fine-tuning Whisper on Malayalam.\n\n"
readme_content += "## Files\n"
readme_content += "- app.py: Main training code\n"
readme_content += "- requirements.txt: Dependencies\n\n"
readme_content += "## Usage\n"
readme_content += "pip install -r requirements.txt\n"
readme_content += "python app.py\n"

with open("README.md", "w") as f:
    f.write(readme_content)

api.upload_file(
    path_or_fileobj="README.md",
    path_in_repo="README.md",
    repo_id=SPACE_ID,
    repo_type="space",
    commit_message="Add README"
)

print("\nSUCCESS!")
print(f"Uploaded: app.py, requirements.txt, README.md")
print(f"\nView at: https://huggingface.co/spaces/{SPACE_ID}")