Restart alembic versions with the table creation

[madhav165]

🐛 Bug-fix PR

📌 Summary

Identified the right way to do alembic migration. First migration script to have table creation commands.

💡 Fix Description

• Base.metadata.create_all() replaced with command.upgrade(cfg, "head") set in gunicorn.config.py.
• Also fixes tests that create multiple randomly named databases

🧪 Verification

Check	Command	Status
Lint suite	make lint	pass
Unit tests	make test	pass

📐 MCP Compliance (if relevant)

• Matches current MCP spec
• No breaking change to MCP clients

✅ Checklist

• Code formatted (make black isort pre-commit)
• No secrets/credentials committed

[madhav165]

Having alembic migration on gunicorn startup works, tested with empty SQLite and PostgreSQL, however not optimal for CI/CD. Need to explore how to do it better for CI/CD.

[crivetimihai]

Fix the migration to be idempotent

In the migration file, wrap column additions in try/except or check existence

from sqlalchemy import inspect

def upgrade():
    bind = op.get_bind()
    inspector = inspect(bind)
    
    # Check if column exists before adding
    columns = [col['name'] for col in inspector.get_columns('gateways')]
    if 'slug' not in columns:
        op.add_column("gateways", sa.Column("slug", sa.String(), nullable=True))

Getting this:

sqlalchemy.exc.OperationalError: (sqlite3.OperationalError) duplicate column name: slug
[SQL: ALTER TABLE gateways ADD COLUMN slug VARCHAR]
(Background on this error at: https://sqlalche.me/e/20/e3q8)