Files
gh-christianlouis-quizzical…/migrations/add_spotify_oauth_columns.py

128 lines
6.5 KiB
Python

import sys
import os
import contextlib
from sqlalchemy import text
# Add project root to Python path
project_root = os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))
if project_root not in sys.path:
sys.path.insert(0, project_root)
from musicround import db, create_app
from musicround.models import User # Import your User model
# Migration name (used for logging and tracking)
migration_name = os.path.splitext(os.path.basename(__file__))[0]
def run_migration():
# Check if an app context already exists (e.g., if called from within Flask app)
from flask import current_app
try:
app = current_app._get_current_object()
app_context_needed = False
except Exception:
app = create_app()
app_context_needed = True
context_manager = app.app_context() if app_context_needed else contextlib.nullcontext()
with context_manager:
try:
print(f"Starting migration: {migration_name}")
# Use raw SQL for schema changes to avoid issues with model definitions
# that might already expect the columns to exist. # Check if spotify_id column exists
inspector = db.inspect(db.engine)
columns = [col['name'] for col in inspector.get_columns('user')]
# COLUMN NAMING STRATEGY EXPLANATION:
# ==================================
# This migration adds 'spotify_id' as the standardized column name for Spotify user IDs.
# The chosen strategy is to use 'spotify_id' consistently throughout the application
# for all Spotify-related user identification, rather than a generic 'oauth_id'.
#
# Reasoning:
# 1. Consistency: All OAuth provider columns follow the pattern '{provider}_id'
# (e.g., google_id, authentik_id, dropbox_id, spotify_id)
# 2. Clarity: 'spotify_id' explicitly indicates this field stores Spotify user IDs
# 3. Maintainability: Future developers can immediately understand the purpose
# 4. Extensibility: Allows for multiple OAuth providers without column name conflicts
#
# This approach avoids generic 'oauth_id' which could be ambiguous when supporting
# multiple OAuth providers. Each provider gets its own dedicated ID column.
if 'spotify_id' in columns and 'spotify_id' not in columns:
# NOTE: This condition will never be true - kept for historical reference
# If there was ever an 'oauth_id' column that needed renaming to 'spotify_id',
# this would be the place to handle it. However, we've chosen to implement
# 'spotify_id' from the start for clarity and consistency.
print("Column 'spotify_id' exists. It will be kept for now. Adding 'spotify_id'.")
if 'spotify_id' not in columns:
print("Adding column 'spotify_id' to 'user' table.")
with db.engine.connect() as connection:
connection.execute(text('ALTER TABLE user ADD COLUMN spotify_id VARCHAR(100)'))
connection.commit()
print("Added 'spotify_id'.")
else:
print("Column 'spotify_id' already exists.")
if 'spotify_token' not in columns:
print("Adding column 'spotify_token' to 'user' table.")
with db.engine.connect() as connection:
connection.execute(text('ALTER TABLE user ADD COLUMN spotify_token TEXT'))
connection.commit()
print("Added 'spotify_token'.")
else:
print("Column 'spotify_token' already exists.")
if 'spotify_refresh_token' not in columns:
print("Adding column 'spotify_refresh_token' to 'user' table.")
with db.engine.connect() as connection:
connection.execute(text('ALTER TABLE user ADD COLUMN spotify_refresh_token TEXT'))
connection.commit()
print("Added 'spotify_refresh_token'.")
else:
print("Column 'spotify_refresh_token' already exists.")
if 'spotify_token_expiry' not in columns:
print("Adding column 'spotify_token_expiry' to 'user' table.")
with db.engine.connect() as connection:
connection.execute(text('ALTER TABLE user ADD COLUMN spotify_token_expiry DATETIME'))
connection.commit()
print("Added 'spotify_token_expiry'.")
else:
print("Column 'spotify_token_expiry' already exists.")
# Add index to spotify_id if it doesn't exist
# Index creation syntax can vary between DBs (e.g., SQLite vs PostgreSQL)
# For SQLite, it's generally: CREATE INDEX IF NOT EXISTS idx_user_spotify_id ON user (spotify_id);
# For SQLAlchemy, it's better to define this in the model and let Alembic/Flask-Migrate handle it,
# but for a manual script:
try:
with db.engine.connect() as connection:
# Check if index exists first (SQLite specific query)
result = connection.execute(text("SELECT name FROM sqlite_master WHERE type='index' AND tbl_name='user' AND name='idx_user_spotify_id';")).fetchone()
if not result:
print("Adding index 'idx_user_spotify_id' to 'user.spotify_id'.")
connection.execute(text('CREATE UNIQUE INDEX idx_user_spotify_id ON user (spotify_id) WHERE spotify_id IS NOT NULL;'))
connection.commit()
print("Added unique index 'idx_user_spotify_id'.")
else:
print("Index 'idx_user_spotify_id' already exists.")
except Exception as e:
print(f"Could not create index on spotify_id (this might be okay if using a different DB or if it exists): {e}")
print(f"Migration {migration_name} completed successfully.")
return True
except Exception as e:
print(f"Error during migration {migration_name}: {e}")
# db.session.rollback() # Not needed with raw SQL execution and individual commits
return False
if __name__ == '__main__':
# This allows running the migration script directly
# Ensure your Flask app and db are initialized correctly
if not run_migration():
sys.exit(1)