[DCP] Remove DB and Table Auto Creation from DCP Service + Modify the schema to match Ingestion Pipeline (#26)

* Removes the logic to create the DB and update the schema from the DCP service. The Ingestion Pipeline will be responsible for this moving forward.

* Also modify the Edge schema to match the new spanner schema from ingestion pipeline. Verified this works with retrieval

* Graph Service changes to reflect schema updates.

* Lint

Author: gmechali

packages/datacommons-api/datacommons_api/api_cli.py

@@ -18,7 +18,7 @@
 from datacommons_api.app import app
 from datacommons_api.core.config import get_config, initialize_config
 from datacommons_api.core.logging import get_logger, setup_logging
-from datacommons_db.session import get_session, initialize_db
+from datacommons_db.session import get_session
 from datacommons_api.services.graph_service import GraphService
 
 setup_logging()
@@ -56,16 +56,6 @@ def start(
         gcp_spanner_database_name=gcp_spanner_database_name,
     )
 
-    # Initialize the database
-    logger.info("Initializing database...")
-    logger.info("GCP Project ID: %s", config.GCP_PROJECT_ID)
-    logger.info("GCP Spanner Instance ID: %s", config.GCP_SPANNER_INSTANCE_ID)
-    logger.info("GCP Spanner Database Name: %s", config.GCP_SPANNER_DATABASE_NAME)
-    initialize_db(
-        config.GCP_PROJECT_ID,
-        config.GCP_SPANNER_INSTANCE_ID,
-        config.GCP_SPANNER_DATABASE_NAME,
-    )
     logger.info("Starting API server...")
     uvicorn.run(
         "datacommons_api.app:app",

packages/datacommons-db/datacommons_db/session.py

@@ -14,15 +14,13 @@
 
 import logging
 
-from sqlalchemy import Engine, create_engine, inspect
+from sqlalchemy import Engine, create_engine
 from sqlalchemy.orm import Session, sessionmaker
 
 from datacommons_db.models.base import Base
 
 logger = logging.getLogger(__name__)
 
-REQUIRED_TABLES = ["Edge", "Node", "Observation"]
-
 
 # DDL for Creating Property Graph
 DDL_PROPERTY_GRAPH = """
@@ -93,34 +91,3 @@ def get_session(project_id: str, instance_id: str, database_name: str) -> Sessio
     engine = get_engine(project_id, instance_id, database_name)
     session = sessionmaker(bind=engine)
     return session()
-
-
-def initialize_db(project_id: str, instance_id: str, database_name: str):
-    """Initialize the Spanner database.
-
-    Args:
-      project_id: GCP project ID
-      instance_id: Cloud Spanner instance ID
-      database_name: Cloud Spanner database name
-    """
-    engine = get_engine(project_id, instance_id, database_name)
-
-    # Check if database is empty by inspecting existing tables
-    inspector = inspect(engine)
-    existing_tables = inspector.get_table_names()
-
-    # Check if all required tables exist
-    missing_tables = [
-        table for table in REQUIRED_TABLES if table not in existing_tables
-    ]
-    if missing_tables:
-        logger.warning(
-            "Missing required tables in database %s: %s", database_name, missing_tables
-        )
-
-    # Only create tables if database is completely empty
-    if not existing_tables or missing_tables:
-        # Import all models so they are properly initialized with the call to Base.metadata.create_all
-        logger.info("Creating tables %s in database %s", REQUIRED_TABLES, database_name)
-        Base.metadata.create_all(engine)
-        create_property_graph(engine)