run_migrations

What it is

A small async script that runs pending SQL migrations by reading *.sql files from a local migrations/ directory and executing their SQL statements against a SQLAlchemy async engine.

Public API

• async def run_migrations()
  • Discovers migration files, parses SQL into statements, and executes them sequentially within a single async_engine.begin() transaction context.
  • Logs progress to stdout.
  • On execution errors:
    • If the error message contains "already exists" (case-insensitive), logs and skips.
    • Otherwise logs the error and continues with remaining statements/migrations.

Configuration/Dependencies

• Filesystem
  • Expects a directory at: Path(__file__).parent / "migrations"
  • Runs all *.sql files in lexicographic order (use filename prefixes to control order).
• Database
  • Requires app.core.database.async_engine to be configured and importable.
  • Uses SQLAlchemy text() and conn.execute(...) for each statement.
• SQL parsing rules
  • Splits statements when a non-empty, non-comment line ends with ;.
  • Skips full-line comments starting with -- and empty lines.
  • Removes inline comments after -- on a line.

Usage

Run as a script

python libs/naas-abi/naas_abi/apps/nexus/apps/api/run_migrations.py

Call from Python

import asyncio
from naas_abi.apps.nexus.apps.api.run_migrations import run_migrations

asyncio.run(run_migrations())

Caveats

• Statement splitting is simplistic:
  • Semicolons must appear at the end of a line to terminate a statement.
  • Does not handle semicolons inside strings/procedural blocks or more complex SQL constructs reliably.
• Errors other than “already exists” are logged but do not stop execution; failures may leave migrations partially applied.
• No migration tracking table/versioning is implemented; it re-runs all *.sql files each time (relying on idempotency or “already exists” handling).