In the website I'm making with Flask, I'm trying to "upgrade" my search bar by using the pg_trgm PostgreSQL extension to match words even after being spelled wrong. However, I'm having trouble figuring out how to use the extension's custom functions alongside SQLAlchemy.
Will I have to use raw SQL to properly perform the queries, or is there a way to do this more cleanly?
Assuming the extension is configured correctly, these functions should be accessible like any other database function. For example, here's how the similarity function might be called using SQLAlchemy core:
import sqlalchemy as sa
engine = sa.create_engine('postgresql+psycopg2:///test', echo=True, future=True)
metadata = sa.MetaData()
tbl = sa.Table('t70546926', metadata,
sa.Column('c1', sa.String, primary_key=True),
sa.Column('c2', sa.String))
tbl.drop(engine, checkfirst=True)
tbl.create(engine)
ins = sa.insert(tbl).values(c1='Alice', c2='Alison')
with engine.begin() as conn:
conn.execute(ins)
query = sa.select(sa.func.similarity(tbl.c.c1, tbl.c.c2).label('similarity_result'))
with engine.connect() as conn:
rows = conn.execute(query)
for row in rows:
print(row.similarity_result)
For Flask-SQLAlchemy, you might do something like this:
result = dbsession.query(sa.func.similarity(val1, val2)).all()
sa.func.SIMILARITY