cache-pg

Highlights from the README

• No new infrastructure. If you run Postgres, you already have the cache store. One table, two indexes, Migrate() creates them.
• Durable. Cached entries survive process restarts (unlike an in-memory cache) — useful for expensive-to-recompute values and idempotency keys.
• Driver-agnostic. Takes a *sql.DB. Bring pgx, lib/pq, or — for tests and embedded use — SQLite. The conformance suite runs in-process on SQLite with no Docker.
• Codegen-free. Plain database/sql, no ORM, no generated client. Easy to audit, easy to vendor.
• Atomic where it matters. SetNX, Incr, Decr, and SetMulti run inside transactions so concurrent callers stay correct.

Code from the README

package main

import (
	"context"
	"database/sql"
	"log"
	"time"

	_ "github.com/jackc/pgx/v5/stdlib"
	pgcache "github.com/ubgo/cache-pg"
)

func main() {
	db, err := sql.Open("pgx", "postgres://localhost:5432/app?sslmode=disable")
	if err != nil {
		log.Fatal(err)
	}

	c := pgcache.New(db)            // Postgres dialect by default
	if err := c.Migrate(context.Background()); err != nil { // idempotent
		log.Fatal(err)
	}
	defer c.Close()

	ctx := context.Background()
	_ = c.Set(ctx, "report:q3", []byte("…"), time.Hour)

	// Reclaim physically-dead rows on a ticker / cron:
	_ = c.Vacuum(ctx)
}

go func() {
	t := time.NewTicker(10 * time.Minute)
	for range t.C {
		_ = c.Vacuum(context.Background())
	}
}()

From the cache-pg README.

At a glance (from the README)