package main

import (
	"context"
	"embed"
	"fmt"
	"log/slog"
	"sort"

	"github.com/jackc/pgx/v5/pgxpool"
)

//go:embed migrations/*.sql
var migrationFS embed.FS

// migrate applies every migrations/*.sql not yet recorded, in filename order, each in its own
// transaction. Applied names are the record — a file that changes after it ran is not re-applied.
func migrate(ctx context.Context, pool *pgxpool.Pool) error {
	_, err := pool.Exec(ctx, `create table if not exists schema_migrations (
		name       text primary key,
		applied_at timestamptz not null default now()
	)`)
	if err != nil {
		return fmt.Errorf("create schema_migrations: %w", err)
	}

	applied := map[string]bool{}
	rows, err := pool.Query(ctx, `select name from schema_migrations`)
	if err != nil {
		return fmt.Errorf("read schema_migrations: %w", err)
	}
	defer rows.Close()
	for rows.Next() {
		var name string
		if err := rows.Scan(&name); err != nil {
			return err
		}
		applied[name] = true
	}
	if err := rows.Err(); err != nil {
		return err
	}
	rows.Close()

	entries, err := migrationFS.ReadDir("migrations")
	if err != nil {
		return err
	}
	names := make([]string, 0, len(entries))
	for _, e := range entries {
		names = append(names, e.Name())
	}
	sort.Strings(names)

	for _, name := range names {
		if applied[name] {
			continue
		}
		sql, err := migrationFS.ReadFile("migrations/" + name)
		if err != nil {
			return err
		}
		tx, err := pool.Begin(ctx)
		if err != nil {
			return err
		}
		if _, err := tx.Exec(ctx, string(sql)); err != nil {
			tx.Rollback(ctx)
			return fmt.Errorf("migration %s: %w", name, err)
		}
		if _, err := tx.Exec(ctx, `insert into schema_migrations (name) values ($1)`, name); err != nil {
			tx.Rollback(ctx)
			return err
		}
		if err := tx.Commit(ctx); err != nil {
			return fmt.Errorf("migration %s: %w", name, err)
		}
		slog.Info("migration applied", "ctx", "startup", "name", name)
	}
	return nil
}

// sweep runs the startup cleanup from docs/spec.md §4.6. An in-process conversion goroutine dies
// with the process, so without this those rows say "converting" forever.
func sweep(ctx context.Context, pool *pgxpool.Pool) error {
	tag, err := pool.Exec(ctx, `update submissions
		set status = 'failed', status_msg = 'interrupted by restart'
		where status in ('queued', 'downloading', 'converting')`)
	if err != nil {
		return err
	}
	if n := tag.RowsAffected(); n > 0 {
		slog.Warn("submissions interrupted by restart", "ctx", "startup", "count", n)
	}

	// ponytail: temp files of swept submissions are unlinked with the row in step 3, once the
	// pipeline exists and there is something to unlink.
	if _, err := pool.Exec(ctx,
		`delete from submissions where created_at < now() - interval '7 days'`); err != nil {
		return err
	}
	if _, err := pool.Exec(ctx, `delete from sessions where expires_at < now()`); err != nil {
		return err
	}

	var failed int
	err = pool.QueryRow(ctx, `select count(*) from submissions where status = 'failed'`).Scan(&failed)
	if err != nil {
		return err
	}
	if failed > 0 {
		// A failed submission is invisible to everyone but its submitter, so a broken pipeline
		// has no other way of announcing itself.
		slog.Warn("failed submissions present", "ctx", "startup", "count", failed)
	}
	return nil
}