Initial commit

internal/storage/db.go

@@ -0,0 +1,140 @@
+package storage
+
+import (
+	"database/sql"
+	"fmt"
+	"log/slog"
+	"os"
+	"path/filepath"
+	"strings"
+	"sync"
+
+	_ "modernc.org/sqlite"
+)
+
+var (
+	mu       sync.RWMutex
+	globalDB *sql.DB
+)
+
+// Init opens (or creates) the SQLite database and runs migrations.
+func Init(dbPath string) error {
+	mu.Lock()
+	defer mu.Unlock()
+
+	if globalDB != nil {
+		return nil
+	}
+
+	dir := filepath.Dir(dbPath)
+	if err := os.MkdirAll(dir, 0o755); err != nil {
+		return fmt.Errorf("create data dir: %w", err)
+	}
+
+	d, err := sql.Open("sqlite", dbPath+"?_pragma=journal_mode(WAL)&_pragma=busy_timeout(5000)&_pragma=foreign_keys(ON)")
+	if err != nil {
+		return fmt.Errorf("open database: %w", err)
+	}
+
+	d.SetMaxOpenConns(1)
+
+	if err := runMigrations(d); err != nil {
+		return fmt.Errorf("run migrations: %w", err)
+	}
+
+	globalDB = d
+	slog.Info("database initialized", "path", dbPath)
+	return nil
+}
+
+// Get returns the global database handle. Panics if Init was not called.
+func Get() *sql.DB {
+	mu.RLock()
+	db := globalDB
+	mu.RUnlock()
+	if db == nil {
+		panic("storage.Get() called before storage.Init()")
+	}
+	return db
+}
+
+// Close closes the global database handle.
+func Close() error {
+	mu.Lock()
+	defer mu.Unlock()
+	if globalDB != nil {
+		err := globalDB.Close()
+		globalDB = nil
+		return err
+	}
+	return nil
+}
+
+func runMigrations(d *sql.DB) error {
+	if _, err := d.Exec(schema); err != nil {
+		return fmt.Errorf("create schema: %w", err)
+	}
+
+	// Idempotent column adds for DBs created before the dedup columns existed.
+	// SQLite errors with "duplicate column name" when the column is already there;
+	// we swallow that specifically.
+	addColumnIfMissing(d, "stories", "url_canonical", "TEXT")
+	addColumnIfMissing(d, "stories", "headline_norm", "TEXT")
+	addColumnIfMissing(d, "post_log", "url_canonical", "TEXT")
+
+	// FTS5 virtual tables don't support IF NOT EXISTS reliably.
+	// Check sqlite_master before creating.
+	var ftsExists int
+	d.QueryRow(`SELECT COUNT(*) FROM sqlite_master WHERE type='table' AND name='stories_fts'`).Scan(&ftsExists)
+	if ftsExists == 0 {
+		if _, err := d.Exec(ftsSchema); err != nil {
+			return fmt.Errorf("create FTS5 table: %w", err)
+		}
+		if _, err := d.Exec(ftsTriggers); err != nil {
+			return fmt.Errorf("create FTS5 triggers: %w", err)
+		}
+		slog.Info("created FTS5 search index")
+	}
+
+	return nil
+}
+
+// RunMaintenance prunes stale data. Called periodically.
+func RunMaintenance(recentWindowHours, classificationLogDays int) {
+	recentCutoff := nowUnix() - int64(recentWindowHours*3600)
+	exec("prune recent_headlines",
+		`DELETE FROM recent_headlines WHERE seen_at < ?`, recentCutoff)
+
+	logCutoff := nowUnix() - int64(classificationLogDays*86400)
+	exec("prune classification_log",
+		`DELETE FROM classification_log WHERE logged_at < ?`, logCutoff)
+
+	// Prune old stories (30 days) and their post logs / reactions
+	storyCutoff := nowUnix() - int64(30*86400)
+	exec("prune old stories",
+		`DELETE FROM stories WHERE seen_at < ? AND classified = 1`, storyCutoff)
+	exec("prune old post_log",
+		`DELETE FROM post_log WHERE posted_at < ?`, storyCutoff)
+	exec("prune old reactions",
+		`DELETE FROM reactions WHERE reacted_at < ?`, storyCutoff)
+
+	exec("wal checkpoint", "PRAGMA wal_checkpoint(TRUNCATE)")
+	exec("optimize", "PRAGMA optimize")
+}
+
+// exec is a fire-and-forget helper that logs errors.
+func exec(label, query string, args ...any) {
+	if _, err := Get().Exec(query, args...); err != nil {
+		slog.Error("db exec failed", "op", label, "err", err)
+	}
+}
+
+func addColumnIfMissing(d *sql.DB, table, column, columnType string) {
+	q := fmt.Sprintf("ALTER TABLE %s ADD COLUMN %s %s", table, column, columnType)
+	if _, err := d.Exec(q); err != nil {
+		// SQLite returns "duplicate column name" when the column already exists.
+		if !strings.Contains(err.Error(), "duplicate column name") {
+			slog.Error("alter table failed", "table", table, "column", column, "err", err)
+		}
+	}
+}