Add bots files handling logic

src/functions/backend/db/bot-files/purge-storage.ts

@@ -0,0 +1,86 @@
+import BunSQLite from "@moduletrace/bun-sqlite";
+import type { BUN_SQLITE_COBBLE_BOT_FILES } from "@/db/types/db";
+import { rmSync } from "node:fs";
+import path from "node:path";
+import type { BotFileScope } from "@/src/dict/bot-file-dict";
+import grabDirNames from "@/src/utils/grab-dir-names";
+import { logError } from "@/src/functions/backend/logger";
+import grabStorageDir from "./grab-storage-dir";
+
+type Params = {
+    scope: BotFileScope;
+    owner_id: number | string;
+    /**
+     * Its files' `media` ids, read before the rows go. Left out, they are
+     * read here (only useful when the catalog rows still exist).
+     */
+    media_ids?: number[];
+};
+
+/**
+ * The `media` ids of a directory's files, read before the bot or workspace
+ * (and with it, the catalog rows) is deleted.
+ */
+export async function grabStorageMediaIds({ scope, owner_id }: { scope: BotFileScope; owner_id: number | string }) {
+    const res = await BunSQLite.sql<BUN_SQLITE_COBBLE_BOT_FILES>({
+        sql: `SELECT media_id FROM bot_files WHERE scope = ? AND owner_id = ?`,
+        values: [scope, owner_id],
+    });
+
+    return (res.payload || []).map((file) => Number(file.media_id)).filter((id) => id > 0);
+}
+
+/**
+ * Removes a directory's files for good when the bot or workspace it
+ * belongs to goes: the disk folder, every file's `media` row, and what is
+ * left of the catalog. A failure is logged, never thrown: deleting the
+ * bot must not stop on a stray file.
+ */
+export default async function purgeStorage({ scope, owner_id, media_ids }: Params) {
+    try {
+        const ids = media_ids || (await grabStorageMediaIds({ scope, owner_id }));
+
+        if (ids.length) {
+            const placeholders = ids.map(() => "?").join(", ");
+
+            await BunSQLite.sql({
+                sql: `DELETE FROM media_paradigms WHERE media_id IN (${placeholders})`,
+                values: ids,
+            }).catch(() => {});
+
+            await BunSQLite.sql({
+                sql: `DELETE FROM media WHERE id IN (${placeholders})`,
+                values: ids,
+            }).catch(() => {});
+        }
+
+        await BunSQLite.sql({
+            sql: `DELETE FROM bot_file_vectors WHERE scope = ? AND owner_id = ?`,
+            values: [scope, owner_id],
+        }).catch(() => {});
+
+        /**
+         * Usually gone already, with the bot or workspace (the cascade).
+         */
+        await BunSQLite.sql({
+            sql: `DELETE FROM bot_files WHERE scope = ? AND owner_id = ?`,
+            values: [scope, owner_id],
+        });
+
+        const dir = grabStorageDir({ scope, owner_id, create: false });
+        const { DATA_DIR } = grabDirNames();
+
+        /**
+         * Never outside the data volume, whatever the path says.
+         */
+        if (path.resolve(dir).startsWith(path.resolve(DATA_DIR))) {
+            rmSync(dir, { recursive: true, force: true });
+        }
+    } catch (error) {
+        await logError({
+            title: `${scope} ${owner_id}'s files couldn't be cleaned up`,
+            error,
+            metadata: { scope, owner_id },
+        }).catch(() => {});
+    }
+}