feat(tools): add pdf extraction for minio ingestion

src/tools/builtin/minio-sync.ts

@@ -12,7 +12,7 @@ type ExecRunner = (
   file: string,
   args: string[],
   options?: { env?: NodeJS.ProcessEnv; maxBuffer?: number },
-) => Promise<{ stdout: string; stderr: string }>;
+) => Promise<{ stdout: string | Buffer; stderr: string | Buffer }>;
 
 interface MinioSyncArgs {
   prefix: string;
@@ -168,13 +168,9 @@ export function createMinioSyncTool(config: BackupConfig, store: MemoryStore, de
           }
 
           const remotePath = `${alias}/${bucket}/${key}`;
-          const { stdout: objectStdout } = await runner('mc', ['cat', remotePath], {
-            env,
-            maxBuffer: 20 * 1024 * 1024,
-          });
-          const text = typeof objectStdout === 'string' ? objectStdout : objectStdout.toString('utf-8');
+          const text = await minioIngestInternals.readObjectText(runner, remotePath, key, env);
 
-          if (!force && !minioIngestInternals.isLikelyText(text)) {
+          if (!force && !minioIngestInternals.isExtractableBinaryObject(key) && !minioIngestInternals.isLikelyText(text)) {
             skipped++;
             continue;
           }