feat(tools): add pdf extraction for minio ingestion

src/tools/builtin/minio-ingest.ts

@@ -1,6 +1,9 @@
 import { promisify } from 'node:util';
 import { execFile } from 'node:child_process';
 import { extname } from 'node:path';
+import { mkdtempSync, rmSync } from 'node:fs';
+import { join } from 'node:path';
+import { tmpdir } from 'node:os';
 import type { BackupConfig } from '../../config/schema.js';
 import type { MemoryStore } from '../../memory/store.js';
 import type { Tool, ToolResult } from '../types.js';
@@ -12,7 +15,7 @@ type ExecRunner = (
   file: string,
   args: string[],
   options?: { env?: NodeJS.ProcessEnv; maxBuffer?: number },
-) => Promise<{ stdout: string; stderr: string }>;
+) => Promise<{ stdout: string | Buffer; stderr: string | Buffer }>;
 
 const TEXT_EXTENSIONS = new Set([
   '.txt',
@@ -29,6 +32,9 @@ const TEXT_EXTENSIONS = new Set([
   '.html',
   '.htm',
 ]);
+const EXTRACTABLE_BINARY_EXTENSIONS = new Set([
+  '.pdf',
+]);
 
 export interface MinioIngestDeps {
   execRunner?: ExecRunner;
@@ -42,12 +48,40 @@ function isLikelyText(content: string): boolean {
 function isLikelyTextObject(objectKey: string): boolean {
   const ext = extname(objectKey).toLowerCase();
   if (!ext) {return true;}
-  return TEXT_EXTENSIONS.has(ext);
+  return TEXT_EXTENSIONS.has(ext) || EXTRACTABLE_BINARY_EXTENSIONS.has(ext);
+}
+
+function isExtractableBinaryObject(objectKey: string): boolean {
+  return EXTRACTABLE_BINARY_EXTENSIONS.has(extname(objectKey).toLowerCase());
+}
+
+async function readObjectText(
+  runner: ExecRunner,
+  remotePath: string,
+  objectKey: string,
+  env: NodeJS.ProcessEnv,
+): Promise<string> {
+  if (!isExtractableBinaryObject(objectKey)) {
+    const { stdout } = await runner('mc', ['cat', remotePath], { env, maxBuffer: 20 * 1024 * 1024 });
+    return toText(stdout);
+  }
+
+  const tempDir = mkdtempSync(join(tmpdir(), 'flynn-minio-ingest-'));
+  const localPath = join(tempDir, 'object.bin');
+  try {
+    await runner('mc', ['cp', remotePath, localPath], { env, maxBuffer: 20 * 1024 * 1024 });
+    const { stdout } = await runner('pdftotext', ['-q', localPath, '-'], { maxBuffer: 20 * 1024 * 1024 });
+    return toText(stdout);
+  } finally {
+    rmSync(tempDir, { recursive: true, force: true });
+  }
 }
 
 export const minioIngestInternals = {
   isLikelyText,
   isLikelyTextObject,
+  isExtractableBinaryObject,
+  readObjectText,
 };
 
 interface MinioIngestArgs {
@@ -139,10 +173,9 @@ export function createMinioIngestTool(config: BackupConfig, store: MemoryStore,
       const remotePath = `${alias}/${bucket}/${objectKey}`;
 
       try {
-        const { stdout } = await runner('mc', ['cat', remotePath], { env, maxBuffer: 20 * 1024 * 1024 });
-        const text = typeof stdout === 'string' ? stdout : stdout.toString('utf-8');
+        const text = await readObjectText(runner, remotePath, objectKey, env);
 
-        if (!force && !isLikelyText(text)) {
+        if (!force && !isExtractableBinaryObject(objectKey) && !isLikelyText(text)) {
           return {
             success: false,
             output: '',
@@ -180,3 +213,6 @@ export function createMinioIngestTool(config: BackupConfig, store: MemoryStore,
     },
   };
 }
+function toText(value: string | Buffer): string {
+  return typeof value === 'string' ? value : value.toString('utf-8');
+}