원 레포랑 완전 분리

workspace/utils/image_converter.py

@@ -0,0 +1,35 @@
+import io
+from pathlib import Path
+from typing import List
+
+import httpx
+
+
+async def prepare_images_from_file(
+    file_url: str, filename: str, max_pages: int = 5, dpi: int = 180
+) -> List[bytes]:
+    """presigned URL → bytes. PDF이면 앞쪽 max_pages 페이지만 이미지로 변환하여 bytes 리스트 반환"""
+    async with httpx.AsyncClient() as client:
+        resp = await client.get(file_url, timeout=None)
+        resp.raise_for_status()
+        file_bytes = resp.content
+
+    ext = Path(filename).suffix.lower()
+    if ext in [".pdf", ".tif", ".tiff"]:
+        try:
+            from pdf2image import convert_from_bytes
+        except ImportError as e:
+            raise RuntimeError(
+                "pdf2image가 필요합니다. `pip install pdf2image poppler-utils`"
+            ) from e
+
+        pil_images = convert_from_bytes(file_bytes, dpi=dpi)
+        images = []
+        for i, im in enumerate(pil_images[:max_pages]):
+            buf = io.BytesIO()
+            im.save(buf, format="PNG")
+            images.append(buf.getvalue())
+        return images
+    else:
+        # 단일 이미지
+        return [file_bytes]