small improvement to checking for image attachments

backend/onyx/connectors/confluence/connector.py

@@ -373,11 +373,24 @@ def _fetch_page_attachments(
             cql=attachment_query,
             expand=",".join(_ATTACHMENT_EXPANSION_FIELDS),
         ):
-            attachment["metadata"].get("mediaType", "")
+            media_type: str = attachment.get("metadata", {}).get("mediaType", "")
+
+            # TODO(rkuo): this check is partially redundant with validate_attachment_filetype
+            # and checks in convert_attachment_to_content/process_attachment
+            # but doing the check here avoids an unnecessary download. Due for refactoring.
+            if not self.allow_images:
+                if media_type.startswith("image/"):
+                    logger.info(
+                        f"Skipping attachment because allow images is False: {attachment['title']}"
+                    )
+                    continue
+
             if not validate_attachment_filetype(
                 attachment,
             ):
-                logger.info(f"Skipping attachment: {attachment['title']}")
+                logger.info(
+                    f"Skipping attachment because it is not an accepted file type: {attachment['title']}"
+                )
                 continue
 
             logger.info(f"Processing attachment: {attachment['title']}")

backend/onyx/connectors/confluence/utils.py

@@ -34,7 +34,11 @@
 from onyx.db.pg_file_store import create_populate_lobj
 from onyx.db.pg_file_store import save_bytes_to_pgfilestore
 from onyx.db.pg_file_store import upsert_pgfilestore
-from onyx.file_processing.extract_file_text import extract_file_text
+from onyx.file_processing.extract_file_text import (
+    OnyxExtensionType,
+    extract_file_text,
+    is_accepted_file_ext,
+)
 from onyx.file_processing.file_validation import is_valid_image_type
 from onyx.file_processing.image_utils import store_image_and_create_section
 from onyx.utils.logger import setup_logger
@@ -59,16 +63,17 @@ def validate_attachment_filetype(
     """
     Validates if the attachment is a supported file type.
     """
-    attachment.get("metadata", {})
     media_type = attachment.get("metadata", {}).get("mediaType", "")
-
     if media_type.startswith("image/"):
         return is_valid_image_type(media_type)
 
     # For non-image files, check if we support the extension
     title = attachment.get("title", "")
     extension = Path(title).suffix.lstrip(".").lower() if "." in title else ""
-    return extension in ["pdf", "doc", "docx", "txt", "md", "rtf"]
+
+    return is_accepted_file_ext(
+        "." + extension, OnyxExtensionType.Plain | OnyxExtensionType.Document
+    )
 
 
 class AttachmentProcessingResult(BaseModel):