fix: improve error logging for index creation and add PARSE_IMAGE_REMOTE setting

application/api/user/routes.py

@@ -70,7 +70,7 @@ try:
     )
     users_collection.create_index("user_id", unique=True)
 except Exception as e:
-    current_app.logger.warning(f"Can't create indexes: {e}", )
+    print("Error creating indexes:", e)
 
 user = Blueprint("user", __name__)
 user_ns = Namespace("user", description="User related operations", path="/")

application/core/settings.py

@@ -30,6 +30,7 @@ class Settings(BaseSettings):
     }
     UPLOAD_FOLDER: str = "inputs"
     PARSE_PDF_AS_IMAGE: bool = False
+    PARSE_IMAGE_REMOTE: bool = False
     VECTOR_STORE: str = (
         "faiss"  #  "faiss" or "elasticsearch" or "qdrant" or "milvus" or "lancedb"
     )

application/parser/file/image_parser.py

@@ -8,6 +8,7 @@ import requests
 from typing import Dict, Union
 
 from application.parser.file.base_parser import BaseParser
+from application.core.settings import settings
 
 
 class ImageParser(BaseParser):
@@ -18,10 +19,13 @@ class ImageParser(BaseParser):
         return {}
 
     def parse_file(self, file: Path, errors: str = "ignore") -> Union[str, list[str]]:
-        doc2md_service = "https://llm.arc53.com/doc2md"
-        # alternatively you can use local vision capable LLM
-        with open(file, "rb") as file_loaded:
-            files = {'file': file_loaded}
-            response = requests.post(doc2md_service, files=files)   
-            data = response.json()["markdown"] 
+        if settings.PARSE_IMAGE_REMOTE:
+            doc2md_service = "https://llm.arc53.com/doc2md"
+            # alternatively you can use local vision capable LLM
+            with open(file, "rb") as file_loaded:
+                files = {'file': file_loaded}
+                response = requests.post(doc2md_service, files=files)   
+                data = response.json()["markdown"] 
+        else:
+            data = ""
         return data