[fix]: fixed pdf parsing

src/__init__.py

@@ -7,9 +7,13 @@ from sentence_transformers import SentenceTransformer
 import torch
 import chromadb
 from src.utils import OpenAIEmbedder, LocalEmbedder
+from src.ghm import initialize_models
 
 load_dotenv()
 
+# Initialize models and configurations at startup
+initialize_models()
+
 def sanitize_html(raw):
     # allow only text and basic tags
     return bleach.clean(raw, tags=[], strip=True)
@@ -68,5 +72,3 @@ def get_embedder():
     else:
         logger.info(f"Using local embedder with model: {EmbeddingConfig.TEXT_MODEL}")
         return LocalEmbedder(model_name=EmbeddingConfig.TEXT_MODEL)
-
-    

src/ghm.py

@@ -1,10 +1,9 @@
 import json
 import os
 
+from loguru import logger
 import requests
 from huggingface_hub import snapshot_download
-from utils import logger
-
 
 def download_json(url):
     response = requests.get(url)
@@ -28,13 +27,16 @@ def download_and_modify_json(url, local_filename, modifications):
         json.dump(data, f, ensure_ascii=False, indent=4)
 
 
-if __name__ == '__main__':
-
+def initialize_models():
+    """
+    Downloads and configures all necessary models and settings.
+    This function is designed to be called once at application startup.
+    """
     mineru_patterns = [
         # "models/Layout/LayoutLMv3/*",
         "models/Layout/YOLO/*",
-        # "models/MFD/YOLO/*",
-        # "models/MFR/unimernet_hf_small_2503/*",
+        "models/MFD/YOLO/*",
+        "models/MFR/unimernet_hf_small_2503/*",
         "models/OCR/paddleocr_torch/*",
         # "models/TabRec/TableMaster/*",
         # "models/TabRec/StructEqTable/*",
@@ -69,3 +71,7 @@ if __name__ == '__main__':
 
     download_and_modify_json(json_url, config_file, json_mods)
     logger.info(f'The configuration file has been configured successfully, the path is: {config_file}')
+
+
+if __name__ == '__main__':
+    initialize_models()