Update streamlit_app.py

streamlit_app.py

@@ -3,6 +3,7 @@ import fitz  # PyMuPDF
 import torch
 import time
 import re
+import os
 import torch.nn.functional as F
 from transformers import (
     AutoTokenizer,
@@ -447,7 +448,10 @@ MODEL_OPTIONS = {
 }
 
 # FastText Configuration
-FASTTEXT_MODEL_PATH = "doc_classifier.bin"
+FASTTEXT_MODEL_PATH = os.path.join(os.path.dirname(__file__), "doc_classifier.bin")
+# Fallback to relative path if __file__ is not available
+if not os.path.exists(FASTTEXT_MODEL_PATH):
+    FASTTEXT_MODEL_PATH = "doc_classifier.bin"
 FASTTEXT_THRESHOLD = 0.45
 
 # Add FastText to model options if available
@@ -485,10 +489,17 @@ def load_model(model_name):
             st.error("FastText is not installed. Please install it using: pip install fasttext")
             return None, None
         try:
+            # Check if model file exists
+            if not os.path.exists(FASTTEXT_MODEL_PATH):
+                st.error(f"❌ FastText model file not found: {FASTTEXT_MODEL_PATH}\n\nPlease ensure 'doc_classifier.bin' is in the same directory as streamlit_app.py")
+                return None, None
             model = fasttext.load_model(FASTTEXT_MODEL_PATH)
             return None, model  # FastText doesn't need a tokenizer
         except FileNotFoundError:
-            st.error(f"FastText model file not found: {FASTTEXT_MODEL_PATH}")
+            st.error(f"❌ FastText model file not found: {FASTTEXT_MODEL_PATH}\n\nPlease ensure 'doc_classifier.bin' is in the repository root.")
+            return None, None
+        except Exception as e:
+            st.error(f"❌ Error loading FastText model: {str(e)}\n\nPlease check that the model file is valid.")
             return None, None
     else:
         tokenizer = AutoTokenizer.from_pretrained(model_name)