Update download_assets.py

download_assets.py

@@ -2,21 +2,14 @@ from huggingface_hub import hf_hub_download
 import os
 
 def download_assets():
-    """Download necessary assets from Hugging Face Hub"""
-    # Create directories if they don't exist
-    os.makedirs('data', exist_ok=True)
-    os.makedirs('vectorstore/db_faiss', exist_ok=True)
+    """Download necessary assets from Hugging Face Hub."""
+    os.makedirs("data", exist_ok=True)
+    os.makedirs("vectorstore/db_faiss", exist_ok=True)
 
-    # Dataset repository ID (type=dataset)
-    repo_id = "MoizK/mindmedic-assets"
+    repo_id   = "MoizK/mindmedic-assets"
     repo_type = "dataset"
+    token     = os.getenv("HUGGINGFACE_API_TOKEN")  # optional
 
-    # Pull your token from the env
-    token = os.getenv("HUGGINGFACE_API_TOKEN")
-    if not token:
-        raise RuntimeError("Please set HUGGINGFACE_HUB_TOKEN in your env")
-
-    # Download PDF files
     pdf_files = [
         "71763-gale-encyclopedia-of-medicine.-vol.-1.-2nd-ed.pdf",
         "Depression-NIM-2024.pdf",
@@ -24,36 +17,36 @@ def download_assets():
         "Doing-What-Matters-in-Times-of-Stress.pdf",
         "Generalized-Anxiety-Disorder-When-Worry-Gets-Out-of-Control.pdf",
         "WHO-mhGAP-Intervention-Guide-v2.pdf",
-        "social-anxiety-disorder-more-than-just-shyness.pdf"
+        "social-anxiety-disorder-more-than-just-shyness.pdf",
     ]
 
-    for pdf in pdf_files:
+    for fname in pdf_files:
         try:
-            path = hf_hub_download(
+            hf_hub_download(
                 repo_id=repo_id,
                 repo_type=repo_type,
-                filename=f"data/{pdf}",
+                filename=f"data/{fname}",
                 local_dir=".",
-                use_auth_token=token
+                local_dir_use_symlinks=False,
+                token=token,
             )
-            print(f"Downloaded {pdf} → {path}")
+            print(f"Downloaded {fname}")
         except Exception as e:
-            print(f"Error downloading {pdf}: {e}")
+            print(f"⚠️ Failed to download {fname}: {e}")
 
-    # Download FAISS index files
-    index_files = ["index.faiss", "index.pkl"]
-    for idx in index_files:
+    for idx in ("index.faiss", "index.pkl"):
         try:
-            path = hf_hub_download(
+            hf_hub_download(
                 repo_id=repo_id,
                 repo_type=repo_type,
                 filename=f"vectorstore/db_faiss/{idx}",
                 local_dir=".",
-                use_auth_token=token
+                local_dir_use_symlinks=False,
+                token=token,
             )
-            print(f"Downloaded {idx} → {path}")
+            print(f"Downloaded {idx}")
         except Exception as e:
-            print(f"Error downloading {idx}: {e}")
+            print(f"⚠️ Failed to download {idx}: {e}")
 
 if __name__ == "__main__":
     download_assets()