Download fixes

app.py

@@ -1,20 +1,16 @@
 import os
-import shutil
 import subprocess
 import signal
 os.environ["GRADIO_ANALYTICS_ENABLED"] = "False"
 import gradio as gr
+import tempfile
 
-from huggingface_hub import create_repo, HfApi
-from huggingface_hub import snapshot_download
-from huggingface_hub import whoami
-from huggingface_hub import ModelCard
-
+from huggingface_hub import HfApi, ModelCard, whoami
 from gradio_huggingfacehub_search import HuggingfaceHubSearch
-
+from pathlib import Path
+from textwrap import dedent
 from apscheduler.schedulers.background import BackgroundScheduler
 
-from textwrap import dedent
 
 HF_TOKEN = os.environ.get("HF_TOKEN")
 
@@ -110,21 +106,25 @@ def process_model(model_id, q_method, use_imatrix, imatrix_q_method, private_rep
             else "*.bin"
         )
 
-        dl_pattern += pattern
-
-        api.snapshot_download(repo_id=model_id, local_dir=model_name, local_dir_use_symlinks=False, allow_patterns=dl_pattern)
-        print("Model downloaded successfully!")
-        print(f"Current working directory: {os.getcwd()}")
-        print(f"Model directory contents: {os.listdir(model_name)}")
-
-        conversion_script = "convert_hf_to_gguf.py"
-        fp16_conversion = f"python llama.cpp/{conversion_script} {model_name} --outtype f16 --outfile {fp16}"
-        result = subprocess.run(fp16_conversion, shell=True, capture_output=True)
-        print(result)
-        if result.returncode != 0:
-            raise Exception(f"Error converting to fp16: {result.stderr}")
-        print("Model converted to fp16 successfully!")
-        print(f"Converted model path: {fp16}")
+        dl_pattern += [pattern]
+
+        with tempfile.TemporaryDirectory(dir="downloads") as tmpdir:
+            # Keep the model name as the dirname so the model name metadata is populated correctly
+            local_dir = Path(tmpdir)/model_name
+            print(local_dir)
+            api.snapshot_download(repo_id=model_id, local_dir=local_dir, local_dir_use_symlinks=False, allow_patterns=dl_pattern)
+            print("Model downloaded successfully!")
+            print(f"Current working directory: {os.getcwd()}")
+            print(f"Model directory contents: {os.listdir(local_dir)}")
+
+            conversion_script = "convert_hf_to_gguf.py"
+            fp16_conversion = f"python llama.cpp/{conversion_script} {local_dir} --outtype f16 --outfile {fp16}"
+            result = subprocess.run(fp16_conversion, shell=True, capture_output=True)
+            print(result)
+            if result.returncode != 0:
+                raise Exception(f"Error converting to fp16: {result.stderr}")
+            print("Model converted to fp16 successfully!")
+            print(f"Converted model path: {fp16}")
 
         imatrix_path = "llama.cpp/imatrix.dat"
 
@@ -257,9 +257,7 @@ def process_model(model_id, q_method, use_imatrix, imatrix_q_method, private_rep
         )
     except Exception as e:
         return (f"Error: {e}", "error.png")
-    finally:
-        shutil.rmtree(model_name, ignore_errors=True)
-        print("Folder cleaned up successfully!")
+
 
 css="""/* Custom CSS to allow scrolling */
 .gradio-container {overflow-y: auto;}