Update app.py

app.py

@@ -17,6 +17,7 @@ import requests
 from charset_normalizer import from_bytes
 import zipfile
 import tempfile
+import webbrowser
 
 # Custom Exception Class
 class GPUQuotaExceededError(Exception):
@@ -27,16 +28,20 @@ EMBEDDING_MODEL_NAME = "sentence-transformers/all-MiniLM-L6-v2"
 CHUNK_SIZE = 500
 BATCH_SIZE = 32
 
-# Persistent storage directories
+# Set Persistent Storage Path
 PERSISTENT_PATH = os.getenv("PERSISTENT_PATH", "/data")
 os.makedirs(PERSISTENT_PATH, exist_ok=True, mode=0o777)
 
+# Define Subdirectories
 TEMP_DIR = os.path.join(PERSISTENT_PATH, "temp")
 os.makedirs(TEMP_DIR, exist_ok=True, mode=0o777)
 
 OUTPUTS_DIR = os.path.join(PERSISTENT_PATH, "outputs")
 os.makedirs(OUTPUTS_DIR, exist_ok=True, mode=0o777)
 
+NPY_CACHE = os.path.join(PERSISTENT_PATH, "npy_cache")
+os.makedirs(NPY_CACHE, exist_ok=True, mode=0o777)
+
 LOG_DIR = os.getenv("LOG_DIR", os.path.join(PERSISTENT_PATH, "logs"))
 os.makedirs(LOG_DIR, exist_ok=True, mode=0o777)
 
@@ -142,8 +147,11 @@ def process_files(files):
         all_embeddings = []
         for i in range(0, len(all_chunks), BATCH_SIZE):
             batch = all_chunks[i:i+BATCH_SIZE]
-            embeddings = handle_gpu_operation(lambda: get_model().encode(batch)) if model else []
-            all_embeddings.extend(embeddings)
+            if model:
+                embeddings = handle_gpu_operation(lambda: model.encode(batch))
+                all_embeddings.extend(embeddings)
+            else:
+                return "Model not initialized. Please check model initialization.", "", ""
 
         # Save results to OUTPUTS_DIR
         embeddings_path = os.path.join(OUTPUTS_DIR, "embeddings.npy")
@@ -168,9 +176,8 @@ def process_files(files):
 def semantic_search(query, top_k=5):
     global model
     if model is None:
-        if not initialize_model():
-            return "Model initialization failed. Please try again."
-
+        return "Model not initialized. Please process files first."
+    
     try:
         # Load saved embeddings from OUTPUTS_DIR
         stored_embeddings = np.load(os.path.join(OUTPUTS_DIR, "embeddings.npy"))
@@ -182,7 +189,7 @@ def semantic_search(query, top_k=5):
 
         # Get query embedding
         if model:
-            query_embedding = handle_gpu_operation(lambda: get_model().encode([query]))[0]
+            query_embedding = model.encode([query])[0]
         else:
             return "Model not initialized. Please process files first."
 
@@ -216,20 +223,20 @@ def search_and_format(query, num_results):
 
 def browse_outputs():
     try:
-        # Attempt to open the OUTPUTS_DIR
-        os.startfile(OUTPUTS_DIR)
-        return "Opened outputs directory successfully"
+        webbrowser.open(f"file://{OUTPUTS_DIR}")
+        return "Opened outputs directory"
     except Exception as e:
         logger.error(f"Error opening file browser: {str(e)}")
         return "Error opening file browser"
 
-def download_results_from_disk():
+def download_results():
     try:
         output_files = [
             os.path.join(OUTPUTS_DIR, "embeddings.npy"),
             os.path.join(OUTPUTS_DIR, "chunks.txt")
         ]
         
+        # Create a temporary zip file
         with tempfile.TemporaryDirectory() as temp_dir:
             zip_path = os.path.join(temp_dir, "results.zip")
             with zipfile.ZipFile(zip_path, 'w') as zipf:
@@ -239,7 +246,7 @@ def download_results_from_disk():
             return zip_path
     except Exception as e:
         logger.error(f"Error creating download: {str(e)}")
-        return "Error creating download file"
+        return None
 
 def create_gradio_interface():
     with gr.Blocks() as demo:
@@ -282,16 +289,14 @@ def create_gradio_interface():
                 outputs=results_output
             )
 
-            # Download Results Button
-            download_results_button = gr.Button("⬇️ Download Results")
-            download_results_button.click(
-                fn=download_results_from_disk,
+            download_button = gr.Button("⬇️ Download Results")
+            download_button.click(
+                fn=download_results,
                 outputs=[gr.File(label="Download Results")]
             )
 
         with gr.Tab("Outputs"):
-            # Browse Outputs Button
-            browse_button = gr.Button("📁 Browse Outputs", variant="primary")
+            browse_button = gr.Button("📁 Browse Outputs")
             browse_button.click(
                 fn=browse_outputs,
                 outputs=None