another path fix for HF

app.py

@@ -72,25 +72,32 @@ def main():
     )
     
     st.title("SongLift LyrGen2 - AI Lyrics Generator")
-    print("===== Application Startup at", datetime.now().strftime("%Y-%m-%d %H:%M:%S"), "=====\n")
     
-    # Run environment tests if in HuggingFace
-    if Settings.is_huggingface():
-        print("\n=== Running Environment Tests ===")
-        tests_passed = test_environment()
-        if not tests_passed:
-            st.error("⚠️ Environment tests failed. Check the logs for details.")
+    # Only run startup once per session
+    if 'initialized' not in st.session_state:
+        print("===== Application Startup at", datetime.now().strftime("%Y-%m-%d %H:%M:%S"), "=====\n")
+        
+        # Run environment tests if in HuggingFace
+        if Settings.is_huggingface():
+            print("\n=== Running Environment Tests ===")
+            tests_passed = test_environment()
+            if not tests_passed:
+                st.error("⚠️ Environment tests failed. Check the logs for details.")
+                return
+            print("✓ Environment tests passed")
+        
+        # Initialize generator
+        try:
+            print("\n=== Initializing Generator ===")
+            generator = LyricGenerator()
+            st.session_state.generator = generator
+            st.session_state.chat_history = []
+            st.session_state.current_lyrics = None
+            st.session_state.initialized = True
+        except Exception as e:
+            st.error(f"Error initializing generator: {str(e)}")
+            print(f"Error: {str(e)}")
             return
-        print("✓ Environment tests passed")
-    
-    # Initialize generator
-    try:
-        print("\n=== Initializing Generator ===")
-        generator = LyricGenerator()
-    except Exception as e:
-        st.error(f"Error initializing generator: {str(e)}")
-        print(f"Error: {str(e)}")
-        return
     
     # Check OpenAI API key at startup
     if Settings.is_huggingface():
@@ -112,15 +119,6 @@ def main():
     
     print(f"Current deployment mode: {Settings.DEPLOYMENT_MODE}")
     
-    # Initialize generator on first run
-    if 'generator' not in st.session_state:
-        generator = initialize_generator()
-        if generator is None:
-            st.stop()
-        st.session_state.generator = generator
-        st.session_state.chat_history = []
-        st.session_state.current_lyrics = None
-    
     # Display chat history
     for message in st.session_state.chat_history:
         user_msg, assistant_msg = message

scripts/test_environment.py

@@ -1,5 +1,6 @@
 import sys
 from pathlib import Path
+import os
 
 # Add project root to path
 project_root = Path(__file__).parent.parent
@@ -19,26 +20,49 @@ def test_openai_connection():
     if not Settings.OPENAI_API_KEY:
         print("❌ OpenAI API key not found!")
         return False
-    print("✓ OpenAI API key found")
+    print(f"✓ OpenAI API key found (length: {len(Settings.OPENAI_API_KEY)})")
     
-    # Test API reachability
+    # Test API reachability with more detail
     try:
-        response = requests.get("https://api.openai.com")
+        print("Testing connection to api.openai.com...")
+        response = requests.get(
+            "https://api.openai.com",
+            timeout=10,
+            verify=True  # Force SSL verification
+        )
         print(f"✓ OpenAI API reachable (Status: {response.status_code})")
+        print(f"Response headers: {dict(response.headers)}")
+    except requests.exceptions.SSLError as e:
+        print(f"❌ SSL Error connecting to OpenAI: {e}")
+        return False
+    except requests.exceptions.ConnectionError as e:
+        print(f"❌ Connection Error: {e}")
+        print("Checking if proxy is needed...")
+        # Try to get environment proxy settings
+        print(f"HTTP_PROXY: {os.environ.get('HTTP_PROXY')}")
+        print(f"HTTPS_PROXY: {os.environ.get('HTTPS_PROXY')}")
+        return False
     except Exception as e:
-        print(f"❌ Cannot reach OpenAI API: {e}")
+        print(f"❌ Cannot reach OpenAI API: {type(e).__name__}: {e}")
         return False
     
-    # Test embeddings
+    # Test embeddings with more detail
     try:
+        print("\nTesting embeddings creation...")
         embeddings = OpenAIEmbeddings(
-            openai_api_key=Settings.OPENAI_API_KEY
+            openai_api_key=Settings.OPENAI_API_KEY,
+            timeout=30,
+            max_retries=2
         )
+        print("Embeddings object created, attempting query...")
         result = embeddings.embed_query("test")
         print(f"✓ Embeddings working (vector size: {len(result)})")
         return True
     except Exception as e:
-        print(f"❌ Embeddings error: {e}")
+        print(f"❌ Embeddings error: {type(e).__name__}: {e}")
+        print("Stack trace:")
+        import traceback
+        traceback.print_exc()
         return False
 
 

src/generator/generator.py

@@ -70,17 +70,23 @@ class LyricGenerator:
             data_dir = Path("/data")
             data_dir.mkdir(exist_ok=True)
             
-            # Create chroma directory
-            chroma_dir = data_dir / "chroma"
-            chroma_dir.mkdir(exist_ok=True)
-            
-            # Download the entire chroma directory
-            print("Downloading Chroma files...")
+            # Find the most recent snapshot
             api = HfApi(token=Settings.HF_TOKEN)
+            files = api.list_repo_files(Settings.HF_DATASET, repo_type="dataset")
+            
+            # Get all snapshot directories
+            snapshots = [f for f in files if 'snapshots' in f and 'chroma' in f]
+            if not snapshots:
+                raise RuntimeError("No snapshots found in dataset")
             
-            # Download all files from the chroma directory
-            for file in api.list_repo_files(Settings.HF_DATASET, repo_type="dataset"):
-                if file.startswith('chroma/'):
+            # Use the most recent snapshot
+            latest_snapshot = sorted(snapshots)[-1]
+            snapshot_dir = data_dir / latest_snapshot
+            print(f"Using snapshot directory: {snapshot_dir}")
+            
+            # Download all files from this snapshot
+            for file in files:
+                if latest_snapshot in file:
                     local_path = hf_hub_download(
                         repo_id=Settings.HF_DATASET,
                         filename=file,
@@ -90,10 +96,9 @@ class LyricGenerator:
                     )
                     print(f"Downloaded: {file}")
             
-            print("All files downloaded successfully")
-            
-            # Set the chroma directory
-            self.chroma_dir = chroma_dir
+            # Set the chroma directory to the snapshot directory
+            self.chroma_dir = snapshot_dir
+            print(f"Chroma directory set to: {self.chroma_dir}")
             
         except Exception as e:
             print(f"\n=== Error in _setup_embeddings_from_hf ===")