fix: resolve template crashes and add health endpoint

app/ai.py

@@ -135,15 +135,30 @@ def set_model(self, model: str) -> None:
 
     def setup_vectorstore(self, file_paths: List[str]) -> Optional[FAISS]:
         """Load documents and create a vector store for retrieval"""
+        if not file_paths:
+            logger.warning("DOC_PATHS is empty. RAG retriever will remain unconfigured.")
+            self.retriever = None
+            return None
+
         all_documents = []
         for file_path in file_paths:
             if os.path.exists(file_path):
-                if file_path.endswith(".pdf"):
-                    loader = PyMuPDFLoader(file_path)
-                else:
-                    loader = TextLoader(file_path)
-                documents = loader.load()
-                all_documents.extend(documents)
+                try:
+                    if file_path.endswith(".pdf"):
+                        loader = PyMuPDFLoader(file_path)
+                    else:
+                        loader = TextLoader(file_path)
+                    documents = loader.load()
+                    all_documents.extend(documents)
+                except Exception as e:
+                    logger.warning(f"Failed to load document '{file_path}': {e}")
+            else:
+                logger.warning(f"Document path does not exist: {file_path}")
+
+        if not all_documents:
+            logger.warning("No documents were loaded from DOC_PATHS. RAG retriever will remain unconfigured.")
+            self.retriever = None
+            return None
 
         embeddings = HuggingFaceEmbeddings(
             model_name="sentence-transformers/all-MiniLM-L6-v2"
@@ -155,6 +170,9 @@ def setup_vectorstore(self, file_paths: List[str]) -> Optional[FAISS]:
 
     def get_relevant_document(self, query: str, threshold: float = 0.5):
         """Get the most relevant document for a query"""
+        if not self.retriever:
+            return None, 0.0
+
         results = self.retriever.invoke(query)
         if results:
             top_result = results[0]
@@ -205,29 +223,38 @@ def debug(self, code: str, context: bool) -> str:
 
     def run(self, question: str) -> str:
         """Process a question through the RAG pipeline"""
-        # build chain components
-        chain_input = {
-            "context": self.retriever | format_docs,
-            "question": RunnablePassthrough()
-        }
-
-        # first chain: prompt -> combine messages -> model -> extract answer
-        first_chain = (
-            chain_input
-            | self.prompt
-            | combine_messages
-            | self.model
-            | extract_answer_from_output
-        )
-
-        doc_result, _ = self.get_relevant_document(question)
-        if doc_result:
-            first_response = first_chain.invoke({
-                "query": question,
-                "context": doc_result.page_content
-            })
+        if not self.retriever:
+            first_chain = (
+                self.prompt
+                | combine_messages
+                | self.model
+                | extract_answer_from_output
+            )
+            first_response = first_chain.invoke({"context": "", "question": question})
         else:
-            first_response = first_chain.invoke(question)
+            # build chain components
+            chain_input = {
+                "context": self.retriever | format_docs,
+                "question": RunnablePassthrough()
+            }
+
+            # first chain: prompt -> combine messages -> model -> extract answer
+            first_chain = (
+                chain_input
+                | self.prompt
+                | combine_messages
+                | self.model
+                | extract_answer_from_output
+            )
+
+            doc_result, _ = self.get_relevant_document(question)
+            if doc_result:
+                first_response = first_chain.invoke({
+                    "query": question,
+                    "context": doc_result.page_content
+                })
+            else:
+                first_response = first_chain.invoke(question)
 
         # second chain for making answer child-friendly
         second_chain = (

app/routes/admin.py

@@ -31,6 +31,7 @@ async def admin_panel(
     denied_keys = db.query(APIKey).filter(APIKey.approved == False, APIKey.is_active == True).all()
 
     return templates.TemplateResponse(
+        request,
         "admin_panel.html", 
         {
             "request": request,

app/routes/auth.py

@@ -30,7 +30,7 @@ async def logout(request: Request):
 @router.get("/oauth-login", response_class=HTMLResponse)
 async def login(request: Request):
     """Render OAuth login page"""
-    return templates.TemplateResponse("oauth_login.html", {"request": request})
+    return templates.TemplateResponse(request, "oauth_login.html", {"request": request})
 
 @router.get("/auth/github")
 async def login_github(request: Request):
@@ -137,7 +137,7 @@ async def auth_callback(provider: str, request: Request, db: Session = Depends(g
 @router.get("/admin-login", response_class=HTMLResponse)
 async def admin_login(request: Request):
     """Render admin login page"""
-    return templates.TemplateResponse("admin_login.html", {"request": request})
+    return templates.TemplateResponse(request, "admin_login.html", {"request": request})
 
 @router.post("/admin-login")
 async def admin_login_submit(
@@ -155,6 +155,7 @@ async def admin_login_submit(
 
     if not key:
         return templates.TemplateResponse(
+            request,
             "admin_login.html", 
             {"request": request, "error": "Invalid API key"}
         )
@@ -167,7 +168,7 @@ async def admin_login_submit(
 @router.get("/request-key", response_class=HTMLResponse)
 async def request_key_form(request: Request):
     """Render API key request page"""
-    return templates.TemplateResponse("request_key.html", {"request": request})
+    return templates.TemplateResponse(request, "request_key.html", {"request": request})
 
 @router.post("/request-key", response_class=HTMLResponse)
 async def submit_key_request(
@@ -192,6 +193,7 @@ async def submit_key_request(
     db.commit()
 
     return templates.TemplateResponse(
+        request,
         "request_key.html", 
         {
             "request": request, 

app/routes/main.py

@@ -13,4 +13,4 @@
 @router.get("/", response_class=HTMLResponse)
 async def root(request: Request):
     """Serve the root welcome page with links to documentation"""
-    return templates.TemplateResponse("welcome.html", {"request": request})
+    return templates.TemplateResponse(request, "welcome.html", {"request": request})

app/routes/web.py

@@ -1,7 +1,9 @@
 """
 Web routes handling HTML responses for Sugar-AI.
 """
-from fastapi import APIRouter, Depends, Request
+import os
+
+from fastapi import APIRouter, Depends, Request, status
 from fastapi.responses import HTMLResponse, RedirectResponse
 from fastapi.templating import Jinja2Templates
 from sqlalchemy.orm import Session
@@ -15,10 +17,26 @@
 # set up templates
 templates = Jinja2Templates(directory=settings.TEMPLATES_DIR)
 
+
+@router.get("/health", status_code=status.HTTP_200_OK)
+async def health_check():
+    """Return service health and essential LLM configuration status."""
+    openai_api_key = os.getenv("OPENAI_API_KEY")
+    local_llm_url = os.getenv("LOCAL_LLM_URL")
+
+    return {
+        "service": "online",
+        "components": {
+            "llm": "configured" if openai_api_key or local_llm_url else "unconfigured",
+            "openai_api_key": "configured" if openai_api_key else "unconfigured",
+            "local_llm_url": "configured" if local_llm_url else "unconfigured",
+        },
+    }
+
 @router.get("/", response_class=HTMLResponse)
 async def root(request: Request):
     """Render welcome page"""
-    return templates.TemplateResponse("welcome.html", {"request": request})
+    return templates.TemplateResponse(request, "welcome.html", {"request": request})
 
 @router.get("/dashboard", response_class=HTMLResponse)
 async def dashboard(request: Request, db: Session = Depends(get_db)):
@@ -64,7 +82,7 @@ async def dashboard(request: Request, db: Session = Depends(get_db)):
     if user and user.get("can_change_model", False):
         admin_url = "/admin"
 
-    return templates.TemplateResponse("dashboard.html", {
+    return templates.TemplateResponse(request, "dashboard.html", {
         "request": request,
         "user": user,
         "api_key": api_key,

main.py

@@ -49,6 +49,8 @@ async def startup_event():
 
     initialized_agent = RAGAgent(model=active_model)
     initialized_agent.retriever = initialized_agent.setup_vectorstore(settings.DOC_PATHS)
+    if not initialized_agent.retriever:
+        logger.warning("RAG retriever is unconfigured. /ask will run without document retrieval.")
 
     # Inject this instance into the API module
     # This updates the 'agent = None' in api.py to be the real loaded model