Read from .env

app/main.py

@@ -9,7 +9,62 @@ from typing import Any, Dict
 from fastapi import FastAPI
 from fastapi.responses import RedirectResponse
 
-# --- Middlewares ---
+# -----------------------------------------------------------------------------
+# Early: load .env (so HF_TOKEN, ADMIN_TOKEN, etc. are available locally)
+# -----------------------------------------------------------------------------
+def _load_env_file(paths: list[str]) -> None:
+    """Load environment variables from the first existing path in `paths`.
+    Prefer python-dotenv if present; otherwise use a tiny fallback parser."""
+    logger = logging.getLogger("uvicorn.error")
+
+    # 1) Try python-dotenv (best)
+    try:
+        from dotenv import load_dotenv  # type: ignore
+        for p in paths:
+            if os.path.exists(p):
+                load_dotenv(dotenv_path=p, override=False)
+                logger.info("Loaded environment from %s", p)
+                return
+        logger.info("No .env file found in %s (skipping)", paths)
+        return
+    except Exception:
+        # 2) Fallback: simple parser
+        for p in paths:
+            if not os.path.exists(p):
+                continue
+            try:
+                with open(p, "r", encoding="utf-8") as f:
+                    for raw in f:
+                        line = raw.strip()
+                        if not line or line.startswith("#"):
+                            continue
+                        if line.startswith("export "):
+                            line = line[len("export ") :].strip()
+                        if "=" not in line:
+                            continue
+                        key, val = line.split("=", 1)
+                        key, val = key.strip(), val.strip()
+                        # strip optional quotes
+                        if (val.startswith('"') and val.endswith('"')) or (
+                            val.startswith("'") and val.endswith("'")
+                        ):
+                            val = val[1:-1]
+                        # do not clobber existing env (Space Secrets)
+                        os.environ.setdefault(key, val)
+                logger.info("Loaded environment from %s (fallback parser)", p)
+                return
+            except Exception as e:
+                logger.warning("Failed loading env from %s: %s", p, e)
+
+    logger.info("No .env loaded (none found / parsers failed)")
+
+# Try typical locations for local dev. HF Spaces will ignore this and use Secrets.
+_load_env_file([".env", "configs/.env", ".env.local", "configs/.env.local"])
+
+
+# -----------------------------------------------------------------------------
+# Middlewares
+# -----------------------------------------------------------------------------
 # Prefer the canonical package name; if your repo uses "middlewares/", this tries both.
 try:
     from .middleware import attach_middlewares  # singular
@@ -22,7 +77,9 @@ except Exception:
                 "attach_middlewares not found; continuing without custom middlewares."
             )
 
-# --- Routers ---
+# -----------------------------------------------------------------------------
+# Routers
+# -----------------------------------------------------------------------------
 from .routers import health, plan, chat
 
 # Optional UI (Home/Chat/Dev). If missing, we gracefully fall back to a JSON root.
@@ -32,6 +89,7 @@ try:
 except Exception:  # pragma: no cover
     HAS_UI = False
 
+
 TAGS_METADATA = [
     {"name": "Health", "description": "Liveness / readiness probes and basic service metadata."},
     {"name": "Planning", "description": "AI plan generation for Matrix Guardian (/v1/plan)."},
@@ -39,12 +97,19 @@ TAGS_METADATA = [
     {"name": "UI", "description": "Minimal web UI (Home, Chat, Dev) if enabled."},
 ]
 
+
 @asynccontextmanager
 async def lifespan(app: FastAPI):
     app.state.started_at = time.time()
     app.state.version = os.getenv("APP_VERSION", "1.0.0")
+
+    # Minimal diagnostics; HF_TOKEN presence matters for inference
+    hf_token_present = bool(os.getenv("HF_TOKEN"))
     logging.getLogger("uvicorn.error").info(
-        "matrix-ai starting (version=%s, port=%s)", app.state.version, os.getenv("PORT", "7860")
+        "matrix-ai starting (version=%s, port=%s, hf_token_present=%s)",
+        app.state.version,
+        os.getenv("PORT", "7860"),
+        "yes" if hf_token_present else "no",
     )
     try:
         yield
@@ -54,6 +119,7 @@ async def lifespan(app: FastAPI):
             "matrix-ai shutting down (uptime=%.2fs)", uptime
         )
 
+
 def create_app() -> FastAPI:
     app = FastAPI(
         title="matrix-ai",
@@ -94,4 +160,5 @@ def create_app() -> FastAPI:
 
     return app
 
+
 app = create_app()

app/routers/chat.py

@@ -1,18 +1,58 @@
-from fastapi import APIRouter, Depends, HTTPException
+# app/routers/chat.py
+from fastapi import APIRouter, Depends, HTTPException, Query
+from pydantic import BaseModel
+from typing import List, Optional
 from ..deps import get_settings
 from ..core.config import Settings
-from ..core.schema import ChatRequest, ChatResponse
-from ..services.chat_service import chat_answer
+from ..services.chat_service import ChatService
 
 router = APIRouter()
 
+class ChatMessage(BaseModel):
+    role: str
+    content: str
+
+class ChatRequest(BaseModel):
+    # accept several shapes so UI/clients don't 422:
+    query: Optional[str] = None
+    question: Optional[str] = None
+    prompt: Optional[str] = None
+    messages: Optional[List[ChatMessage]] = None
+
+    def as_text(self) -> str:
+        if self.query:
+            return self.query
+        if self.question:
+            return self.question
+        if self.prompt:
+            return self.prompt
+        if self.messages:
+            # prefer last user message
+            for m in reversed(self.messages):
+                if m.role.lower() == "user":
+                    return m.content
+            # fallback to last message
+            if self.messages:
+                return self.messages[-1].content
+        raise ValueError("Body must include 'query'/'question'/'prompt' or 'messages'")
+
+class ChatResponse(BaseModel):
+    answer: str
+
 @router.post("/chat", response_model=ChatResponse)
-async def v1_chat(
-    req: ChatRequest,
-    settings: Settings = Depends(get_settings)
-):
-    """Answers questions about the MatrixHub ecosystem using RAG."""
+async def chat(req: ChatRequest, settings: Settings = Depends(get_settings)):
     try:
-        return await chat_answer(req, settings=settings)
-    except Exception as e:
-        raise HTTPException(status_code=500, detail=f"Failed to process chat request: {e}")
+        text = req.as_text()
+    except ValueError as e:
+        raise HTTPException(status_code=422, detail=str(e))
+    svc = ChatService(settings)
+    answer = await svc.answer(text)
+    return ChatResponse(answer=answer)
+
+# Handy for curl/browser tests:
+#   GET /v1/chat?query=hello
+@router.get("/chat", response_model=ChatResponse)
+async def chat_get(query: str = Query(...), settings: Settings = Depends(get_settings)):
+    svc = ChatService(settings)
+    answer = await svc.answer(query)
+    return ChatResponse(answer=answer)

app/services/chat_service.py

@@ -1,10 +1,23 @@
-# Placeholder for Stage-2 RAG chat service
-from ..core.schema import ChatRequest, ChatResponse
+# app/services/chat_service.py
+from __future__ import annotations
 from ..core.config import Settings
+from ..core.inference.client import HFClient
 
-async def chat_answer(req: ChatRequest, settings: Settings) -> ChatResponse:
-    """Placeholder chat function."""
-    return ChatResponse(
-        answer="The RAG chat service is not yet enabled in Stage-1.",
-        sources=[]
-    )
+SYSTEM_PROMPT = (
+    "You are MATRIX-AI, a concise, helpful assistant for the Matrix EcoSystem. "
+    "Answer clearly and briefly. If unsure, say so."
+)
+
+class ChatService:
+    def __init__(self, settings: Settings):
+        self.settings = settings
+        self.client = HFClient(model=settings.model.name)
+
+    async def answer(self, query: str) -> str:
+        prompt = f"{SYSTEM_PROMPT}\n\nUser: {query}\nAssistant:"
+        text = await self.client.generate(
+            prompt=prompt,
+            max_new_tokens=self.settings.model.max_new_tokens,
+            temperature=self.settings.model.temperature,
+        )
+        return (text or "").strip()

app/templates/base.html

@@ -156,6 +156,8 @@
       <a href="/chat">Chat</a>
       <a href="/dev">Dev</a>
       <a href="/docs" target="_blank" rel="noreferrer">API Docs</a>
+      <a href="https://github.com/agent-matrix/matrix-ai" target="_blank" rel="noreferrer" title="Give me a star on GitHub!">GitHub</a>
+
     </nav>
   </header>