update full files

Dockerfile

@@ -10,9 +10,16 @@ RUN apt-get update && apt-get install -y \
     make \
     && rm -rf /var/lib/apt/lists/*
 
+# إنشاء مستخدم غير جذر لتشغيل التطبيق
+RUN useradd -m -u 1000 appuser
+
 # إعداد مجلد العمل
 WORKDIR /app
 
+# إنشاء مجلد التخزين المؤقت وتغيير الأذونات
+RUN mkdir -p /app/.cache/huggingface && \
+    chown -R appuser:appuser /app
+
 # نسخ ملفات التطبيق
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
@@ -25,9 +32,12 @@ RUN chmod +x setup.sh
 # تحميل النموذج
 RUN ./setup.sh
 
-# تعيين المتغيّر لتجنب التحذير
+# تعيين المتغيرات البيئية
 ENV HF_HOME=/app/.cache/huggingface
-ENV TRANSFORMERS_CACHE=/app/.cache/huggingface
+ENV PYTHONUNBUFFERED=1
+
+# تغيير المستخدم إلى appuser
+USER appuser
 
 # تشغيل التطبيق
 CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "8080"]

README.md

@@ -1,5 +1,5 @@
 ---
-title: Api Mg
+title: MGZON Smart Assistant
 emoji: 🏃
 colorFrom: purple
 colorTo: gray
@@ -8,5 +8,38 @@ sdk_version: 0.115.0
 app_file: app.py
 pinned: false
 license: apache-2.0
-short_description: MGZON/api-mg
+short_description: API for integrating T5 and Mistral-7B (GGUF) models in Hugging Face Spaces
 ---
+
+# MGZON Smart Assistant
+
+This project provides a FastAPI-based API for integrating two language models:
+- **MGZON-FLAN-T5**: A pre-trained T5 model fine-tuned to respond to questions containing keywords like "mgzon", "flan", or "t5".
+- **Mistral-7B-GGUF**: A Mistral-7B model in GGUF format for answering general questions.
+
+![Demo](https://huggingface.co/spaces/MGZON/api-mg/resolve/main/demo.png)
+
+## Setup
+- **Docker**: The image is built using `python:3.10-slim` with development tools (`gcc`, `g++`, `cmake`, `make`) installed to support building `llama-cpp-python`.
+- **System Requirements**: Dependencies are installed from `requirements.txt`, including `transformers`, `torch`, `fastapi`, and `llama-cpp-python`.
+- **Model Download**: The Mistral-7B GGUF model is downloaded via `setup.sh` using `huggingface_hub`.
+- **Environment Variables**:
+  - `HF_HOME` and `TRANSFORMERS_CACHE` are set to `/app/.cache/huggingface`.
+  - `HF_TOKEN` (secret) is required to access models from the Hugging Face Hub.
+
+## How to Run
+1. Build the Docker image using the provided `Dockerfile`.
+2. Ensure the `HF_TOKEN` is set in the Hugging Face Spaces settings.
+3. Run the application using `uvicorn` on port 8080.
+
+## Endpoint
+- **POST /ask**:
+  - **Input**: JSON containing `question` (the query) and `max_new_tokens` (optional, default=150).
+  - **Output**: JSON containing `model` (name of the model used) and `response` (the answer).
+
+## Example Usage
+```bash
+curl -X POST "https://mgzon-api-mg.hf.space/ask" \
+-H "Content-Type: application/json" \
+-d '{"question": "What is MGZON?", "max_new_tokens": 100}'
+```

app.py

@@ -4,27 +4,19 @@ from pydantic import BaseModel
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 from llama_cpp import Llama
 
-# -------------------------------------------------
 # إعداد مسار الـ cache
-# -------------------------------------------------
-CACHE_DIR = "/app/.cache/huggingface"  # مسار موحد لـ Hugging Face Spaces
-os.makedirs(CACHE_DIR, exist_ok=True)
+CACHE_DIR = os.environ.get("HF_HOME", "/app/.cache/huggingface")
 
 # تأكد من أن المكتبتين تقرأ المتغيّرات البيئية
-os.environ["TRANSFORMERS_CACHE"] = CACHE_DIR
 os.environ["HF_HOME"] = CACHE_DIR
 
-# -------------------------------------------------
 # إنشاء التطبيق
-# -------------------------------------------------
 app = FastAPI(
     title="MGZON Smart Assistant",
     description="دمج نموذج T5 المدرب مع Mistral‑7B (GGUF) داخل Space"
 )
 
-# -------------------------------------------------
-# 1️⃣ تحميل نموذج T5 المدرب من Hub
-# -------------------------------------------------
+# تحميل نموذج T5 المدرب من Hub
 T5_REPO = "MGZON/mgzon-flan-t5-base"
 try:
     t5_tokenizer = AutoTokenizer.from_pretrained(T5_REPO, cache_dir=CACHE_DIR)
@@ -32,9 +24,7 @@ try:
 except Exception as e:
     raise RuntimeError(f"فشل تحميل نموذج T5 من {T5_REPO}: {str(e)}")
 
-# -------------------------------------------------
-# 2️⃣ تحميل ملف Mistral .gguf
-# -------------------------------------------------
+# تحميل ملف Mistral .gguf
 gguf_path = os.path.abspath("models/mistral-7b-instruct-v0.1.Q4_K_M.gguf")
 if not os.path.exists(gguf_path):
     raise RuntimeError(
@@ -52,23 +42,18 @@ try:
 except Exception as e:
     raise RuntimeError(f"فشل تحميل نموذج Mistral من {gguf_path}: {str(e)}")
 
-# -------------------------------------------------
 # تعريف شكل الطلب (JSON)
-# -------------------------------------------------
 class AskRequest(BaseModel):
     question: str
     max_new_tokens: int = 150
 
-# -------------------------------------------------
 # نقطة النهاية /ask
-# -------------------------------------------------
 @app.post("/ask")
 def ask(req: AskRequest):
     q = req.question.strip()
     if not q:
         raise HTTPException(status_code=400, detail="Empty question")
 
-    # منطق اختيار النموذج
     try:
         if any(tok in q.lower() for tok in ["mgzon", "flan", "t5"]):
             # نموذج T5

setup.sh

@@ -8,6 +8,7 @@ fi
 
 # إنشاء مجلد لتخزين النموذج
 mkdir -p models
+chown -R appuser:appuser models
 
 # تحميل ملف .gguf إذا لم يكن موجودًا مسبقًا
 python - <<PY