Upload 3 files

app.py

@@ -0,0 +1,143 @@
+import os, sys, time, asyncio
+from typing import List, Dict
+import gradio as gr
+from dotenv import load_dotenv
+
+# Windows event loop (prevents asyncio warnings on Win)
+if sys.platform.startswith("win"):
+    try:
+        asyncio.set_event_loop_policy(asyncio.WindowsProactorEventLoopPolicy())
+    except Exception:
+        pass
+
+# config from .env or defaults
+load_dotenv()
+APP_Name    = os.getenv("APP_Name", "Ghaymah GenAI chatbots")
+APP_Version = os.getenv("APP_Version", "0.1.0")
+API_KEY     = os.getenv("API_KEY", "")
+
+# Models from .env or fallback to your set
+MODELS = [m.strip() for m in os.getenv("Models", "").split(",") if m.strip()] or [
+    "gemma-3-4b-it",
+    "QwQ-32B",
+    "DeepSeek-V3-0324",
+    "Qwen/Qwen3-32B",
+    "zai-org/GLM-4.5-Air",
+    "moonshotai/Kimi-K2-Instruct",
+]
+
+# Friendly descriptions & logo
+MODEL_INFO = {
+    "gemma-3-4b-it": "Google Gemma-3 4B Instruct — light, fast, solid reasoning.",
+    "QwQ-32B": "QwQ-32B — reasoning-focused; strong long-form answers.",
+    "DeepSeek-V3-0324": "DeepSeek V3 (0324) — versatile, great multi-step reasoning.",
+    "Qwen/Qwen3-32B": "Qwen3-32B — multilingual, good code & math.",
+    "zai-org/GLM-4.5-Air": "GLM-4.5-Air — efficient generalist, good latency.",
+    "moonshotai/Kimi-K2-Instruct": "Kimi K2 Instruct — long-context, helpful writing.",
+}
+LOGO_PATH = "download.jpeg"  # change to your image if different
+
+# ── OpenAI-compatible client ──────────────────────────────────────────────────
+from openai import OpenAI
+BASE_URL = "https://genai.ghaymah.systems"
+client = OpenAI(api_key=API_KEY, base_url=BASE_URL) if API_KEY else None
+
+SYSTEM_SEED = "You are Ghaymah Assistant. Be concise and helpful."
+
+# Helpers 
+BACKOFF = [5, 10, 20]  # basic retry for 429s
+
+def safe_chat_complete(model: str, messages: List[Dict], max_tokens: int = 800) -> str:
+    if not client:
+        return "⚠️ Missing API_KEY in .env"
+    attempt = 0
+    while True:
+        try:
+            resp = client.chat.completions.create(
+                model=model,
+                messages=messages,
+                max_tokens=max_tokens,
+                temperature=0.3,
+                timeout=90,
+            )
+            return resp.choices[0].message.content or ""
+        except Exception as e:
+            msg = str(e)
+            if ("429" in msg or "Rate" in msg) and attempt < len(BACKOFF):
+                time.sleep(BACKOFF[attempt]); attempt += 1
+                continue
+            return f"Request failed for `{model}`: {e}"
+
+def init_state():
+    return {"messages": [{"role": "system", "content": SYSTEM_SEED}]}
+
+# Gradio app
+with gr.Blocks(title=APP_Name) as demo:
+    state = gr.State(init_state())
+
+    gr.Markdown(f"# {APP_Name}  \n<span style='opacity:.7'>v{APP_Version}</span>")
+
+    with gr.Row():
+        # Left: Chat
+        with gr.Column(scale=3):
+            chat = gr.Chatbot(label="Chat", height=520, type="messages", value=[])
+            user_in = gr.Textbox(label="Your message", placeholder="Type here…", lines=2)
+            with gr.Row():
+                send_btn = gr.Button("Send", variant="primary")
+                clear_btn = gr.Button("Clear")
+
+        # Right: Model selector + logo + info
+        with gr.Column(scale=1, min_width=320):
+            model_choice = gr.Radio(
+                choices=MODELS,
+                value=MODELS[0],
+                label="Models",
+                info="Select Your Model Here",
+            )
+            gr.Image(LOGO_PATH, show_label=False, container=False)
+            info_md = gr.Markdown(MODEL_INFO.get(MODELS[0], ""))
+
+            def _update_info(m: str) -> str:
+                title = f"**{m}**"
+                desc = MODEL_INFO.get(m, "")
+                return f"{title}\n\n{desc}"
+            model_choice.change(_update_info, model_choice, info_md)
+
+    # Step 1: push the user message into the chat stream
+    def on_submit(msg, chat_messages):
+        if not msg:
+            return "", (chat_messages or [])
+        updated = (chat_messages or []) + [{"role": "user", "content": msg}]
+        return "", updated
+
+    def bot_step(chat_messages, chosen_model, st):
+        msgs = [{"role": "system", "content": SYSTEM_SEED}]
+        # only include last 2 visible messages
+        for m in (chat_messages or [])[-2:]:
+            role, content = m.get("role"), m.get("content")
+            if role in ("user", "assistant") and isinstance(content, str):
+                msgs.append({"role": role, "content": content})
+
+        reply = safe_chat_complete(chosen_model, msgs, max_tokens=800)
+        updated = (chat_messages or []) + [{"role": "assistant", "content": reply}]
+        st = st or init_state()
+        st["messages"] = msgs + [{"role": "assistant", "content": reply}]
+        return updated, st
+
+
+    # Clear
+    def on_clear():
+        return [], init_state()
+
+    # Wire events
+    user_in.submit(on_submit, [user_in, chat], [user_in, chat]) \
+           .then(bot_step, [chat, model_choice, state], [chat, state])
+
+    send_btn.click(on_submit, [user_in, chat], [user_in, chat]) \
+            .then(bot_step, [chat, model_choice, state], [chat, state])
+
+    clear_btn.click(on_clear, outputs=[chat, state])
+
+if __name__ == "__main__":
+    demo.queue()
+    demo.launch(debug=True)

requirements.txt

@@ -0,0 +1,3 @@
+openai
+dotenv
+gradio