Update app.py

app.py

@@ -1,9 +1,9 @@
+# app.py
 import os
 import secrets
 import logging
 import asyncio
 import html
-
 from dataclasses import dataclass
 from typing import Any, Optional, Tuple
 
@@ -11,6 +11,9 @@ import gradio as gr
 from transformers import pipeline
 from dotenv import load_dotenv
 from pydantic import BaseModel
+from fastapi import FastAPI, Request
+from fastapi.responses import JSONResponse
+import uvicorn
 
 # ----------------- Configuration & Models -----------------
 load_dotenv()
@@ -38,25 +41,20 @@ class APIResponse(BaseModel):
     error: Optional[str] = None
 
 
-# ----------------- Enhanced Logger -----------------
+# ----------------- Logger -----------------
 def setup_logger() -> logging.Logger:
     cfg = Config()
     log_level = getattr(logging, cfg.LOG_LEVEL.upper(), logging.INFO)
-
     logger = logging.getLogger("gemma_saas")
     if not logger.handlers:
         logger.setLevel(log_level)
         formatter = logging.Formatter("%(asctime)s - %(name)s - %(levelname)s - %(message)s")
-
-        file_handler = logging.FileHandler("gemma_saas.log")
-        file_handler.setFormatter(formatter)
-
-        stream_handler = logging.StreamHandler()
-        stream_handler.setFormatter(formatter)
-
-        logger.addHandler(file_handler)
-        logger.addHandler(stream_handler)
-
+        fh = logging.FileHandler("gemma_saas.log")
+        fh.setFormatter(formatter)
+        sh = logging.StreamHandler()
+        sh.setFormatter(formatter)
+        logger.addHandler(fh)
+        logger.addHandler(sh)
     return logger
 
 
@@ -71,17 +69,23 @@ class ModelManager:
         self.model_loaded = False
 
     async def initialize(self) -> None:
+        """
+        Inicializa o pipeline. Usa HF_TOKEN (variável de ambiente ou Config).
+        Evita passar `use_auth_token` em model_kwargs.
+        """
         if not self.config.HF_TOKEN:
             logger.error("Token do Hugging Face não encontrado. O carregamento do modelo irá falhar.")
             return
 
         try:
             logger.info(f"A carregar o modelo: {self.config.MODEL_NAME}...")
+            # garante env var como fallback
             os.environ.setdefault("HF_TOKEN", self.config.HF_TOKEN)
 
             loop = asyncio.get_event_loop()
 
             def load_pipeline():
+                # Passa token diretamente (substitui use_auth_token)
                 return pipeline(
                     "text-generation",
                     model=self.config.MODEL_NAME,
@@ -98,11 +102,7 @@ class ModelManager:
 
     async def generate(self, request: GenerationRequest) -> Tuple[bool, str, int]:
         if not self.model_loaded or self.pipeline is None:
-            return (
-                False,
-                "❌ O modelo não está disponível. Por favor, verifique os logs do servidor.",
-                0,
-            )
+            return False, "❌ O modelo não está disponível. Por favor, verifique os logs do servidor.", 0
 
         try:
             if not request.prompt.strip():
@@ -132,7 +132,7 @@ class ModelManager:
 
                 generated_text = outputs[0].get("generated_text", "")
                 if generated_text.startswith(prompt_text):
-                    generated_text = generated_text[len(prompt_text) :]
+                    generated_text = generated_text[len(prompt_text):]
 
                 tokens_used = 0
                 if tokenizer and hasattr(tokenizer, "encode"):
@@ -161,11 +161,12 @@ class GemmaService:
         await self.model_manager.initialize()
 
     async def generate_text(self, api_key: str, prompt: str, **kwargs) -> APIResponse:
-        if not api_key or not api_key.startswith("gsk-"):
+        # Validação simples da gsk-... chave da app
+        if not api_key or not isinstance(api_key, str) or not api_key.startswith("gsk-"):
             return APIResponse(success=False, error="Chave de API inválida ou ausente.")
         try:
-            request = GenerationRequest(prompt=prompt, **kwargs)
-            success, text, tokens_used = await self.model_manager.generate(request)
+            req = GenerationRequest(prompt=prompt, **kwargs)
+            success, text, tokens_used = await self.model_manager.generate(req)
             if success:
                 return APIResponse(success=True, data={"generated_text": text, "tokens_used": tokens_used})
             else:
@@ -175,22 +176,16 @@ class GemmaService:
             return APIResponse(success=False, error="Ocorreu um erro interno no serviço.")
 
 
-# ----------------- Enhanced UI -----------------
+# ----------------- Gradio UI -----------------
 class GradioInterface:
     def __init__(self, service: GemmaService):
         self.service = service
 
     def create_custom_css(self) -> str:
-        # Importa Material Icons e adiciona ícones via pseudo-elementos nos botões
         return """
-        /* importar Material Icons */
         @import url('https://fonts.googleapis.com/css2?family=Material+Icons&display=swap');
 
-        :root {
-            --dark-bg: #0a0a0a; --panel-bg: #1a1a1a; --border-color: #333;
-            --text-color: #f0f0f0; --text-light: #a0a0a0; --accent-orange: #FF4500;
-            --accent-orange-hover: #FF6347; --code-bg: #282c34;
-        }
+        :root { --dark-bg:#0a0a0a; --panel-bg:#1a1a1a; --border-color:#333; --text-color:#f0f0f0; --text-light:#a0a0a0; --accent-orange:#FF4500; --accent-orange-hover:#FF6347; --code-bg:#282c34; }
         .gradio-container { background: var(--dark-bg) !important; color: var(--text-color); }
         #main_layout { background: transparent; border: none !important; box-shadow: none !important; gap: 2rem; }
         #right_panel, #left_panel { background: var(--panel-bg); border: 1px solid var(--border-color); border-radius: 16px; padding: 2rem !important; }
@@ -201,99 +196,25 @@ class GradioInterface:
         #api_key_input textarea, #prompt_input textarea { background-color: #2C2C2C !important; border-color: var(--border-color) !important; color: var(--text-color) !important; border-radius: 12px !important; }
         #send_button { background: var(--accent-orange); color: white; border: none; border-radius: 12px !important; transition: background-color 0.3s ease; position: relative; padding-left: 3rem; }
         #send_button:hover { background-color: var(--accent-orange-hover); }
-        #generate_button {
-            background: linear-gradient(135deg, var(--accent-orange), var(--accent-orange-hover)); color: white !important;
-            font-size: 1.1rem !important; font-weight: bold !important; border: none; border-radius: 12px !important;
-            padding: 1rem 1.25rem !important; box-shadow: 0 4px 15px rgba(255, 69, 0, 0.4); transition: all 0.3s ease; position: relative; padding-left: 3rem;
-        }
-        #generate_button:hover { transform: translateY(-2px); box-shadow: 0 6px 20px rgba(255, 69, 0, 0.6); }
-        h2, h3 { color: white; border-bottom: 1px solid var(--border-color); padding-bottom: 0.75rem; margin-bottom: 1.5rem; font-weight: 600; }
+        #generate_button { background: linear-gradient(135deg, var(--accent-orange), var(--accent-orange-hover)); color: white !important; font-size: 1.1rem !important; font-weight: bold !important; border: none; border-radius: 12px !important; padding: 1rem 1.25rem !important; box-shadow: 0 4px 15px rgba(255,69,0,0.4); transition: all 0.3s ease; position: relative; padding-left: 3rem; }
+        #generate_button:hover { transform: translateY(-2px); box-shadow: 0 6px 20px rgba(255,69,0,0.6); }
         .code-snippet { background-color: var(--code-bg); color: #abb2bf; padding: 1.5rem; border-radius: 12px; font-family: 'Courier New', monospace; white-space: pre-wrap; word-wrap: break-word; border: 1px solid var(--border-color); }
-        .code-snippet .keyword { color: #c678dd; } .code-snippet .string { color: #98c379; } .code-snippet .number { color: #d19a66; }
         .gr-slider { color: var(--text-light); }
 
-        /* estilo para usar as Material Icons como ligatures */
-        .material-icon {
-            font-family: 'Material Icons', sans-serif;
-            font-weight: normal;
-            font-style: normal;
-            font-size: 20px;
-            line-height: 1;
-            letter-spacing: normal;
-            text-transform: none;
-            display: inline-block;
-            white-space: nowrap;
-            word-wrap: normal;
-            direction: ltr;
-            -webkit-font-feature-settings: 'liga';
-            -webkit-font-smoothing: antialiased;
-        }
-
-        /* adicionar ícones antes dos botões (usando ligatures) */
-        #send_button::before {
-            content: "send"; /* ligature do ícone */
-            font-family: 'Material Icons', sans-serif;
-            position: absolute;
-            left: 12px;
-            top: 50%;
-            transform: translateY(-50%);
-            font-size: 18px;
-            line-height: 1;
-            opacity: 0.95;
-        }
-
-        #generate_button::before {
-            content: "auto_awesome";
-            font-family: 'Material Icons', sans-serif;
-            position: absolute;
-            left: 12px;
-            top: 50%;
-            transform: translateY(-50%);
-            font-size: 18px;
-            line-height: 1;
-            opacity: 0.95;
-        }
-
-        /* ícone para o botão de gerar chave (se usar outro botão, adapte o id) */
-        #generate_button[aria-label], #generate_button[title] { /* fallback */
-            padding-left: 3rem;
-        }
-
-        /* ícone ao lado do exemplo de código (vpn_key) */
-        #right_panel .code-snippet::before {
-            content: "vpn_key";
-            font-family: 'Material Icons', sans-serif;
-            display: inline-block;
-            margin-right: 0.5rem;
-            vertical-align: middle;
-            font-size: 18px;
-            opacity: 0.9;
-        }
+        #send_button::before { content: "send"; font-family: 'Material Icons', sans-serif; position:absolute; left:12px; top:50%; transform:translateY(-50%); font-size:18px; opacity:0.95; }
+        #generate_button::before { content: "auto_awesome"; font-family: 'Material Icons', sans-serif; position:absolute; left:12px; top:50%; transform:translateY(-50%); font-size:18px; opacity:0.95; }
         """
 
     async def create_interface(self) -> gr.Blocks:
-        with gr.Blocks(css=self.create_custom_css(), theme=None) as app:
+        with gr.Blocks(css=self.create_custom_css(), theme=None) as demo:
             with gr.Row(elem_id="main_layout", equal_height=False):
                 with gr.Column(scale=2):
                     with gr.Column(elem_id="left_panel"):
-                        output_display = gr.Markdown(
-                            elem_id="output_display",
-                            value="<p style='color: #a0a0a0;'>A sua resposta aparecerá aqui...</p>",
-                        )
+                        output_display = gr.Markdown(elem_id="output_display", value="<p style='color: #a0a0a0;'>A sua resposta aparecerá aqui...</p>")
                         with gr.Column(elem_id="input_area"):
-                            api_key_input = gr.Textbox(
-                                label="A Sua Chave de API",
-                                placeholder="Cole a sua chave gsk-... aqui",
-                                type="password",
-                                elem_id="api_key_input",
-                            )
+                            api_key_input = gr.Textbox(label="A Sua Chave de API", placeholder="Cole a sua chave gsk-... aqui", type="password", elem_id="api_key_input")
                             with gr.Row():
-                                prompt_input = gr.Textbox(
-                                    show_label=False,
-                                    placeholder="Digite a sua mensagem...",
-                                    elem_id="prompt_input",
-                                    scale=10,
-                                )
+                                prompt_input = gr.Textbox(show_label=False, placeholder="Digite a sua mensagem...", elem_id="prompt_input", scale=10)
                                 send_button = gr.Button("➤ Enviar", elem_id="send_button", scale=2)
 
                 with gr.Column(scale=1):
@@ -312,54 +233,42 @@ class GradioInterface:
 
             def handle_key_generation():
                 key = f"gsk-{secrets.token_urlsafe(24).replace('_', '').replace('-', '')}"
-                code_html = f"""
+                code_html = f'''
                 <div class="code-snippet">
-                    <div><span class="keyword">import</span> requests</div>
-                    <div>&nbsp;</div>
-                    <div>url = <span class="string">"https://GuXSs.hf.space/run/generate"</span></div>
-                    <div>payload = {{</div>
-                    <div>&nbsp;&nbsp;&nbsp;&nbsp;<span class="string">"api_key"</span>: <span class="string">"{key}"</span>,</div>
-                    <div>&nbsp;&nbsp;&nbsp;&nbsp;<span class="string">"prompt"</span>: <span class="string">"Escreva um haikai sobre o universo"</span>,</div>
-                    <div>&nbsp;&nbsp;&nbsp;&nbsp;<span class="string">"max_tokens"</span>: <span class="number">50</span></div>
-                    <div>}}</div>
-                    <div>&nbsp;</div>
-                    <div>response = requests.post(url, json=payload)</div>
-                    <div><span class="keyword">print</span>(response.json())</div>
+                  <div><span class="keyword">import</span> requests</div>
+                  <div>&nbsp;</div>
+                  <div>url = <span class="string">"https://GuXSs.hf.space/api/generate"</span></div>
+                  <div>payload = {{</div>
+                  <div>&nbsp;&nbsp;&nbsp;&nbsp;<span class="string">"api_key"</span>: <span class="string">"{key}"</span>,</div>
+                  <div>&nbsp;&nbsp;&nbsp;&nbsp;<span class="string">"prompt"</span>: <span class="string">"Escreva um haikai sobre o universo"</span>,</div>
+                  <div>&nbsp;&nbsp;&nbsp;&nbsp;<span class="string">"max_tokens"</span>: <span class="number">50</span></div>
+                  <div>}}</div>
+                  <div>&nbsp;</div>
+                  <div>response = requests.post(url, json=payload)</div>
+                  <div><span class="keyword">print</span>(response.json())</div>
                 </div>
-                """
+                '''
                 return key, gr.update(value=code_html)
 
             async def handle_generation(api_key, prompt, temp, max_tokens, top_k, top_p, btn):
+                # função que Gradio chama quando botão é pressionado na UI
                 if not api_key:
-                    yield (
-                        "<p style='color: #FFCC00;'>Por favor, insira a sua chave de API para começar.</p>",
-                        gr.update(value="➤ Enviar", interactive=True),
-                    )
+                    yield "<p style='color: #FFCC00;'>Por favor, insira a sua chave de API para começar.</p>", gr.update(value="➤ Enviar", interactive=True)
                     return
                 if not prompt:
-                    yield (
-                        "<p style='color: #FFCC00;'>Por favor, digite um prompt.</p>",
-                        gr.update(value="➤ Enviar", interactive=True),
-                    )
+                    yield "<p style='color: #FFCC00;'>Por favor, digite um prompt.</p>", gr.update(value="➤ Enviar", interactive=True)
                     return
 
                 yield "<p style='color: #a0a0a0;'>A gerar resposta...</p>", gr.update(value="A gerar...", interactive=False)
 
-                response = await self.service.generate_text(
-                    api_key=api_key,
-                    prompt=prompt,
-                    temperature=temp,
-                    max_tokens=int(max_tokens),
-                    top_k=int(top_k),
-                    top_p=top_p,
-                )
-
+                response = await self.service.generate_text(api_key=api_key, prompt=prompt, temperature=temp, max_tokens=int(max_tokens), top_k=int(top_k), top_p=top_p)
                 if response.success:
                     formatted_text = html.escape(response.data["generated_text"]).replace("\n", "<br>")
                     yield formatted_text, gr.update(value="➤ Enviar", interactive=True)
                 else:
                     yield f"<p style='color: #FF4500;'>{response.error}</p>", gr.update(value="➤ Enviar", interactive=True)
 
+            # conectar o callback da UI — nome API usado por Gradio será "generate"
             send_button.click(
                 handle_generation,
                 inputs=[api_key_input, prompt_input, temp_slider, max_tokens_slider, top_k_slider, top_p_slider, send_button],
@@ -368,29 +277,95 @@ class GradioInterface:
             )
 
             key_button.click(handle_key_generation, outputs=[api_key_input, api_example_display])
+            demo.load(lambda: gr.update(value="<p style='color: #a0a0a0;'>Clique em 'Gerar Nova Chave' para ver um exemplo de código.</p>"), [], [api_example_display])
 
-            app.load(
-                lambda: gr.update(value="<p style='color: #a0a0a0;'>Clique em 'Gerar Nova Chave' para ver um exemplo de código.</p>"),
-                [],
-                [api_example_display],
-            )
+        return demo
 
-        return app
 
+# ----------------- FastAPI + endpoints -----------------
+def create_fastapi_app(gradio_blocks: gr.Blocks, service: GemmaService) -> FastAPI:
+    fast_app = FastAPI(title="Gemma Service (Gradio + API)")
 
-# ----------------- Main Application -----------------
-async def main():
+    # monta a UI Gradio na raiz "/" (usa mount_gradio_app)
     try:
-        service = GemmaService()
-        await service.initialize()
+        # função disponibilizada por versões recentes do gradio
+        gr.mount_gradio_app(fast_app, gradio_blocks, path="/")
+    except Exception as exc:
+        logger.warning("Não foi possível montar Gradio com mount_gradio_app: %s. A UI pode não funcionar embutida.", exc)
+
+    @fast_app.post("/api/generate")
+    async def api_generate(req: Request):
+        """
+        Endpoint REST "amigável" que aceita JSON:
+        { "api_key": "...", "prompt": "...", "max_tokens": 128, "temperature": 0.7, "top_k": 50, "top_p": 0.95 }
+        """
+        try:
+            body = await req.json()
+        except Exception:
+            return JSONResponse(status_code=400, content={"success": False, "error": "Payload inválido (JSON esperado)."})
+
+        api_key = body.get("api_key")
+        prompt = body.get("prompt", "")
+        max_tokens = int(body.get("max_tokens", 512))
+        temperature = float(body.get("temperature", 0.7))
+        top_k = int(body.get("top_k", 50))
+        top_p = float(body.get("top_p", 0.95))
+
+        resp = await service.generate_text(api_key=api_key, prompt=prompt, max_tokens=max_tokens, temperature=temperature, top_k=top_k, top_p=top_p)
+        status = 200 if resp.success else 400
+        return JSONResponse(status_code=status, content=resp.dict())
+
+    @fast_app.post("/run/generate")
+    async def gradio_compatible_generate(req: Request):
+        """
+        Endpoint compatível com o formato 'Gradio' (data array).
+        Exemplo:
+        { "data": [ "gsk-..", "prompt...", 128, 0.7, 50, 0.95 ] }
+        """
+        try:
+            body = await req.json()
+        except Exception:
+            return JSONResponse(status_code=400, content={"success": False, "error": "Payload inválido (JSON esperado)."})
+
+        data = body.get("data")
+        if not isinstance(data, list):
+            return JSONResponse(status_code=400, content={"success": False, "error": "Campo 'data' inválido. Esperado array."})
+
+        # mapear por posições (compatível com a UI)
+        try:
+            api_key = data[0]
+            prompt = data[1] if len(data) > 1 else ""
+            max_tokens = int(data[2]) if len(data) > 2 else 512
+            temperature = float(data[3]) if len(data) > 3 else 0.7
+            top_k = int(data[4]) if len(data) > 4 else 50
+            top_p = float(data[5]) if len(data) > 5 else 0.95
+        except Exception as e:
+            return JSONResponse(status_code=400, content={"success": False, "error": f"Erro ao parsear 'data': {e}"})
+
+        resp = await service.generate_text(api_key=api_key, prompt=prompt, max_tokens=max_tokens, temperature=temperature, top_k=top_k, top_p=top_p)
+        status = 200 if resp.success else 400
+        return JSONResponse(status_code=status, content=resp.dict())
+
+    return fast_app
+
+
+# ----------------- Entrypoint -----------------
+async def build_and_run():
+    service = GemmaService()
+    await service.initialize()
+
+    interface = GradioInterface(service)
+    gradio_blocks = await interface.create_interface()
 
-        interface = GradioInterface(service)
-        app = await interface.create_interface()
+    fast_app = create_fastapi_app(gradio_blocks, service)
 
-        app.launch(server_name="0.0.0.0", server_port=7860, share=False, debug=False)
-    except Exception as e:
-        logger.critical(f"Falha ao iniciar a aplicação: {e}", exc_info=True)
+    # Quando executado localmente com "python app.py", usamos uvicorn para servir.
+    # No Hugging Face Spaces, o arquivo app.py será automaticamente usado (uvicorn não é necessário manualmente),
+    # mas manter este bloco para execução local.
+    return fast_app
 
 
 if __name__ == "__main__":
-    asyncio.run(main())
+    # Constrói app (inicializa modelo) e executa uvicorn
+    fast_app = asyncio.run(build_and_run())
+    uvicorn.run(fast_app, host="0.0.0.0", port=int(os.getenv("PORT", 7860)))