Update app.py

app.py

@@ -1,111 +1,82 @@
 import os
-import time
-import shutil
-from pydub import AudioSegment
-from openai import OpenAI
 import gradio as gr
-from fastapi import FastAPI, File, UploadFile
+from fastapi import FastAPI, UploadFile, Form, HTTPException
+from openai import OpenAI
 
-# ========================
-# 🔐 基本設定
-# ========================
-PASSWORD = os.getenv("APP_PASSWORD", "defaultpass")
-MAX_SIZE = 25 * 1024 * 1024
+# 初始化 FastAPI + Gradio
+app = FastAPI()
 client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
 
-# FastAPI App for捷徑 API
-app = FastAPI()
+# 載入安全金鑰（請在 Hugging Face → Settings → Secrets 裡設 APP_PASSWORD）
+APP_PASSWORD = os.getenv("APP_PASSWORD", None)
+
+# === API endpoint ===
+@app.post("/api/transcribe")
+async def transcribe_api(file: UploadFile, token: str = Form(...)):
+    # 驗證 token
+    if not APP_PASSWORD:
+        raise HTTPException(status_code=500, detail="Server misconfiguration: APP_PASSWORD not set.")
+    if token != APP_PASSWORD:
+        raise HTTPException(status_code=403, detail="Forbidden: invalid token.")
+
+    # 儲存臨時音訊檔
+    contents = await file.read()
+    temp_path = f"/tmp/{file.filename}"
+    with open(temp_path, "wb") as f:
+        f.write(contents)
+
+    # 語音轉文字
+    with open(temp_path, "rb") as audio_file:
+        transcript = client.audio.transcriptions.create(
+            model="whisper-1",
+            file=audio_file
+        )
 
-# ========================
-# 🎧 音訊轉錄核心
-# ========================
-def split_audio_if_needed(path: str):
-    size = os.path.getsize(path)
-    if size <= MAX_SIZE:
-        return [path]
-    audio = AudioSegment.from_file(path)
-    n = int(size / MAX_SIZE) + 1
-    chunk_ms = len(audio) / n
-    parts = []
-    for i in range(n):
-        fn = f"chunk_{i+1}.wav"
-        audio[int(i * chunk_ms):int((i + 1) * chunk_ms)].export(fn, format="wav")
-        parts.append(fn)
-    return parts
-
-def transcribe_core(path: str, model: str = "whisper-1"):
-    if path.lower().endswith(".mp4"):
-        fixed = path[:-4] + ".m4a"
-        try:
-            shutil.copy(path, fixed)
-            path = fixed
-            print("🔧 已自動修正 mp4 → m4a")
-        except Exception as e:
-            print(f"⚠️ mp4→m4a 轉檔失敗：{e}")
-
-    chunks = split_audio_if_needed(path)
-    txts = []
-    for f in chunks:
-        with open(f, "rb") as af:
-            t = client.audio.transcriptions.create(
-                model=model, file=af, response_format="text"
-            )
-            txts.append(t)
-    full = "\n".join(txts)
-    summ = client.chat.completions.create(
+    text = transcript.text.strip()
+
+    # 簡短摘要
+    summary_prompt = f"請幫我用中文摘要以下內容：\n\n{text}"
+    summary = client.chat.completions.create(
         model="gpt-4o-mini",
-        messages=[{"role": "user", "content": f"請用繁體中文摘要以下內容：\n{full}"}],
-        temperature=0.4,
+        messages=[{"role": "user", "content": summary_prompt}]
     ).choices[0].message.content.strip()
-    return full, summ
 
-# ========================
-# 🌐 FastAPI 端點（捷徑用）
-# ========================
-@app.post("/api/transcribe")
-async def api_transcribe(file: UploadFile = File(...)):
-    """供 iPhone 捷徑上傳音訊並取得 JSON"""
-    temp = file.filename
-    with open(temp, "wb") as f:
-        f.write(await file.read())
-    text, summary = transcribe_core(temp)
-    os.remove(temp)
     return {"text": text, "summary": summary}
 
-@app.get("/health")
-def health():
-    """捷徑可先 ping 這個確認服務運作中"""
-    return {"status": "ok", "time": int(time.time())}
-
-# ========================
-# 💬 Gradio 介面
-# ========================
-def transcribe_with_pw(password, file):
-    if password.strip() != PASSWORD:
-        return "❌ 密碼錯誤", "", ""
-    if not file:
-        return "⚠️ 未選擇檔案", "", ""
-    text, summary = transcribe_core(file.name)
-    return "✅ 完成", text, summary
-
-with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    gr.Markdown("## 🎧 LINE 語音轉錄與摘要工具（支援 .m4a / .mp4）")
-    pw = gr.Textbox(label="輸入密碼", type="password")
-    f = gr.File(label="上傳音訊檔 (.m4a/.mp3/.wav/.mp4)")
-    run = gr.Button("開始轉錄 🚀")
-    s = gr.Textbox(label="狀態", interactive=False)
-    t = gr.Textbox(label="逐字稿", lines=10)
-    su = gr.Textbox(label="摘要", lines=8)
-
-    run.click(transcribe_with_pw, [pw, f], [s, t, su])
-
-# ========================
-# 🚀 啟動（單一埠）
-# ========================
-# 讓 Gradio 介面掛載到 FastAPI
-gr.mount_gradio_app(app, demo, path="/")
-
-# Hugging Face 自動綁定 port=7860，不用手動設定
+
+# === Gradio 前端 ===
+def transcribe_ui(audio):
+    if audio is None:
+        return "請上傳音訊檔案", ""
+    with open(audio, "rb") as f:
+        transcript = client.audio.transcriptions.create(
+            model="whisper-1",
+            file=f
+        )
+    text = transcript.text.strip()
+
+    summary_prompt = f"請幫我用中文摘要以下內容：\n\n{text}"
+    summary = client.chat.completions.create(
+        model="gpt-4o-mini",
+        messages=[{"role": "user", "content": summary_prompt}]
+    ).choices[0].message.content.strip()
+
+    return text, summary
+
+
+demo = gr.Interface(
+    fn=transcribe_ui,
+    inputs=gr.Audio(type="filepath", label="上傳音訊"),
+    outputs=[
+        gr.Textbox(label="轉錄結果"),
+        gr.Textbox(label="AI 摘要")
+    ],
+    title="LINE 語音轉錄與摘要 (安全版)",
+    description="上傳 LINE 語音或其他音訊，進行自動轉錄與摘要"
+)
+
+# 掛載 Gradio 到 FastAPI
+app = gr.mount_gradio_app(app, demo, path="/")
+
 if __name__ == "__main__":
-    import uvicorn
-    uvicorn.run(app, host="0.0.0.0", port=7860)
+    demo.launch(server_name="0.0.0.0", server_port=7860)