Create app.py

app.py

@@ -0,0 +1,66 @@
+# file: app.py
+import os
+from pydub import AudioSegment
+from openai import OpenAI
+import gradio as gr
+
+client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
+MAX_SIZE = 25 * 1024 * 1024
+
+def split_audio_if_needed(input_path):
+    size = os.path.getsize(input_path)
+    if size <= MAX_SIZE:
+        return [input_path]
+
+    audio = AudioSegment.from_file(input_path)
+    num_chunks = int(size / MAX_SIZE) + 1
+    chunk_length = len(audio) / num_chunks
+
+    chunk_files = []
+    for i in range(num_chunks):
+        start = int(i * chunk_length)
+        end = int((i + 1) * chunk_length)
+        chunk = audio[start:end]
+        chunk_filename = f"chunk_{i+1}.wav"
+        chunk.export(chunk_filename, format="wav")
+        chunk_files.append(chunk_filename)
+    return chunk_files
+
+def transcribe_and_summarize(file):
+    if not file:
+        return "請上傳音訊檔。", ""
+    chunks = split_audio_if_needed(file)
+    transcripts = []
+    for idx, f in enumerate(chunks, 1):
+        with open(f, "rb") as audio_file:
+            text = client.audio.transcriptions.create(
+                model="whisper-1",
+                file=audio_file,
+                response_format="text"
+            )
+        transcripts.append(text)
+    full_text = "\n".join(transcripts)
+
+    response = client.chat.completions.create(
+        model="gpt-4o-mini",
+        messages=[
+            {"role": "system", "content": "你是一位精準且擅長摘要的助手。"},
+            {"role": "user", "content": "請用繁體中文摘要以下內容：\n" + full_text}
+        ],
+        temperature=0.4,
+    )
+    summary = response.choices[0].message.content.strip()
+    return full_text, summary
+
+iface = gr.Interface(
+    fn=transcribe_and_summarize,
+    inputs=gr.Audio(type="filepath", label="上傳音檔 (.m4a, .aac, .wav)"),
+    outputs=[
+        gr.Textbox(label="完整轉錄文字", lines=10),
+        gr.Textbox(label="摘要結果", lines=10),
+    ],
+    title="語音轉錄與摘要工具",
+    description="上傳音檔後，自動轉錄成文字並生成摘要（繁體中文）。"
+)
+
+iface.launch()