Create app.py

app.py

@@ -0,0 +1,216 @@
+import gradio as gr
+import openai
+from pydub import AudioSegment
+import os
+import re
+
+# 轉譯音檔函式，使用 OpenAI Whisper 模型
+def transcribe(filename, api_key):
+    client = openai.OpenAI(api_key=api_key)
+    with open(filename, "rb") as audio_file:
+        transcript_txt = client.audio.transcriptions.create(
+            model="whisper-1",
+            file=audio_file,
+            response_format="text"
+        )
+    return transcript_txt
+
+# 檢查檔案大小並分段處理（若檔案過大）
+def transcribe_large_audio(filename, api_key, segment_length_ms=30 * 60 * 1000):
+    def get_file_size_in_mb(file_path):
+        return os.path.getsize(file_path) / (1024 * 1024)
+
+    def split_audio_file(file_path, segment_length_ms=30 * 60 * 1000):
+        audio = AudioSegment.from_file(file_path, format="mp3")
+        segment_filenames = []
+        for i in range(0, len(audio), segment_length_ms):
+            end = min(i + segment_length_ms, len(audio))
+            segment = audio[i:end]
+            segment_filename = f"{file_path}_part{len(segment_filenames) + 1}.mp3"
+            segment.export(segment_filename, format="mp3", bitrate="36k")
+            segment_filenames.append(segment_filename)
+        return segment_filenames
+
+    transcript_txt = ""
+
+    if get_file_size_in_mb(filename) > 25:
+        audio_chunks = split_audio_file(filename)
+        for chunk_filename in audio_chunks:
+            transcript_txt += transcribe(chunk_filename, api_key)
+            os.remove(chunk_filename)
+    else:
+        transcript_txt = transcribe(filename, api_key)
+
+    return transcript_txt
+
+# 字數統計函式
+def count_words(text):
+    # 移除空白字符後計算字數
+    cleaned_text = re.sub(r'\s+', '', text)
+    return len(cleaned_text)
+
+# 自動斷句與標點符號功能
+def auto_punctuate(text, api_key):
+    openai.api_key = api_key
+    prompt = """請幫我將以下逐字稿加入適當的標點符號和段落分隔，使文本更容易閱讀：
+    原文：
+    """ + text
+    
+    completion = openai.chat.completions.create(
+        model="gpt-4",
+        messages=[{"role": "user", "content": prompt}]
+    )
+    return completion.choices[0].message.content
+
+# OpenAI 翻譯文本的函式
+def openai_translate_text(text, target_lang, api_key):
+    openai.api_key = api_key
+    language_mapping = {
+        "繁體中文": "繁體中文",
+        "英文": "英文",
+        "日文": "日文",
+        "韓文": "韓文",
+        "法文": "法文",
+        "德文": "德文",
+        "西班牙文": "西班牙文"
+    }
+    
+    prompt = f"請將以下文本翻譯成{language_mapping[target_lang]}，保持原文的語氣和風格：\n\n{text}"
+    completion = openai.chat.completions.create(
+        model="gpt-4",
+        messages=[{"role": "user", "content": prompt}]
+    )
+    return completion.choices[0].message.content
+
+# OpenAI 生成摘要的函式（根據不同受眾）
+def openai_generate_summary(text, audience_type, api_key):
+    openai.api_key = api_key
+    audience_prompts = {
+        "學生": "請將以下會議內容整理成適合學生閱讀的摘要，重點放在學習價值和知識傳遞：",
+        "老師": "請將以下會議內容整理成適合教師參考的摘要，重點放在教學應用和教育意義：",
+        "會議": "請將以下會議內容整理成正式的會議摘要，重點放在決策、行動項目和重要討論：",
+        "主管": "請將以下會議內容整理成適合管理層閱讀的摘要，重點放在策略決策和關鍵績效：",
+        "技術人員": "請將以下會議內容整理成適合技術團隊閱讀的摘要，重點放在技術細節和實作方向：",
+        "行銷人員": "請將以下會議內容整理成適合行銷團隊閱讀的摘要，重點放在市場策略和推廣重點：",
+        "一般員工": "請將以下會議內容整理成適合一般員工閱讀的摘要，重點放在執行重點和日常工作相關內容：",
+        "客戶": "請將以下會議內容整理成適合客戶閱讀的摘要，重點放在服務優化和價值傳遞：",
+        "投資者": "請將以下會議內容整理成適合投資者閱讀的摘要，重點放在財務表現和未來展望：",
+        "研究人員": "請將以下會議內容整理成適合研究人員閱讀的摘要，重點放在研究方法和數據分析："
+    }
+    
+    prompt = f"{audience_prompts[audience_type]}\n\n{text}"
+    completion = openai.chat.completions.create(
+        model="gpt-4",
+        messages=[{"role": "user", "content": prompt}]
+    )
+    return completion.choices[0].message.content
+
+# 處理音檔轉文字
+def process_audio(audio_file, api_key):
+    if not audio_file or not api_key:
+        return "請確保上傳音檔和輸入API金鑰", "", 0
+    
+    try:
+        # 轉譯音檔
+        transcript = transcribe_large_audio(audio_file, api_key)
+        # 計算字數
+        word_count = count_words(transcript)
+        return transcript, f"字數統計：{word_count} 字", word_count
+    except Exception as e:
+        return f"處理失敗：{str(e)}", "", 0
+
+# 處理標點與段落
+def process_punctuation(text, api_key):
+    if not text or not api_key:
+        return "請確保有文本內容和API金鑰"
+    try:
+        return auto_punctuate(text, api_key)
+    except Exception as e:
+        return f"標點處理失敗：{str(e)}"
+
+# 處理翻譯
+def process_translation(text, target_lang, api_key):
+    if not all([text, target_lang, api_key]):
+        return "請確保所有必要欄位都已填寫"
+    try:
+        return openai_translate_text(text, target_lang, api_key)
+    except Exception as e:
+        return f"翻譯失敗：{str(e)}"
+
+# 處理摘要
+def process_summary(text, audience_type, api_key):
+    if not all([text, audience_type, api_key]):
+        return "請確保所有必要欄位都已填寫"
+    try:
+        return openai_generate_summary(text, audience_type, api_key)
+    except Exception as e:
+        return f"摘要生成失敗：{str(e)}"
+
+# 建立 Gradio 介面
+with gr.Blocks() as demo:
+    gr.Markdown("會議音檔轉文字處理系統")
+    
+    with gr.Row():
+        audio_file_input = gr.Audio(type="filepath", label="上傳音檔")
+        api_key_input = gr.Textbox(label="輸入 OpenAI API 金鑰", type="password")
+    
+    with gr.Row():
+        transcript_output = gr.Textbox(label="原始逐字稿", lines=5)
+        word_count_output = gr.Textbox(label="字數統計")
+    
+    with gr.Row():
+        punctuated_output = gr.Textbox(label="加入標點符號後的文本", lines=5)
+    
+    with gr.Row():
+        target_lang_input = gr.Dropdown(
+            choices=["繁體中文", "英文", "日文", "韓文", "法文", "德文", "西班牙文"],
+            label="選擇目標語言",
+            value="繁體中文"
+        )
+        translated_output = gr.Textbox(label="翻譯結果", lines=5)
+    
+    with gr.Row():
+        audience_type_input = gr.Dropdown(
+            choices=[
+                "學生", "老師", "會議", "主管", "技術人員", 
+                "行銷人員", "一般員工", "客戶", "投資者", "研究人員"
+            ],
+            label="選擇摘要類型",
+            value="會議"
+        )
+        summary_output = gr.Textbox(label="客製化摘要", lines=5)
+    
+    with gr.Row():
+        transcribe_button = gr.Button("1. 開始轉譯")
+        punctuate_button = gr.Button("2. 添加標點符號")
+        translate_button = gr.Button("3. 翻譯文本")
+        summary_button = gr.Button("4. 生成摘要")
+
+    # 連接按鈕和函式
+    transcribe_button.click(
+        process_audio,
+        inputs=[audio_file_input, api_key_input],
+        outputs=[transcript_output, word_count_output]
+    )
+    
+    punctuate_button.click(
+        process_punctuation,
+        inputs=[transcript_output, api_key_input],
+        outputs=punctuated_output
+    )
+    
+    translate_button.click(
+        process_translation,
+        inputs=[punctuated_output, target_lang_input, api_key_input],
+        outputs=translated_output
+    )
+    
+    summary_button.click(
+        process_summary,
+        inputs=[translated_output, audience_type_input, api_key_input],
+        outputs=summary_output
+    )
+
+# 啟動介面
+if __name__ == "__main__":
+    demo.launch()