Upload app.py

app.py

@@ -1,5 +1,7 @@
 import gradio as gr
 from pydub import AudioSegment
+import google.generativeai as genai
+from google.generativeai.types import HarmCategory, HarmBlockThreshold
 import json
 import uuid
 import io
@@ -10,12 +12,6 @@ import pypdf
 import os
 import time
 from typing import List, Dict, Tuple
-import openai
-import logging
-
-# At the beginning of your script, set up logging
-logging.basicConfig(level=logging.INFO)
-logger = logging.getLogger(__name__)
 
 class PodcastGenerator:
     def __init__(self):
@@ -28,7 +24,7 @@ class PodcastGenerator:
         引數：
             prompt (str): 用於生成Podcast劇本的使用者輸入文字。
             language (str): Podcast指劇本所需的語言。
-            api_key (str): 用於訪問 SambaNova API 服務的 API 金鑰。
+            api_key (str): 用於訪問 Gemini AI 服務的 API 金鑰。
 
         返回：
             Dict: 包含以 JSON 格式生成Podcast劇本的字典。
@@ -36,7 +32,7 @@ class PodcastGenerator:
         異常：
             gr.Error: 如果 API 金鑰或速率限制出現問題。
 
-        此方法使用 SambaNova API 根據使用者的輸入生成Podcast劇本。
+        此方法使用 Gemini AI 模型根據使用者的輸入生成Podcast劇本。
         它處理語言選擇，使用適當的配置設定 AI 模型，並處理生成的響應。
         """
         # 定義一個示例JSON結構，用於指導AI生成類似格式的Podcast劇本
@@ -255,51 +251,51 @@ class PodcastGenerator:
         # 設定使用者提示，包含使用者輸入的內容
         user_prompt = f"Please generate a podcast script based on the following user input:\n{prompt}"
 
-        # 配置 SambaNova API client
-        if not api_key:
-            api_key = os.getenv("YOUR_API_TOKEN")
-        client = openai.OpenAI(
-            api_key=api_key,
-            base_url="https://api.sambanova.ai/v1",
+        # 設定訊息列表，包含使用者提示
+        messages = [
+            {"role": "user", "parts": [user_prompt]}
+        ]
+
+        # 配置Google Generative AI
+        genai.configure(api_key=api_key)
+
+        # 設定生成配置
+        generation_config = {
+            "temperature": 1,
+            "max_output_tokens": 8192,
+            "response_mime_type": "application/json",
+        }
+
+        # 建立生成模型實例
+        model = genai.GenerativeModel(
+            model_name="gemini-1.5-flash-002",
+            generation_config=generation_config,
+            safety_settings={
+                HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: HarmBlockThreshold.BLOCK_NONE,
+                HarmCategory.HARM_CATEGORY_HARASSMENT: HarmBlockThreshold.BLOCK_NONE,
+                HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: HarmBlockThreshold.BLOCK_NONE,
+                HarmCategory.HARM_CATEGORY_HATE_SPEECH: HarmBlockThreshold.BLOCK_NONE
+            },
+            system_instruction=system_prompt
         )
 
         # 嘗試生成內容
         try:
-            response = client.chat.completions.create(
-                model='Meta-Llama-3.1-405B-Instruct',
-                messages=[
-                    {"role": "system", "content": system_prompt},
-                    {"role": "user", "content": user_prompt}
-                ],
-                temperature=1
-            )
-            logger.info(f"API Response: {response}")
-            
-            if response.choices and len(response.choices) > 0:
-                generated_text = response.choices[0].message.content
-            else:
-                logger.warning("No content generated from the API")
-                raise ValueError("No content generated from the API")
-
+            response = await model.generate_content_async(messages)
         except Exception as e:
-            logger.error(f"Error generating script: {str(e)}")
             # 處理可能的錯誤
             if "API key not valid" in str(e):
-                raise gr.Error("Invalid API key. Please provide a valid SambaNova API key.")
+                raise gr.Error("Invalid API key. Please provide a valid Gemini API key.")
             elif "rate limit" in str(e).lower():
-                raise gr.Error("Rate limit exceeded for the API key. Please try again later or provide your own SambaNova API key.")
+                raise gr.Error("Rate limit exceeded for the API key. Please try again later or provide your own Gemini API key.")
             else:
-                raise gr.Error(f"Failed to generate podcast script: {str(e)}")
+                raise gr.Error(f"Failed to generate podcast script: {e}")
 
         # 列印生成的Podcast指令碼
-        print(f"Generated podcast script:\n{generated_text}")
+        print(f"Generated podcast script:\n{response.text}")
 
-        # 嘗試解析JSON，如果失敗則返回原始文本
-        try:
-            return json.loads(generated_text)
-        except json.JSONDecodeError:
-            print("Warning: Generated text is not valid JSON. Returning raw text.")
-            return {"raw_text": generated_text}
+        # 返回解析後的JSON資料
+        return json.loads(response.text)
 
     async def tts_generate(self, text: str, speaker: int, speaker1: str, speaker2: str) -> str:
         """
@@ -446,9 +442,6 @@ async def process_input(input_text: str, input_file, language: str, speaker1: st
 
     # 定義語音名稱對映
     voice_names = {
-        "臺女1 - Chinese Taiwanese (Taiwan)": "zh-TW-HsiaoChenNeural",
-        "臺女2 - Chinese Taiwanese (Taiwan)": "zh-TW-HsiaoYuNeural",
-        "臺男 - Chinese Taiwanese (Taiwan)": "zh-TW-YunJheNeural",
         "Andrew - English (United States)": "en-US-AndrewMultilingualNeural",
         "Ava - English (United States)": "en-US-AvaMultilingualNeural",
         "Brian - English (United States)": "en-US-BrianMultilingualNeural",
@@ -489,9 +482,10 @@ iface = gr.Interface(
         gr.File(label="Or Upload a PDF or TXT file"),
         gr.Dropdown(label="Language", choices=[
             "Auto Detect",
-            "Chinese Taiwanese", "Afrikaans", "Albanian", "Amharic", "Arabic", "Armenian", "Azerbaijani",
+            "Afrikaans", "Albanian", "Amharic", "Arabic", "Armenian", "Azerbaijani",
             "Bahasa Indonesian", "Bangla", "Basque", "Bengali", "Bosnian", "Bulgarian",
-            "Burmese", "Catalan", "Chinese Cantonese", "Chinese Mandarin", "Croatian", "Czech", "Danish", "Dutch", "English",
+            "Burmese", "Catalan", "Chinese Cantonese", "Chinese Mandarin",
+            "Chinese Taiwanese", "Croatian", "Czech", "Danish", "Dutch", "English",
             "Estonian", "Filipino", "Finnish", "French", "Galician", "Georgian",
             "German", "Greek", "Hebrew", "Hindi", "Hungarian", "Icelandic", "Irish",
             "Italian", "Japanese", "Javanese", "Kannada", "Kazakh", "Khmer", "Korean",