本地Jupyter运行Gemini API教程遇两类错误求助

我正在学习Gemini API的Python版Jupyter Notebook教程，该教程在Google Colab中运行正常，但在本地Python 3.11的Anaconda Jupyter环境中运行时出现两个异常，以下是问题详情及排查解决建议：

错误1：执行模型列表代码时触发AttributeError #

触发场景 #

执行以下代码时出错：

for m in genai.list_models():
  if 'generateContent' in m.supported_generation_methods:
    print(m.name)

错误日志 #

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In[28], line 1
----> 1 for m in genai.list_models():
      2   if 'generateContent' in m.supported_generation_methods:
      3     print(m.name)

File /opt/anaconda3/envs/base_ML_AI/lib/python3.11/site-packages/google/generativeai/models.py:191, in list_models(page_size, client, request_options)
    188     client = get_default_model_client()
    190 for model in client.list_models(page_size=page_size, **request_options):
--> 191     model = type(model).to_dict(model)
    192     yield model_types.Model(**model)
AttributeError: to_dict

排查与解决建议 #

• 检查并升级google-generativeai库：Colab环境通常保持最新版本，本地环境可能版本滞后。执行以下命令更新到最新稳定版：

  pip install --upgrade google-generativeai
  

• 清理依赖冲突：Anaconda环境可能存在旧版Google客户端库冲突，更新核心依赖确保兼容性：

  pip install --upgrade google-api-core google-auth
  

• 手动处理模型对象：若升级后仍有问题，显式转换模型对象格式绕过to_dict方法：

  import google.generativeai as genai
  from google.generativeai.client import get_default_model_client
  
  client = get_default_model_client()
  for model in client.list_models():
      model_dict = dict(model)
      m = genai.types.Model(**model_dict)
      if 'generateContent' in m.supported_generation_methods:
          print(m.name)
  

错误2：提取响应文本时触发TypeError #

触发场景 #

执行to_markdown(response.text)时出错。

错误日志 #

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Cell In[42], line 1
----> 1 to_markdown(response.text)

File /opt/anaconda3/envs/base_ML_AI/lib/python3.11/site-packages/google/generativeai/types/generation_types.py:407, in BaseGenerateContentResponse.text(self)
    400 if not parts:
    401     raise ValueError(
    402         "The `response.text` quick accessor only works when the response contains a valid "
    403         "`Part`, but none was returned. Check the `candidate.safety_ratings` to see if the "
    404         "response was blocked."
    405     )
--> 407 if len(parts) != 1 or "text" not in parts[0]:
    408     raise ValueError(
    409         "The `response.text` quick accessor only works for "
    410         "simple (single-`Part`) text responses. This response is not simple text. "
   (...)
    413         "instead."
    414     )
    415 return parts[0].text

TypeError: argument of type 'Part' is not iterable

排查与解决建议 #

• 升级google-generativeai库：该错误多为旧版本库中Part对象的迭代判断逻辑bug导致，升级到最新版本即可修复。
• 手动提取响应文本：若暂时无法升级，绕过response.text快捷属性，直接从响应结构中提取：

  # 替代response.text的手动提取方式
  text = response.candidates[0].content.parts[0].text
  to_markdown(text)
  

• 验证响应结构：打印response.candidates[0].content.parts确认返回内容为纯文本类型Part，排除因安全拦截或格式不匹配导致的异常。

内容的提问来源于stack exchange，提问作者Janek Podwysocki