Description
在 Windows 10 + Docker Desktop 环境下，rdagent fin_quant 自动拼接的挂载路径为 /tmp/full:D:\workspace\qlib_workspace\workspace_cache:rw，导致 Docker 报错“too many colons”。我已测试 docker run -v E:/workspace/qlib_workspace/workspace_cache:/test ubuntu ls /test 没有问题，.env 里路径也写成了绝对路径和 Linux 路径，但依然报错。请官方修复 Windows 下的路径拼接逻辑。
(rdagent) E:>rdagent fin_quant
2025-06-26 14:14:42.610 | INFO | rdagent.oai.backend.litellm::41 - backend='rdagent.oai.backend.LiteLLMAPIBackend' chat_model='deepseek/deepseek-chat' embedding_model='litellm_proxy/BAAI/bge-large-en-v1.5' reasoning_effort=None reasoning_think_rm=True log_llm_chat_content=True use_azure=False chat_use_azure=False embedding_use_azure=False chat_use_azure_token_provider=False embedding_use_azure_token_provider=False managed_identity_client_id=None max_retry=10 retry_wait_seconds=20 dump_chat_cache=False use_chat_cache=False dump_embedding_cache=False use_embedding_cache=False prompt_cache_path='E:\prompt_cache.db' max_past_message_include=10 timeout_fail_limit=10 violation_fail_limit=1 use_auto_chat_cache_seed_gen=False init_chat_cache_seed=42 openai_api_key='' chat_openai_api_key=None chat_openai_base_url=None chat_azure_api_base='' chat_azure_api_version='' chat_max_tokens=None chat_temperature=0.5 chat_stream=True chat_seed=None chat_frequency_penalty=0.0 chat_presence_penalty=0.0 chat_token_limit=100000 default_system_prompt="You are an AI assistant who helps to answer user's questions." system_prompt_role='system' embedding_openai_api_key='' embedding_openai_base_url='' embedding_azure_api_base='' embedding_azure_api_version='' embedding_max_str_num=50 use_llama2=False llama2_ckpt_dir='Llama-2-7b-chat' llama2_tokenizer_path='Llama-2-7b-chat/tokenizer.model' llams2_max_batch_size=8 use_gcr_endpoint=False gcr_endpoint_type='llama2_70b' llama2_70b_endpoint='' llama2_70b_endpoint_key='' llama2_70b_endpoint_deployment='' llama3_70b_endpoint='' llama3_70b_endpoint_key='' llama3_70b_endpoint_deployment='' phi2_endpoint='' phi2_endpoint_key='' phi2_endpoint_deployment='' phi3_4k_endpoint='' phi3_4k_endpoint_key='' phi3_4k_endpoint_deployment='' phi3_128k_endpoint='' phi3_128k_endpoint_key='' phi3_128k_endpoint_deployment='' gcr_endpoint_temperature=0.7 gcr_endpoint_top_p=0.9 gcr_endpoint_do_sample=False gcr_endpoint_max_token=100 chat_use_azure_deepseek=False chat_azure_deepseek_endpoint='' chat_azure_deepseek_key='' chat_model_map={}
2025-06-26 14:14:43.840 | INFO | rdagent.utils.env:prepare:714 - Building the image from dockerfile: D:\Anaconda\envs\rdagent\lib\site-packages\rdagent\scenarios\qlib\docker
⠋ Successfully tagged local_qlib:latest
2025-06-26 14:14:43.888 | INFO | rdagent.utils.env:prepare:732 - Finished building the image from dockerfile: D:\Anaconda\envs\rdagent\lib\site-packages\rdagent\scenarios\qlib\docker
2025-06-26 14:14:43.894 | INFO | rdagent.utils.env:prepare:897 - We are downloading!
2025-06-26 14:14:44.459 | INFO | rdagent.utils.env:_f:791 - GPU Devices are available.
2025-06-26 14:14:44.465 | WARNING | rdagent.utils.env:__run_ret_code_with_retry:197 - Error while running the container: Error while running the container: 500 Server Error for http+docker://localnpipe/v1.49/containers/create: Internal Server Error ("mount denied:
the source path "/tmp/full:E:\workspace\qlib_workspace\workspace_cache:rw"
too many colons"), current try index: 1, 4 retries left.