Does the 'pr-agent' version (0.3.1) not support dashscope models?

[luoyq-Rockey]

Git provider

Gitlab

System Info

'pr-agent' version : 0.3.1
pc : windows 11
IDE:pycharm 2025
python version: 3.13
local gitlab version: 18.5.1

Bug details

I upgraded my local pr-agent from 0.3.0 to 0.3.1, and then made a code comparison between these two versions. I found that version 0.3.1 is in the code of pr-agent\pr_agent\algo\ai_handlers\litellm_ai_handler.py. Support dashscope models has been cancelled

`(.venv) PS D:\Code\pr-agent-0.31\pr_agent> python cli.py --pr_url=http://172.0.0.1:8081/root/lyq_test/-/merge_requests/10 improve --verbose
D:\Code\pr-agent-0.31.venv\Lib\site-packages\azure_init_.py:5: UserWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html. The pkg_resources package is slated for removal as early as 2025-11-30. Refrain from using this package or pin to Setuptools<81.
import pkg_resources
2025-11-24 12:47:36.316 | INFO | pr_agent.tools.pr_code_suggestions:run:99 - Generating code suggestions for PR...
2025-11-24 12:47:40.756 | INFO | pr_agent.tools.pr_code_suggestions:prepare_prediction_main:698 - Number of PR chunk calls: 1

Give Feedback / Get Help: https://github.com/BerriAI/litellm/issues/new
LiteLLM.Info: If you need to debug this error, use `litellm._turn_on_debug()'.

2025-11-24 12:47:41.482 | WARNING | pr_agent.algo.ai_handlers.litellm_ai_handler:chat_completion:396 - Error during LLM inference: litellm.NotFoundError: NotFoundError: DashscopeException - model 'deepseek-v3.1' not found

Give Feedback / Get Help: https://github.com/BerriAI/litellm/issues/new
LiteLLM.Info: If you need to debug this error, use `litellm._turn_on_debug()'.

2025-11-24 12:47:41.496 | WARNING | pr_agent.algo.ai_handlers.litellm_ai_handler:chat_completion:396 - Error during LLM inference: litellm.NotFoundError: NotFoundError: DashscopeException - model 'deepseek-v3.1' not found
2025-11-24 12:47:41.496 | WARNING | pr_agent.algo.pr_processing:retry_with_fallback_models:333 - Failed to generate prediction with dashscope/deepseek-v3.1
2025-11-24 12:47:41.501 | ERROR | pr_agent.tools.pr_code_suggestions:run:185 - Failed to generate code suggestions for PR, error: Failed to generate prediction with any model of ['dashscope/deepseek-v3.1']`

If I copy the difference code in pr-agent\pr_agent\algo\ai_handlers\litellm_ai_handler.py to 0.3.1, the same error will still occur

Is this a defect or will dashscope no longer be supported?

[qodo-merge-for-open-source]

💡 Auto-generated documentation-based answer:

Based on the error message you're seeing (NotFoundError: DashscopeException - model 'deepseek-v3.1' not found), this appears to be a model identification issue rather than a complete lack of dashscope support.

The error suggests that litellm (which PR-Agent uses for model interactions) is not recognizing the model identifier you're using. Here are the key points:

1. Model Name Format: Ensure you're using the correct model identifier format. For dashscope models, the format should typically be dashscope/model-name. Your configuration shows dashscope/deepseek-v3.1, but you may need to verify the exact model name with dashscope's documentation.

2. Required Environment Variables: According to the litellm documentation, dashscope models require specific environment variables to be set. Make sure you have configured:

   • The dashscope API key
   • Any other required dashscope-specific parameters

3. Configuration: In your .secrets.toml or environment variables, you need to set the dashscope credentials. The exact variable names may vary, so check the litellm documentation for the current requirements.

4. Troubleshooting Steps:

   • Verify the exact model name is correct (e.g., is it deepseek-v3.1 or deepseek-v3 or something else?)
   • Ensure all required API keys and environment variables are properly set
   • Check that the model is actually available through your dashscope account

The documentation notes: "Failing to set the needed keys of a specific model will usually result in litellm not identifying the model type, and failing to utilize it."

To get more specific help, it would be useful to know:

• What dashscope API credentials you have configured
• The exact model identifier you're trying to use
• Whether this model worked in version 0.3.0 with the same configuration

Relevant Sources:

• https://qodo-merge-docs.qodo.ai//usage-guide/changing_a_model#changing-a-model-in-pr-agent
• https://qodo-merge-docs.qodo.ai//usage-guide/changing_a_model
• https://qodo-merge-docs.qodo.ai//installation/locally#i-get-an-error-when-running-the-docker-image-what-should-i-do

[luoyq-Rockey]

All my configurations are the same as those of v0.30. It works normally on V0.30, but the above error will occur on V0.31