Add support model deepseek/deepseek-reasoner

[KennyDizi]

PR Type

Enhancement

---

Description

• Added support for the deepseek/deepseek-reasoner model with a token limit of 64K.

• Updated logic to combine system and user prompts for deepseek-reasoner model.

• Improved logging to specify the model name when combining prompts.

---

Changes walkthrough 📝

Relevant files

Enhancement

__init__.py Add `deepseek/deepseek-reasoner` model with token limit    pr_agent/algo/init.py Added deepseek/deepseek-reasoner model with a 64K token limit. +1/-0      litellm_ai_handler.py Enhance prompt handling for `deepseek-reasoner` model        pr_agent/algo/ai_handlers/litellm_ai_handler.py Included deepseek-reasoner in logic for combining system and user prompts. Updated logging to include model name when combining prompts. +2/-2

---

Need help?

Type /help how to ... in the comments thread for any question about Qodo Merge usage.

Check out the documentation for more information.

[qodo-merge-pro-for-open-source]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 1 🔵⚪⚪⚪⚪

🏅 Score: 95

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Code Clarity Consider extracting the model type check into a separate function for better readability and maintainability, since it now handles both O1 and deepseek-reasoner models if (model_type.startswith(O1_MODEL_PREFIX)) or ("deepseek-reasoner" in model): user = f"{system}\n\n\n{user}"

[qodo-merge-pro-for-open-source]

PR Code Suggestions ✨

Explore these optional code suggestions:

Category	Suggestion	Score
General	Improve code readability and maintainability The condition for combining prompts should be more explicit and separated for better maintainability. Create separate conditions for O1 models and deepseek-reasoner. pr_agent/algo/ai_handlers/litellm_ai_handler.py [203-206] -if (model_type.startswith(O1_MODEL_PREFIX)) or ("deepseek-reasoner" in model): +needs_combined_prompts = ( + model_type.startswith(O1_MODEL_PREFIX) or + "deepseek-reasoner" in model +) +if needs_combined_prompts: user = f"{system}\n\n\n{user}" system = "" get_logger().info(f"Using model {model}, combining system and user prompts") Apply this suggestion Suggestion importance[1-10]: 4 Why: The suggestion improves code readability by extracting the complex condition into a descriptive variable, making the logic clearer. However, the impact is relatively minor as the original code is already functional and reasonably clear.	4

[mrT23]

ok, but maybe in the future it worth further refactoring - a list of models that do not have 'system' message support, as this is getting bigger

[KennyDizi]

That's a good point! Do you have any ideas about this? I would like to improve the code.

[mrT23]

i think that in the place where we define models, we can define an extra list of 'user-message-only' models

and use that in the handler

[KennyDizi]

That sounds good. I will do that in the next PR.

[mrT23]

p.s - that's also the feedback from Qodo Merge

Improve code readability and maintainability

---

The condition for combining prompts should be more explicit and separated for better
maintainability. Create separate conditions for O1 models and deepseek-reasoner.

pr_agent/algo/ai_handlers/litellm_ai_handler.py [203-206]

-if (model_type.startswith(O1_MODEL_PREFIX)) or ("deepseek-reasoner" in model):
+needs_combined_prompts = (
+    model_type.startswith(O1_MODEL_PREFIX) or
+    "deepseek-reasoner" in model
+)
+if needs_combined_prompts:
     user = f"{system}\n\n\n{user}"
     system = ""
     get_logger().info(f"Using model {model}, combining system and user prompts")```

    
  
<ul class="contains-task-list">
<li class="task-list-item enabled"><input type="checkbox" id="" class="task-list-item-checkbox"> <strong>Apply this suggestion</strong> </li>
</ul>
<details><summary>Suggestion importance[1-10]: 4</summary>
<p dir="auto">Why: The suggestion improves code readability by extracting the complex condition into a descriptive variable, making the logic clearer. However, the impact is relatively minor as the original code is already functional and reasonably clear.</p>
</details></details>