fix: add add_generation_prompt parameter to ChatCompletionRequest

[zelong518]

Fix: Add add_generation_prompt parameter to ChatCompletionRequest

Problem Description

When sending requests through the router with continue_final_message: true set but without explicitly setting add_generation_prompt, the following error occurs:

continue_final_message and add_generation_prompt are not compatible. 
Use continue_final_message when you want the model to continue the final 
message, and add_generation_prompt when you want to add a header that will 
prompt it to start a new assistant message instead.

Root Cause

The ChatCompletionRequest struct in the router was missing the add_generation_prompt parameter. When this parameter was not included in the user's request, the router would not forward it to the backend, causing the backend to use its default value of true. If continue_final_message: true was also set, these two parameters would conflict.

This issue does not occur when sending requests directly to the backend model, as all parameters can be controlled directly.

Solution

Added the add_generation_prompt field to the ChatCompletionRequest struct with a default value of true (aligned with vLLM). This ensures:

1. The router can properly receive and forward the add_generation_prompt parameter
2. When users explicitly set add_generation_prompt: false, it can be used compatibly with continue_final_message: true
3. The default behavior is consistent with vLLM

Changes

• Added add_generation_prompt: bool field to ChatCompletionRequest struct
• Set default value to true using #[serde(default = "default_true")] to align with vLLM behavior

Testing

Verified the fix using the following curl command:

curl -X POST "http://127.0.0.1:30000/v1/chat/completions" \
     -H "Content-Type: application/json" \
     -d '{
           "model": "IQuestLoopCoder-Instruct",
           "messages": [
             {
               "role": "user",
               "content": "tell me a common saying"
             },
             {
               "role": "assistant",
               "content": "Here is a common saying about apple. An apple a day, keeps"
             }
           ],
           "add_generation_prompt": false,
           "continue_final_message": true,
           "echo": false
         }'

The request now works correctly without parameter conflict errors.

[Prowindy]

The fix looks good. could you fix buildkite?

[zelong518]

ok，i will