apunkt/plano
1
0
Fork 0
mirror of https://github.com/katanemo/plano.git synced 2026-07-17 16:31:04 +02:00
Code Issues Projects Releases Packages Wiki Activity

adding support for model aliases in archgw

Browse source
This commit is contained in:
Salman Paracha 2025-09-14 22:30:57 -07:00
parent 1e8c81d8f6
commit f13b420146
13 changed files with 1438 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

19
crates/hermesllm/src/apis/openai.rs
View file

@ -104,6 +104,20 @@ pub struct ChatCompletionsRequest {
// pub web_search: Option<bool>, // GOOD FIRST ISSUE: Future support for web search
}
impl ChatCompletionsRequest {
/// Suppress max_tokens if the model is o3, o3-*, openrouter/o3, or openrouter/o3-*
pub fn suppress_max_tokens_if_o3(&mut self) {
let model = self.model.as_str();
let is_o3 = model == "o3"
|| model.starts_with("o3-")
|| model == "openrouter/o3"
|| model.starts_with("openrouter/o3-");
if is_o3 {
self.max_tokens = None;
}
}
}
// ============================================================================
// CHAT COMPLETIONS API TYPES
// ============================================================================
@ -530,7 +544,10 @@ impl TryFrom<&[u8]> for ChatCompletionsRequest {
type Error = OpenAIStreamError;
fn try_from(bytes: &[u8]) -> Result<Self, Self::Error> {
serde_json::from_slice(bytes).map_err(OpenAIStreamError::from)
let mut req: ChatCompletionsRequest = serde_json::from_slice(bytes).map_err(OpenAIStreamError::from)?;
// Use the centralized suppression logic
req.suppress_max_tokens_if_o3();
Ok(req)
}
}

6
crates/hermesllm/src/clients/transformer.rs
View file

@ -97,7 +97,7 @@ impl TryFrom<AnthropicMessagesRequest> for ChatCompletionsRequest {
let openai_tools = req.tools.map(|tools| convert_anthropic_tools(tools));
let (openai_tool_choice, parallel_tool_calls) = convert_anthropic_tool_choice(req.tool_choice);
Ok(ChatCompletionsRequest {
let mut _chat_completions_req: ChatCompletionsRequest = ChatCompletionsRequest {
model: req.model,
messages: openai_messages,
temperature: req.temperature,
@ -109,7 +109,9 @@ impl TryFrom<AnthropicMessagesRequest> for ChatCompletionsRequest {
tool_choice: openai_tool_choice,
parallel_tool_calls,
..Default::default()
})
};
_chat_completions_req.suppress_max_tokens_if_o3();
Ok(_chat_completions_req)
}
}