apunkt/plano
1
0
Fork 0
mirror of https://github.com/katanemo/plano.git synced 2026-05-02 12:22:43 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

fixed changes related to max_tokens and processing http error codes like 400 properly (#574)

Browse source 
Co-authored-by: Salman Paracha <salmanparacha@MacBook-Pro-257.local>
This commit is contained in:
Salman Paracha 2025-09-25 17:00:37 -07:00 • committed by GitHub
parent 7ce8d44d8e
commit 03c2cf6f0d
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
6 changed files with 157 additions and 23 deletions
Download patch file Download diff file Expand all files Collapse all files

8
crates/hermesllm/src/clients/transformer.rs
View file

@ -102,7 +102,7 @@ impl TryFrom<AnthropicMessagesRequest> for ChatCompletionsRequest {
messages: openai_messages,
temperature: req.temperature,
top_p: req.top_p,
max_tokens: Some(req.max_tokens),
max_completion_tokens: Some(req.max_tokens),
stream: req.stream,
stop: req.stop_sequences,
tools: openai_tools,
@ -142,7 +142,9 @@ impl TryFrom<ChatCompletionsRequest> for AnthropicMessagesRequest {
model: req.model,
system: system_prompt,
messages,
max_tokens: req.max_tokens.unwrap_or(DEFAULT_MAX_TOKENS),
max_tokens: req.max_completion_tokens
.or(req.max_tokens)
.unwrap_or(DEFAULT_MAX_TOKENS),
container: None,
mcp_servers: None,
service_tier: None,
@ -1079,7 +1081,7 @@ mod tests {
assert_eq!(openai_req.model, "claude-3-sonnet-20240229");
assert_eq!(openai_req.messages.len(), 2); // system + user message
assert_eq!(openai_req.max_tokens, Some(1024));
assert_eq!(openai_req.max_completion_tokens, Some(1024));
assert_eq!(openai_req.temperature, Some(0.7));
assert_eq!(openai_req.top_p, Some(0.9));
assert_eq!(openai_req.stream, Some(false));

5
crates/hermesllm/src/providers/request.rs
View file

@ -360,6 +360,9 @@ mod tests {
assert_eq!(openai_req.model, openai_req2.model);
assert_eq!(openai_req.messages[0].role, openai_req2.messages[0].role);
assert_eq!(openai_req.messages[0].content.extract_text(), openai_req2.messages[0].content.extract_text());
assert_eq!(openai_req.max_tokens, openai_req2.max_tokens);
// After roundtrip, deprecated max_tokens should be converted to max_completion_tokens
let original_max_tokens = openai_req.max_completion_tokens.or(openai_req.max_tokens);
let roundtrip_max_tokens = openai_req2.max_completion_tokens.or(openai_req2.max_tokens);
assert_eq!(original_max_tokens, roundtrip_max_tokens);
}
}