apunkt/plano
1
0
Fork 0
mirror of https://github.com/katanemo/plano.git synced 2026-06-17 15:25:17 +02:00
Code Issues Projects Releases Packages Wiki Activity

fix tracing

Browse source
This commit is contained in:
Adil Hafeez 2025-01-14 11:56:10 -08:00
parent 35065e2e41
commit 9570b167db
2 changed files with 13 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

4
crates/common/src/tokenizer.rs
View file

@ -1,4 +1,4 @@
use log::debug;
use log::trace;
#[derive(thiserror::Error, Debug, PartialEq, Eq)]
#[allow(dead_code)]
@ -9,7 +9,7 @@ pub enum Error {
#[allow(dead_code)]
pub fn token_count(model_name: &str, text: &str) -> Result<usize, Error> {
debug!("getting token count model={}", model_name);
trace!("getting token count model={}", model_name);
// Consideration: is it more expensive to instantiate the BPE object every time, or to contend the singleton?
let bpe = tiktoken_rs::get_bpe_from_model(model_name).map_err(|_| Error::UnknownModel {
model_name: model_name.to_string(),

17
crates/llm_gateway/src/stream_context.rs
View file

@ -312,9 +312,11 @@ impl HttpContext for StreamContext {
}
fn on_http_response_body(&mut self, body_size: usize, end_of_stream: bool) -> Action {
debug!(
trace!(
"on_http_response_body [S={}] bytes={} end_stream={}",
self.context_id, body_size, end_of_stream
self.context_id,
body_size,
end_of_stream
);
if !self.is_chat_completions_request {
@ -398,9 +400,10 @@ impl HttpContext for StreamContext {
let body = if self.streaming_response {
let chunk_start = 0;
let chunk_size = body_size;
debug!(
trace!(
"streaming response reading, {}..{}",
chunk_start, chunk_size
chunk_start,
chunk_size
);
let streaming_chunk = match self.get_http_response_body(0, chunk_size) {
Some(chunk) => chunk,
@ -520,9 +523,11 @@ impl HttpContext for StreamContext {
}
}
debug!(
trace!(
"recv [S={}] total_tokens={} end_stream={}",
self.context_id, self.response_tokens, end_of_stream
self.context_id,
self.response_tokens,
end_of_stream
);
Action::Continue