Skip to content

Commit 29f9780

Browse files
committed
Remove chunk metrics calculation for reasoning content
- deepseek sends reasoning in such way: [Pasted text], the issue with `apps/editor/src/utils/send-llm-message.ts` is that when run for deepseek, it calls juggles between on_chunk and on_thinking_chunk constantly causing UI jittering
1 parent decbdb9 commit 29f9780

1 file changed

Lines changed: 0 additions & 2 deletions

File tree

‎apps/editor/src/utils/send-llm-message.ts‎

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -308,7 +308,6 @@ export const send_llm_message = async (params: {
308308
if (new_reasoning_content) {
309309
api_thoughts += new_reasoning_content
310310
handle_thinking_chunk(new_reasoning_content)
311-
handle_chunk_metrics(new_reasoning_content)
312311
}
313312

314313
process_content(new_content)
@@ -344,7 +343,6 @@ export const send_llm_message = async (params: {
344343
if (typeof reasoning_content == 'string') {
345344
api_thoughts += reasoning_content
346345
handle_thinking_chunk(reasoning_content)
347-
handle_chunk_metrics(reasoning_content)
348346
}
349347
}
350348
}

0 commit comments

Comments
 (0)