APEX/BrowserUse_and_ComputerUse_skills
5
0
Fork 0
Code Issues 4 Pull requests 1 Projects 1 Releases Packages Wiki Activity Actions

Merge PR #403: Fix context overrun crash with local LLM backends

Browse source 
Authored by ch3ronsa. Fixes #348.

Adds 'context size' (LM Studio) and 'context window' (Ollama) to
context-length error detection phrases so local backend 400 errors
trigger compression instead of aborting. Also removes 'error code: 400'
from the non-retryable error list as defense in depth.
This commit is contained in:
teknium1 2026-03-04 17:48:44 -08:00
commit 3220bb8aaa
1 changed files with 11 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

22
run_agent.py
View file

@ -3333,23 +3333,24 @@ class AIAgent:
} }
# Check for context-length errors BEFORE generic 4xx handler. # Check for context-length errors BEFORE generic 4xx handler.
# OpenRouter returns 400 (not 413) for "maximum context length" # Local backends (LM Studio, Ollama, llama.cpp) often return
# errors — if we let the generic 4xx handler catch those first, # HTTP 400 with messages like "Context size has been exceeded"
# it aborts immediately instead of attempting compression+retry. # which must trigger compression, not an immediate abort.
is_context_length_error = any(phrase in error_msg for phrase in [ is_context_length_error = any(phrase in error_msg for phrase in [
'context length', 'maximum context', 'token limit', 'context length', 'context size', 'maximum context',
'too many tokens', 'reduce the length', 'exceeds the limit', 'token limit', 'too many tokens', 'reduce the length',
'exceeds the limit', 'context window',
'request entity too large', # OpenRouter/Nous 413 safety net 'request entity too large', # OpenRouter/Nous 413 safety net
]) ])
if is_context_length_error: if is_context_length_error:
print(f"{self.log_prefix}⚠️ Context length exceeded - attempting compression...") print(f"{self.log_prefix}⚠️ Context length exceeded - attempting compression...")
original_len = len(messages) original_len = len(messages)
messages, active_system_prompt = self._compress_context( messages, active_system_prompt = self._compress_context(
messages, system_message, approx_tokens=approx_tokens messages, system_message, approx_tokens=approx_tokens
) )
if len(messages) < original_len: if len(messages) < original_len:
print(f"{self.log_prefix} 🗜️ Compressed {original_len} → {len(messages)} messages, retrying...") print(f"{self.log_prefix} 🗜️ Compressed {original_len} → {len(messages)} messages, retrying...")
continue # Retry with compressed messages continue # Retry with compressed messages
@ -3370,11 +3371,10 @@ class AIAgent:
# Check for non-retryable client errors (4xx HTTP status codes). # Check for non-retryable client errors (4xx HTTP status codes).
# These indicate a problem with the request itself (bad model ID, # These indicate a problem with the request itself (bad model ID,
# invalid API key, forbidden, etc.) and will never succeed on retry. # invalid API key, forbidden, etc.) and will never succeed on retry.
# Note: 413 and context-length errors are excluded — handled above # Note: 413 and context-length errors are excluded — handled above.
# via compression.
is_client_status_error = isinstance(status_code, int) and 400 <= status_code < 500 and status_code != 413 is_client_status_error = isinstance(status_code, int) and 400 <= status_code < 500 and status_code != 413
is_client_error = (is_client_status_error or any(phrase in error_msg for phrase in [ is_client_error = (is_client_status_error or any(phrase in error_msg for phrase in [
'error code: 400', 'error code: 401', 'error code: 403', 'error code: 401', 'error code: 403',
'error code: 404', 'error code: 422', 'error code: 404', 'error code: 422',
'is not a valid model', 'invalid model', 'model not found', 'is not a valid model', 'invalid model', 'model not found',
'invalid api key', 'invalid_api_key', 'authentication', 'invalid api key', 'invalid_api_key', 'authentication',
@ -3397,7 +3397,7 @@ class AIAgent:
"failed": True, "failed": True,
"error": str(api_error), "error": str(api_error),
} }
if retry_count >= max_retries: if retry_count >= max_retries:
print(f"{self.log_prefix}❌ Max retries ({max_retries}) exceeded. Giving up.") print(f"{self.log_prefix}❌ Max retries ({max_retries}) exceeded. Giving up.")
logging.error(f"{self.log_prefix}API call failed after {max_retries} retries. Last error: {api_error}") logging.error(f"{self.log_prefix}API call failed after {max_retries} retries. Last error: {api_error}")