engineering/opencompletion.com
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Improve code auto-fix: prefer Qwen Coder with Hermes fallback

Browse source 
Update /api/fix-code endpoint to try MODEL_3 (Qwen Coder) first for
better code generation, with graceful fallback to MODEL_1 (Hermes) if
MODEL_3 is unavailable or returns errors.

This ensures the auto-fix feature works reliably even when specialized
code models are temporarily unavailable (e.g., 502 errors).
This commit is contained in:
Russell Ballestrini 2025-11-30 12:04:51 -05:00
parent 6400a5eacb
commit 31fc7946a0
1 changed files with 8 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

10
app.py
View file

@ -868,8 +868,14 @@ def fix_code():
if not code or not language:
return jsonify({"error": "Code and language are required"}), 400
# Use MODEL_1 (Hermes) to fix the code
client, model = get_openai_client_and_model("MODEL_1")
# Try MODEL_3 (Qwen Coder) first for better code fixing, fall back to MODEL_1 (Hermes)
try:
client, model = get_openai_client_and_model("MODEL_3")
print(f"[INFO] Using MODEL_3 for code fixing: {model}")
except Exception as e:
print(f"[WARN] MODEL_3 not available, falling back to MODEL_1: {e}")
client, model = get_openai_client_and_model("MODEL_1")
print(f"[INFO] Using MODEL_1 for code fixing: {model}")
system_prompt = f"""You are an expert {language} programmer and debugger. Your task is to fix code that has errors.