Handle Open Assistant error event_type

[tanchekwei]

Handle below error response:

data: {"event_type": "error", "error": "Request failed during generation: Server error: transport error", "message": {"id": "123", "parent_id": "123", "content": null, "chat_id": "123", "created_at": "2023-07-17T03:44:08.659000", "role": "assistant", "state": "aborted_by_worker", "score": 0, "reports": [], "work_parameters": {"model_config": {"model_id": "OpenAssistant/oasst-sft-6-llama-30b", "max_input_length": 1024, "max_total_length": 1792, "quantized": false}, "sampling_parameters": {"top_k": 50, "top_p": 0.95, "typical_p": null, "temperature": 0.75, "repetition_penalty": 1.2, "max_new_tokens": 1024}, "do_sample": true, "seed": 123, "system_prompt": null, "plugins": [], "plugin_max_depth": 4}, "safe_content": null, "safety_level": null, "safety_label": null, "safety_rots": null, "used_plugin": null}}

And add try-catch block

And update done to true after stream closed.

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
chatall	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Jul 17, 2023 11:45am