Issue with maintaining conversation_id for continuing the same chat session in OpenaiChat

[kimkibom]

I'm using the g4f.ChatCompletion.create method to interact with OpenaiChat with model gpt-4o in a conversational context. However, every time I run the script, a new conversation is created instead of continuing with the same conversation_id.

Here's the code I'm using:

response = g4f.ChatCompletion.create(
    model="gpt-4o",
    provider=OpenaiChat,
    messages=messages,
    history_disabled=False,
    conversation_id="your-conversation-id"
)

I want to continue the conversation with the same conversation_id, but the debug logs show a new conversation being created each time. For example, the log shows:

OpenaiChat: New conversation: 67f40fb7-ea08-8005-ae7e-7XXXXXXXXXX1

This happens every time I run the Python script, even though I pass the same conversation_id. I'm unsure whether I missed something or if there’s a specific way I need to handle session persistence or conversation_id reuse to keep the conversation context intact.

Could someone point out what I might be missing or doing wrong?

[hlohaus]

Please include the conversation ID in the conversation field. Enable return_conversation=True and include it in the "conversation" field of your subsequent request. A parent message ID is also required. @kimkibom

[kimkibom]

Based on your answer, here is the format that you expected, right?

response = g4f.ChatCompletion.create(
    model="gpt-4o",
    provider=OpenaiChat,
    messages=messages,
    return_conversation=True,
    conversation_id="your-conversation-id",
    parent_message_id="your-parent-message-id"
)

However, is there an alternative way to keep the conversation context intact without manually specifying the conversation_id and parent_message_id every time? Or is this a limitation that requires specific configuration or setup?"

[hlohaus]

@kimkibom the API do it automatically, when you pass everytime a self generated conversation_id.

2. The internal conversation should be a JsonConversation object, when you are using python.