intelligentnode / intellichat Goto Github PK

[✓] Add one key to the settings.
[✓] Get the one key from the query and enable the settings if sent.
[✓] Add the new cohere model to the list.
[✓] Update chat API to pass in the one_key

• update intellinode.
• add Anthropic as provider.

• Create the chatbot API.
• Add settings to choose the model provider.
• Connect the UI to the API.

• By default the login should be turned off.
• Update the readme with enablement instructions.
• If the login on, the user cannot send messages.

Using ChatContext when selecting replicate as the model throws an Error; the RemoteEmbedModel only accepts openai and cohere as the SupportedEmbedModels for providers. Should we always pass OpenAI as the provider to ChatContext regardless to the selected model?

Create a folder "intellinext" and install the modules.

Azure Openai provides an extra layer of privacy, this is supported in the intelliNode chatbot and we want to integrate this with the chat UI.

The UI impact:

• Add option in the models dropbox called: Azure Opneai
• Show 4 text boxes when this option is selected:
  [ ] API key.
  [ ] Resource name.
  [ ] Chatbot model name.
  [ ] Embedding model name.

Azure Integration Sample:
https://github.com/intelligentnode/IntelliNode/blob/main/samples/command_sample/test_azure_chatbot.js

Note that the default chatbot mode should be the regular openai, this is added as another option similar to replicate.

• Set max number (n) of message sent in a request
• Make the max number of messages configurable from the chat settings UI

• update the chat mutationFn to include the maximum number of messages in the request payload
• update the chat route to use the new ChatContext function
• update the chat route to pass the ChatContext function response to the addMessages util function instead of the payload's messages

Add a show case video for the multi providers functionality.

Allow the user to select the language model from the chat settings UI based on the selected model provider.

Intellinode provides a chatbot with Sagemaker llama backend as described in this wiki.

The required work to update the UI with Sagemaker is as follows:

• Add Sagemaker option in the dropbox.
• Display a drop-down list with one selection: llama.
• Display a textbox with url title.

The backend:
When they send a message with SageMaker selection, call the relevant module from intellinode.

• Configure eslint and prettier.
• Configure tailwind.
• install intellinode and basic type definitions for chat.
• Configure shadcn and radixui.
• Write a minimal Github CI file for typescript type checking and linting.

Create an initial UI skeleton for the blog assistant page and components

The link to generate the start button:
https://buttons.github.io/

Create an initial UI skeleton for the chatbot page and components

• Persist the chat settings' state in localStorage.
• Add tooltips where needed.
• Add conditional rendering depending on the selected chat provider:
• Display the relevant API key field based on the selected provider,
• Display an option to select whether to use chat context (embeddings) or not when Replicate is selected,
• Display the OpenAI API key field when both Replicate provider and chat context are selected,
• Chat context is enabled by default when using OpenAI.

• Update the chat settings UI to include the APIKeys fields.
• Update the request to include the keys in the request body.
• Update the chat API route to accept an option APIKeys property and use those if provided instead of the keys defined in environment variables.

Validate all model selection works and the keys sent as expected.