Major Refactor of Research Assistant Web App

[victordibia]

Why are these changes needed?

• Organize all backend logic into a library structure named autogenra . This will help support future extension (e.g., logic for a playground)
• extract all db functions into a db manager class
• extract all autogen functionality into an autogenchat manager class
• triage setup of autogen agents via a chatmanager class that constructs agents using flow class. The flow class takes in a set of declarative specifications and returns autogen agents. This is important as these specifications can be mapped in the future to some UI where agents and agent flows are specified and can be run easily.
• extract all data models in to a a clean datamodel file reused across the entire app stack.

Other Changes

• The need for explicit @execute is removed as this was an artifact of the earlier demo.
• personalization is not included at the moment (can be added in a future update)
• teachability is not added at the moment (can be added in a future update)

@pcdeadeasy @gagb

Related issue number

Checks

• I've included any doc changes needed for https://microsoft.github.io/autogen/. See https://microsoft.github.io/autogen/docs/Contribute#documentation to build and test documentation locally.
• I've added tests (if relevant) corresponding to the changes introduced in this PR.
• I've made sure all auto checks have passed.

[afourney]

Alright. I got it working, and it looks great. A few comments (mostly already addressed in other channels).

1. I've got the latest Ubuntu LTS, and it installs a pretty old version of Node. So, I got this error:

Gatsby requires Node.js 14.15.0 or higher (you have v12.22.9).
Upgrade Node to the latest stable release:
https://gatsby.dev/upgrading-node-js

I was able to resolve this by upgrading Node:

sudo npm cache clean -f
sudo npm install -g n
sudo n stable

2. The version I tried was missing frontend/.env.production. Adding that, and doing a fresh npm run build solved a problem with undefined paths.

3. The UI does not read from OAI_CONFIG_LIST. Instead it looks to OPENAI_API_KEY. This probably needs to be corrected before release, since the CONFIG_LIST offers all the nifty options we depend on.

4. When I typed my first message "Hello", the agents got in an infinite loop:

user_proxy (to primary_assistant):

Hello

--------------------------------------------------------------------------------
primary_assistant (to user_proxy):

Hello! How can I assist you today?

--------------------------------------------------------------------------------
user_proxy (to primary_assistant):

As an AI, I don't require assistance. Do you have any questions or is there something you need help with?

--------------------------------------------------------------------------------
primary_assistant (to user_proxy):

As an AI, I'm here to assist you. How can I help you today?

--------------------------------------------------------------------------------
user_proxy (to primary_assistant):

As an AI, I don't require assistance. Do you have any questions or is there something you need help with?

--------------------------------------------------------------------------------
primary_assistant (to user_proxy):

As an AI, I'm here to assist you. How can I help you today?

--------------------------------------------------------------------------------
user_proxy (to primary_assistant):

As an AI, I don't require assistance. Do you have any questions or is there something you need help with?

--------------------------------------------------------------------------------
primary_assistant (to user_proxy):

As an AI, I'm here to assist you. How can I help you today?

--------------------------------------------------------------------------------
user_proxy (to primary_assistant):

As an AI, I don't require assistance. Do you have any questions or is there something you need help with?

--------------------------------------------------------------------------------
primary_assistant (to user_proxy):

As an AI, I'm here to assist you. How can I help you today?

--------------------------------------------------------------------------------
user_proxy (to primary_assistant):

As an AI, I don't require assistance. Do you have any questions or is there something you need help with?

--------------------------------------------------------------------------------
primary_assistant (to user_proxy):

As an AI, I'm here to assist you. How can I help you today?

--------------------------------------------------------------------------------
user_proxy (to primary_assistant):

As an AI, I don't require assistance. Do you have any questions or is there something you need help with?

--------------------------------------------------------------------------------
primary_assistant (to user_proxy):

As an AI, I'm here to assist you. How can I help you today?

--------------------------------------------------------------------------------
user_proxy (to primary_assistant):

As an AI, I don't require assistance. Do you have any questions or is there something you need help with?

--------------------------------------------------------------------------------
primary_assistant (to user_proxy):

As an AI, I'm here to assist you. How can I help you today?

--------------------------------------------------------------------------------
user_proxy (to primary_assistant):

As an AI, I don't require assistance. Do you have any questions or is there something you need help with?

--------------------------------------------------------------------------------
primary_assistant (to user_proxy):

As an AI, I'm here to assist you. How can I help you today?

--------------------------------------------------------------------------------
user_proxy (to primary_assistant):

As an AI, I don't require assistance. Do you have any questions or is there something you need help with?

--------------------------------------------------------------------------------
primary_assistant (to user_proxy):

As an AI, I'm here to assist you. How can I help you today?

--------------------------------------------------------------------------------
user_proxy (to primary_assistant):

As an AI, I don't require assistance. Do you have any questions or is there something you need help with?

We probably need to keep the user_proxy to just executing code, and TERMINATING when there's nothing to do. Infinite loops shouldn't happen (but arise in many circumstances with Autogen, unless configured just right)

[gagb]

(Minor separate PR is okay) Does this require sudo? If so, should we warn the users? (Eg, by adding a comment in this script)