feat: PPT-568 Added Controller for ChatBot

[naqvis]

No description provided.

[stakach]

I'm thinking we rename openai.cr to chat_gpt.cr
just as i feel it describes the functionality better and matches the class name

Also we should make the API keys configurable per-domain. Similar to what we do in webrtc.cr

      authority = current_authority.not_nil!
      authority.internals["webrtc_ice"]?.try(&.to_json) || WEBRTC_DEFAULT_ICE_CONFIG

just with openai / azure credentials

[stakach]

also seeing this error on our dev server when opening a websocket

level=[I] time=2023-10-17T01:18:17Z program=EventBus source=action-controller client_ip=119.18.0.136 request_id=e16aa21c-b88e-44e2-8b7b-6becac4f5923 user_id=user-C2KyPySJYghf6g event=response method=GET path=/api/engine/v2/chatgpt/chat/sys-ELNF3KjDuK?bearer_token=[FILTERED] status=101 duration=2.5878
level=[D] time=2023-10-17T01:18:17Z program=EventBus source=db message="Executing query" query=SELECT * FROM "user" WHERE "id" = $1 LIMIT 1 args=["user-C2KyPySJYghf6g"]
level=[D] time=2023-10-17T01:18:17Z program=EventBus source=db message="Executing query" query=BEGIN args=[]
level=[D] time=2023-10-17T01:18:17Z program=EventBus source=db message="Executing query" query=ROLLBACK args=[]
Unhandled exception in spawn: PlaceOS::Model::Chat has an invalid field. `summary` is required (PgORM::Error::RecordInvalid)
  from app/lib/pg-orm/src/pg-orm/persistence.cr:210:13 in '__create'
  from app/lib/pg-orm/src/pg-orm/persistence.cr:107:42 in 'save!'
  from app/lib/pg-orm/src/pg-orm/base.cr:19:5 in 'create!:user_id:system_id:summary'
  from app/src/placeos-rest-api/controllers/openai.cr:42:9 in 'chat'
  from app/src/placeos-rest-api/controllers/openai.cr:4:3 in 'ws_chat__system_id'
  from app/src/placeos-rest-api/controllers/openai.cr:4:3 in '->'
  from app/lib/opentelemetry-instrumentation/src/opentelemetry/instrumentation/crystal/http_server.cr:230:19 in 'process'
  from usr/share/crystal/src/http/server.cr:521:5 in 'handle_client'
  from app/lib/opentelemetry-instrumentation/src/opentelemetry/instrumentation/crystal/http_server.cr:86:9 in 'handle_client'
  from usr/share/crystal/src/http/server.cr:451:5 in '->'
  from usr/share/crystal/src/fiber.cr:146:11 in 'run'
  from usr/share/crystal/src/fiber.cr:98:34 in '->'
  from ???

[stakach]

I feel like we should only create a chat if the user actually asks a question
maybe add an on_summary callback that provides a summary of the request and first response once available

[naqvis]

I feel like we should only create a chat if the user actually asks a question maybe add an on_summary callback that provides a summary of the request and first response once available

Yeah, but this was coded per our initial design in which call to web socket was done with initial question and payload. I removed the payload and chat, but kept the initial logic. But I agree it make more sense to create the model only after receiving the very first chat message.

[naqvis]

I'm thinking we rename openai.cr to chat_gpt.cr just as i feel it describes the functionality better and matches the class name

Also we should make the API keys configurable per-domain. Similar to what we do in webrtc.cr

      authority = current_authority.not_nil!
      authority.internals["webrtc_ice"]?.try(&.to_json) || WEBRTC_DEFAULT_ICE_CONFIG

just with openai / azure credentials

Agree and will make these changes

[stakach]

The only thing I think that needs tweaking is this chat resume
It would be nice to have the same description with the capabilities embedded

So we need to make the LLM.new_chat request call for chat resume so the data can be available