- go to labs\03-LAB-03\1-azure-deployment
- modify azuredeploy.parameters.json to give resources a good name. e.g. arg-syd-aiapp1day
- modify deploy-core.ps1
- sub for demo sub: 9df3a442-42f1-40dd-8547-958c3e01597a
- rg is in this format, change suffix if needed. replace all placeholders : arg-syd-aiapp1day-anko
- run the commands in deploy-core.ps1 one by one to create azure resources
- AzureOpenAI apikey: the apikey is same for all models. look up the old apikey in the document, search and replace all in the vs code.
- mongodb connection string: only need to replace the host name part. search and replace all in the vs code.
- check the .env file and update with new keys and urls.
- Playground app
apps\playground: no config change needed. use vs code to publish the site. - API app
apps\api: replace .env values (follow .env.demo). use vs code to publish the site. - Chatbot app
apps\chatbot: replace .env values, point to API app url. use vs code to publish the site.
- playground: enter apikey and chat is working. https://arg-syd-aiapp1day-playground.azurewebsites.net
- api: check swagger endpoint is working. https://arg-syd-aiapp1day-api.azurewebsites.net/docs
- chatbot: check chat is working. https://arg-syd-aiapp1day-chat.azurewebsites.net
- if any of them not working, check the deployment log tab in the azure portal.
- the first time deployment could take long for orxy to restore all the packages
- use db name as aiapp1day_demo
- go through the setup steps
- follow below steps to run completed scripts, basically follow lab2 part 2 & 3
- once done, the chatbot should be able to answer sku questions
- labs\03-LAB-03\2-Load-Data\completed> npm install
- labs\03-LAB-03\2-Load-Data\completed> node .\convert.js
- labs\03-LAB-03\2-Load-Data\completed> node .\import.js
- labs\03-LAB-03\2-Load-Data\completed> npm install
- labs\03-LAB-03\3-Vector-Search\completed> node .\vectorize.js