johnregan3 / tars Goto Github PK

https://www.youtube.com/watch?v=8z8Cobsvc9k

Reference: https://www.youtube.com/watch?v=QGLF3UbDf7g

Provide TARS with a way to search the internet for answers and more recent information than is currently available to OpenAI's models. For instance, the ability for TARS to consider the upcoming weather forecast when recommending plans would be really helpful (and frankly, pretty darn cool).

A Knowledge Base (KB) is a reference guide for TARS to store important information, like names of family members, important dates, etc.

Issue: http://localhost:5173 has no app

This is a great project, may I ask where did I go wrong? Everything went smoothly, but the application was not seen on the webpage in the end:

Env:

Macbook Air M1
postgres (PostgreSQL) 15.2

✅Clone this repo.
✅Run composer install
✅Run npm install
✅php artisan migrate:fresh --seed
✅npm run build
✅npm run dev

.env:

OPENAI_API_KEY=sk-
OPENAI_ORGANIZATION=
CHAT_USER_NAME=Cooper
CHAT_USER_ID=1
CHAT_TARS_NAME=TARS
CHAT_TARS_ID=2

DB_CONNECTION=pgsql
DB_HOST=mysql
DB_PORT=5432
DB_DATABASE=chat_database
DB_USERNAME=chat_user
DB_PASSWORD=password

CACHE_DRIVER=redis
APP_URL=http://localhost:8080

Terminal:

VITE v4.1.4 ready in 269 ms

➜ Local: http://localhost:5173/
➜ Network: use --host to expose
➜ press h to show help

LARAVEL v9.52.4 plugin v0.7.4

➜ APP_URL: http://localhost:8080

Graph database or vector database can help semantic operations, like search or vertical navigation on the knowledge graph.

Maybe pinecone.io

Check Shapiro's Chatbot with infinite memory using OpenAI and Pinecone video.

I like having the web UI, as well as the familiar toolset of JS/PHP for most tasks (for instance, routing and authentication, etc), but I should investigate if there are speed benefits to processing most of the Controller work via Python, and how practical that implementation would be.