Unfortunately the project isn't online anymore because my free trial ran out and I couldn't find a way to send the provided credits to my existing subscription
Inspiration
My inspiration for the project was noticing how historical events are described, in sources like textbooks and Wikipedia, differently than in first hand sources. I don't mean to say contemporary writing of historical events is inaccurate, but it has a different feel and style to it. When we describe things in hindsight we tend to narrativize the material differently than people at the time would've. I found a dataset of old newspaper articles and I thought it could be an interesting source to use with vector searching, for more 'authentic' feeling history research.
What it does
It works the same as the tutorial except it searches a dataset of old newspaper articles. This allows the ai to interpret historical questions somewhat more like how they would be perceived by someone at the time. It also can retrieve obscure information that chatGPT hasn't memorized. I only used a small portion of the articles from 1945 because the whole dataset is way too big given the constraints I have.
Challenges I ran into
There were a number of steps in the tutorial that didn't work properly for me, thankfully I was able to resolve most of the issues by reading the Discord channel. I did have to do a lot of troubleshooting with different Azure features, docker container errors, and others.
Accomplishments that I'm proud of
This is my first 'hackathon' submission, my goal was just to submit a legal submission. It ended up being a lot more complicated than I thought it would be, every page of the tutorial had 5+ techniques and systems I've never used before, and as I said a number of steps didn't work right away. I'm satisfied I was able to finish it and I learned quite a lot of new things.
What I learned
what RAG is, using Azure platform, using docker, using langchain, using mongoDB, using environment variables, using json, more advanced compartmentalization concepts, python documenting and style concepts, what a container is, basics of how apis work, basics of modifying shell scripts, basics of debugging professional libraries, basics of command prompt, basics of powershell, deploying website with github,
Log in or sign up for Devpost to join the conversation.