This lesson's interactive features are locked, please to keep using them

What Is Search?

Click to play video

Imagine if you had to scroll through every single movie on Netflix to find the movie you're looking for.

Search is a non-negotiable feature of nearly all content-based web applications. In this course, we'll learn how it works, then focus on using AI, LLMs, and search together for an even better user experience. Good search is about much more than exact keyword matches. We want to be able to:

Find relevant results even if the search terms don't match exactly
Match based on semantic meaning, not just keywords
Understand user intent and context
Provide succinct results that go beyond the raw data

To do that, we'll learn about RAG, or "Retrieval Augmented Generation." "Retrieval" is the first component in RAG:

R: Retrieve information via a search algorithm
A: Augment our instructions with the search results
G: Generate better, richer, and more accurate information using LLMs

What Will We Build?

We're going to build pieces of a make-believe product called Hoopla, a Netflix-like streaming service. Specifically, we'll build the technologies that power its search bar.

Assignment

Hoopla needs movie data to search through. Let's download the dataset we'll use throughout this course.

Use uv (installation docs) to create a new project. It will create the directory and initialize a Git repo. In our example code, we've named the project directory rag-search-engine, since it's only one part of the Hoopla product.
```
uv init rag-search-engine
cd rag-search-engine
```
Create a virtual environment at the top level of your project directory:
```
uv venv
```
Make sure the venv directory is in your .gitignore file. uv init usually adds it for you.
Activate the virtual environment:
```
source .venv/bin/activate
```
You should see (rag-search-engine) at the beginning of your terminal prompt; for example, mine is:
```
(rag-search-engine) wagslane@MacBook-Pro-2 rag-search-engine %
```
Always make sure that your virtual environment is activated when running the code or using the Boot.dev CLI in this course.
Create a data directory in your project root.
Download the movie dataset here and save it as data/movies.json.
Add /data/ to your .gitignore file to keep large data files out of version control.

Run and submit the CLI tests.