clrs-practice-exam/README.md

# CLRS practice exam generator

Generates practice exams by randomly selecting questions and their answers from the CLRS textbook

## Why?

[Interleaved practice](https://en.wikipedia.org/wiki/Varied_practice) is when you mix practicing a bunch of subjects together. Textbooks are designed such that all the questions relevant to a particular topic are grouped together. While this is good for reference, it's not all that good for practice.

## How?

[Peng-Yu Chen](https://pengyuc.com/) and [contributors to the walkccc/CLRS](https://github.com/walkccc/CLRS/graphs/contributors) have collected solutions to many (but not all!) of the problems in the third edition of CLRS. They have also conviently and seemingly accidentally structured them in a way that lends itself well to being parsed into a database. I've written a script that automatically sorts all of these problems into an SQLite database, and another you can use to query it and generate nice-looking HTML practice exams.

This is probably best described as something that barely works. As in, it's something I duct taped together while bored in class and bothered to do very little error-handling.

Not to mention, `walkccc/CLRS` is incomplete as far as data sets go. It's the largest and most well-formated that I'm aware of, anyways.

### How can I generate new practice exams?

This setup assumes you have Python 3, Git, and the `cmarkgfm` Python package. Before you get started, make sure you have the `cmarkgfm` package installed. On Debian:

```bash
$ sudo apt install python3-cmarkgfm
```

Or with pip, call:
```bash
$ pip install cmarkgfm
```

Next, you can run this script if you're on some kind of Unix machine. Something similar on Windows but idk how that stuff works.

```bash
# 1. Make yourself a workspace
mkdir work
cd work

# 2. Get the problems
git clone https://github.com/walkccc/CLRS

# 3. Get the scripts
git clone https://git.nats.solutions/nat/clrs-practice-exam
cd clrs-practice-exam

# 4. Clean the data
python clean-data.py ../CLRS/docs

# 5. Generate the practice exam
python practice-exam-generator.py "select * from problem where chapter in (3,4,7,12,15,16,22,24,26,34) order by random() limit 20" --template template.html
```

### Specifics

`practice-exam-generator.py` takes a positional argument that's an SQL query. The query in the example above is the one I used to generate the example exams, and it's kind of but not super reflective of the stuff I've covered in my algorithms course.

The relevant table in the database is made like this:

```sql
create table if not exists problem (
    problem_number number not null,
    question text not null,
    answer text not null,
    chapter number not null,
    section number,
    starred number not null
)
```

So, you can use all of this information to construct your SQL query.

Both scripts take an `-h` argument that breaks down what you can pass to it to tweak things based on your setup.

The `template.html` document should probably include [KaTeX](https://katex.org/) lest the many, many LaTeX equations be broken. This template must include the string "%$%content" somewhere in it; it'll be replaced with all the selected problems when the exam is being generated

## "License"

This software (that is, the scripts; not including anything you'll find in `./examples`) is a gift from me to you. You form a relationship with all those from whom you accept gifts, and with this relationship comes certain expectations. Namely:

* When you share this gift with others, you will share it in the same spirit as I share it with you.
* You will not use this gift to hurt people, any living creatures, or the planet

The questions were written by Michelle Bodnar and Andrew Lohr and used here on a fair-use basis. The answers are organized in the [walkccc/CLRS](https://github.com/walkccc/CLRS) repository, by [Peng-Yu Chen](https://github.com/walkccc) and [contributors](https://github.com/walkccc/CLRS/graphs/contributors). Their repository is shared under the MIT license.