Advent of Code DEV Megathread Language Bot

In 2019, on the DEV megathread for the Advent of Code challenge, the thread author kept a tally of how many people had completed the previous day's challenge in each language.

I'm doing the same thing now, but updating them is quickly becoming more time consuming than planned. It's time for a bot!

The Idea

For each article in my DEV AoC megathread, this bot should go get the comments, and, ideally, output the markdown table that I can copy paste into the article.

There are some comments that don't have code blocks classified by language. And some people comment just that they solved in particular language, maybe with an image of their code. In order to not duplicate work, the bot should run each comment through a checker that checks for any possible language words that might signify that there's more to the prose of the article about how the person solved things. It should show me the text of the comment and ask me to classify the language.

Then it should save those decisions so it doesn't have to ask me again.

Installation

The short version is clone the repo, create a virtual environment, install the dependencies, and you're good to go.

git clone https://github.com/rpalo/aoc_language_bot.git
cd aoc_language_bot.git
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Usage

Currently two subcommands are supported: fetching and showing.

Fetching

Fetching downloads the articles and comments and builds a database via a semi-user-interactive mode. Any comments it's not sure about, it'll ask you what you think. Answer with the language you think it is, and if it's a comment that doesn't have a submission in it, leave your answer blank.

python3 aoc_language_bot/aoc_language_bot.py fetch

By default, it expects a list of post URLs in the same directory you call it from called urls.txt. Each URL on its own line with no other characters. You can customize this file path with the -u flag.

python3 aoc_language_bot/aoc_language_bot.py -u data/my_urls.txt fetch

Similarly, it expects and subsequently force-creates a database file, which is, by default called aoc_language_bot_history.json. You can customize this path with the -H/--history_file option.

python3 aoc_language_bot/aoc_language_bot.py -H data/db.json fetch

If the file doesn't exist, it will be created.

Note that both of these flags have to come before the fetch subcommand.

Also, you can use the -d/--delete flag to blow the existing database away without having to hunt down and delete it yourself.

python3 aoc_language_bot/aoc_language_bot.py -H data/db.json fetch --delete

Showing

The bot outputs the markdown into your terminal so you can copy/paste it right into your blog post.

python3 aoc_language_bot/aoc_language_bot.py show 1

You must specify which day you wish to choose. The order of the URLs in the URL file specifies the days (i.e. the first URL is for Day 1).

You can optionally specify an "alias" file if the automated finder needs correcting a lot. By default, it expects a file called aliases.json in the same directory you ran the command, but you can customize that with the -a/--alias_file option.

python3 aoc_language_bot/aoc_language_bot.py show 1 -a data/nicks.json

Aliases are simple key-value pairs of common misspellings to how they should properly look. Here's an example:

{
  "Javascript": "JavaScript",
  "Php": "PHP",
  "Csharp": "C#"
}

Contributing

I welcome any contributions. This was just a fun little project to make my life a little easier, but I think it turned out good. Feel free to fork it up, open a pull request, or open an issue to talk about improving things.

Code of Conduct

This project has a Code of Conduct. Please read it before contributing or opening an issue.

Development Notes

If someone makes two different language submissions in the same comment, it will currently not be detected. The auto finder should be extended to compare all of the code blocks in the comment and do a hashset of all of the languages included.