AO3 Backup-to-Website

What is this?

This is really two-and-a-half projects glued together. The first project scrapes the AO3 works for a single user and generates markdown files for backup and storage. The second project builds a static site based on those markdown files for hosting those works.

Uses Scrapy for scraping AO3 and Hugo for generating a static site.

Scrapes AO3 and deploys to Github Pages using actions.

How can I use this?

This is not currently designed to be usable for anyone not comfortable with the command line and modifying configuration files.

Running locally

This is for if you don't want to futz around with getting GitHub actions to work or using GitHub in particular.

Scraping AO3

1. Run git clone --recursive https://github.com/thedeadparrot/ao3backup.git
2. Set up a python virtualenv.
3. Run pip install -r ao3scrape/requirements.txt
4. Modify ao3scrape/settings.py to set AO3_USER to your own AO3 username.
5. Run git rm backup/content/posts/* in ordert to delete all of my fic and clear the way for your own.
6. From within the ao3scrape subdirectory, run scrapy crawl ao3. This will create the markdown files backup/content/posts.

Building the static site

1. Install Hugo.
2. Modify backup/config.toml with your own name and links and other configuration you might be interested in.
3. Run hugo from within the backup directory.
4. Your static site should now be in backup/public, which can then be hosted wherever you wish.

Running on GitHub

These instructions assume a familiarity with both GitHub and git.

1. Fork the repo.
2. Delete the markdown files in backup/content/posts.
3. Modify the settings in backup/config.toml.
4. Set up your fork to use GitHub Pages and the source 'GitHub Actions'.
5. Set up your fork to use GitHub Actions and 'all actions and reusable workflows'.
6. From your fork, run the action Scrape AO3 and attempt to commit, inserting the username when prompted.
7. Once that action has run (mine takes about 30 minutes), check your pull requests for the new markdown files and merge if they look good.
8. Run Deploy Hugo site to Pages to build and deploy your site.

Limitations

Since I built this for me, I have not addressed every possible use case. There are a few that I know for a fact we don't handle.

• This doesn't scrape anything that is private or ao3-locked. It is welcomed and encouraged to make your own markdown files for those works if you would like to archive them.
• Series ordering for the most part doesn't work. This is mostly the fault of Hugo.
• Also the fault of Hugo is that it doesn't handle tags/fandoms/pairings that end in a period. I chop them off the ends, so that's why your tags might look weird.
• Since this doesn't have any underlying understanding of synonymous tags, some of the tag wrangling magic of AO3 is gone. How it shows up on your work is how it shows up in your generated site.
• There is no single-chapter view for works.
• I'm sure I'm missing a bunch of other caveats. Sorry.