kevin.garden is a site sown from files, directories, and .DS_Store.
kevin.garden is also the code that generates itself.
- Images (jpeg, png, webp, gif, apng, svg, bmp, ico), videos (mp4, webm), and audio (mp3, wav, ogg, m4a) are displayed as their html equivalents: <img>, <video>, <audio>.
- Markdown (md) files are parsed into their relevant html.
- A file with no extension has its name and contents displayed. An empty extensionless file makes for a great visual header.
- All other file types are direct links with no formatting.
- Respects .gitignore and .gardenignore denylists.
- Parses from the .DS_Store file coordinates, directory view options, and background colors.
Please read this entire section if you're interested in forking kevin.garden.
how do we turn a directory into a garden?
usage: node cultivate.js DIR
options: -h, --help print helpSince I have my code and my content in the same place, everywhere here, it's a bit messy if you want to fork this repo. To start fresh, all you need is to copy package.json, src/ and views/. I didn't want to hide src/ and views/ on kevin.garden, so those folders will have residual index.htmls inside them; pardon the dust.
In the views/ directory are two ejs templates, natural.ejs, and formal.ejs. The script selects the natural template when "Sort By > None" is selected for the directory AND every file to be displayed has been moved at least once. Otherwise, it chooses the formal template.
By default, npm run cultivate runs node src/cultivate.js . (on the current directory). Beware where you run this! The maximum recursion depth is set to 3 but even that can easily wreak havoc in the wrong directory.
Before you go off cultivating everything, be sure to create a .gardenignore file with .* so that hidden files and directories (such as .git/) are not visited. My .gardenignore also includes:
*.html
CNAME
README.md
package-lock.json
package.json
And please make sure node_modules/ is ignored somewhere (cultivate.js also reads .gitignore files, so there is great), otherwise index.htmls will find their way into many, many unwanted cracks (folders). Unless you want that. An important note is that the script only reads .gardenignore and .gitignore once, at the root directory it was invoked in. It will not listen to additional .gitignores within directories.
I have found that the best way to have the .DS_Store file update, and thus the cultivate.js script to know where you have moved your files, is to first make all your updates and movements, and then delete the index.html file before re-running the script. Deleting a file in a directory seems to trigger an update to .DS_Store which in turn updates the location data of its contents.
Also important to know is that the "Sort By" and background color properties of the current directory is stored in the .DS_Store file of the previous directory. This is not an issue when visiting subsequent directories, but does mean that the script, in the beginning, will try to read the .DS_Store file in the directory before the root directory. This happens in cultivateHelper().
cultivate.js uses Node's spawnSync() to run the .DS_Store parser, an amazing Python script by hanwenzhu. Make sure to have python 3 installed. If python3 is not your python 3 command be sure to change it in parseDS_Store().
- I have plans to add max resursion depth as a command line argument. I have not done that yet. For now, manually set it, or pass it in in
cultivateHelper().