Optimization

One of the first problem i ran is how convert the obsidian's link format into a working system for Next.

I came to the conclusion that i need to parse every markdown files loaded to get the name, path and other infos. From what i saw, some others alternatives came to the same conclusion, but how to optimize this to not have to parse at each page load ?

The first solution from Mindstone is to execute a function that write a jsonfile with all the data needed, the file is generated on first fetch and used after that. This is probably one of the best and easiest solution, but i wanted to try something else. Some hosting method don't really support read/write files between builds, or that are not created on build time.

So i decided to use the next's api routes and caching methods manage the data-fetching, it's for me the most reliable method. And for further extend, i can imagine a day where the data will be fetcher from a database or a server, it open a lot of possibilities.

In reality, i din't test if it's really that effective, i don't have so much data to note any difference, and with all the build output options, maybe this is not the best way to do, i'll see that later.

So in short, i have 2 main routes :

• /api/getData - parse all the files and return the array of the files with the needed datas
  • This is the one which has to be cached
• /api/getFile - with the slug in parameter from the next [...slug] params on the body, fetch the /api/getData and return the specified file

The object representing the file :

{
  "id": "id-path-to-the-file",
  "name": "Name of the file without .md",
  "aliases": ["array", "of", "aliases"],
  "data": {}, // the data from the file
  "content": "the markdown content of the file",
  "path": "/path/to/the/file"
}