.search_index.gz

Congratulations! You have found the search index. One of the filesystem's dirty little secrets.

I needed a place to store an index of each user's files so I could make them searchable. Now, there are search databases like ElasticSearch and such, but that's a lot of work to set up and maintain.. Instead I opted to simply put the full path of every file in your filesystem in a text file. That's what you're looking at here. That can add up to a lot of data, but since the paths usually have a lot of repetitive elements it compresses incredibly well. You'd be hard-pressed to grow this index over even 1 MB. Honestly, this search system is incredibly efficient, I'd be surprised if EleasticSearch could even match it.

This file is updated 10 minutes after the last time you modify a file on your filesystem. So if you're constantly uploading and deleting files your search index might never update and you will be left with stale search results.

If you delete this file then search will stop working until the file is regenerated 10 minutes later. If you replace this file with a different file then that file will be overwritten 10 minutes later. And if you replace this file with a directory with the same name then search will stop working completely until you delete the directory (yes, I tested this case).

Each time you type a search term in the search dialog this file gets decompressed and searched on the fly. There is no trickery here, the file simply gets read line by line. Modern CPUs are incredibly good at searching for text. I benchmarked it once, I don't remember the exact numbers but it was somewhere along the lines of one gigabyte of text per second. Fast enough to be unnoticeable even if you have millions of files in your filesystem.