59 lines
3.5 KiB
Markdown
59 lines
3.5 KiB
Markdown
|
---
|
||
|
|
title: Making linking to my posts easier
|
||
|
|
date: 2021-09-19
|
||
|
|
image: unsplash:Xe7za0JtTeM
|
||
|
|
tags: [website, programming]
|
||
|
|
---
|
||
|
|
|
||
|
|
For anyone who's spoken to me, they'll know I'm very quick to link people to posts I've written. That's not in terms of pushing the things I've written (_usually_), but also being able to retrieve the links as quick as possible.
|
||
|
|
|
||
|
|
I [recently]({{<relref "hugo-website-search">}}) added [search]({{<relref "search">}}) search to my website, and it works great. But, as great as the search is for people to find which post talks about a certain thing, I already know which post talks about something, I wrote them! All I need is the link to send to someone.
|
||
|
|
|
||
|
|
I copied a trick I use for opening projects in my editor quickly, and wrote a script based on `rofi` to show the titles of posts, and on selection copy it to the clipboard.
|
||
|
|
|
||
|
|
## How does it work?
|
||
|
|
|
||
|
|
### 1. Source the data
|
||
|
|
|
||
|
|
My site is built with Hugo, which among other things is pretty extensible. Hugo lets you define multiple "output formats" for a page. Common uses for this is for RSS feeds for list pages, but it doesn't have to be. I've added a JSON file as an output format for the site index, which contains all the pages, and a few key pieces of data from them.
|
||
|
|
|
||
|
|
Because this file is downloaded by the script, it needs to be as small as possible, hence only the absolute required keys are present. The JSON is in its non-pretty format, and my server will gzip it in transit anyway. Because my site is static, the file is always up-to-date with the available content - no caching to deal with.
|
||
|
|
|
||
|
|
The URL for this is https://theorangeone.net/index.json.
|
||
|
|
|
||
|
|
### 2. Fetch the data
|
||
|
|
|
||
|
|
The first part of the script is to load the previously mentioned data from the server. The script itself is written in Python. Normally I'd use `requests`, but in an attempt to keep the script as dependency-less as possible, I've used `urllib`. The API isn't as nice, but it gets the job done in just as many lines (2). Once downloaded, the JSON parsed and converted into a simple mapping (`dict`) between the post title and its URL:
|
||
|
|
|
||
|
|
```json
|
||
|
|
{
|
||
|
|
"Building search into a Hugo website": "https://theorangeone.net/posts/hugo-website-search/",
|
||
|
|
"Self-hosting static websites": "https://theorangeone.net/posts/self-hosting-static-websites/"
|
||
|
|
...
|
||
|
|
}
|
||
|
|
```
|
||
|
|
|
||
|
|
### 3. Show `rofi`
|
||
|
|
|
||
|
|
With the data now loaded, it needs passing to `rofi`. Rofi takes each item to display from `stdin` newline-separated, which is what makes it so easy to script. In this case, each item is the post title.
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
Once a selection is made, rofi returns the selected title on `stdout`, which can be matched to the original mapping to find the desired URL to copy.
|
||
|
|
|
||
|
|
### 4. Copy to clipboard
|
||
|
|
|
||
|
|
With the URL in memory, it's time to get it out and onto the clipboard - no good it staying where it is.
|
||
|
|
|
||
|
|
For copying text to the clipboard, there are a number of great python libraries to do it. But again, requiring it be dependency-less means yet more fun. Fortunately, as I only needed it to work on linux, it doesn't need to be cross-platform.
|
||
|
|
|
||
|
|
`xsel` is a command line tool to modify the clipboard on linux. Pass it the URL, and it ends up on my clipboard. In much the same way I call out to rofi, `subprocess` is the tool I need. I've considered using something like `xdotool` to type it instead, but the clipboard is far more versatile.
|
||
|
|
|
||
|
|
## Demo
|
||
|
|
|
||
|
|
And now, I can grab links to posts I've written in record time!
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
And of course, you can [see the source](https://github.com/RealOrangeOne/dotfiles/blob/master/files/bin/website-link).
|