chore: update dependencies
fix: koblenz-press-releases.lua, selectors.go
fix: apnews.lua
RSS all the things!
ratt is a tool for converting websites to rss/atom feeds. It uses lua config files which define the extraction of the feed data by using css selectors or lua functions.
Config files are in lua format:
--for automatic extraction, ratt checks all config files and matches the regex
ratt.add(
--regex
"https://github.com/trending",
--css selectors table
{
--settings for all http requests for the website
httpsettings = {
cookie = {},
header = {},
useragent = "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/90.0.4430.72 Safari/537.36"
},
--css selectors to get the feed data
feed = {
title = ".h1",
},
--css selectors to get item data
item = {
--the item container
container = "article.Box-row",
--selector can be a function which get's the item container selection object
title = function(sel, _)
return sel:find("h1.h3 > a[data-hydro-click]"):text():gsub("%s+", "")
end,
link = function(sel, _)
return "https://github.com" .. sel:find("a[data-hydro-click]"):attr("href")
end,
description = "p.color-fg-muted",
}
}
)
Config files are lua files. ratt has some confs embedded. When calling eg: ratt https://1337x.to/top-100
ratt will try to find the config for the website url, it searches the embedded config files, the current directory and in ~/.config/ratt/*.lua
.
First install go
, git
and scdoc
, then:
git clone https://git.sr.ht/~ghost08/ratt
cd ratt
sudo make install
Install on Arch Linux from AUR with your favorite helper:
yay -S ratt-git
File bugs and TODOs through the issue tracker or send an email to ~ghost08/ratt@todo.sr.ht. For general discussion, use the mailing list: ~ghost08/ratt@lists.sr.ht.
Just calling ratt with the url of the web page.
ratt https://github.com/trending/go
man ratt.5
That's a very good question. I'm happy you asked :)
You might feed the feed directly to photon, which is a modern RSS/Atom reader. photon will play you the media from your feed. It uses mpv and youtube-dl to automaticaly play videos, download torrents, view images and much more :)
So try this out:
ratt https://1337x.to/top-100 | photon -
If a css selector isn't enough to select the needed data, every feed and item attribute can be a lua function.
The function gets two arguments by default:
sel
is the selection object of the feed/item container on which it can be queried for the selectors
index
number of the item processed
The Lua script will get some modules to help with the extraction:
goquery
is a module imported by default and it is a subset of the famous goquery library
gojq
is a module imported by default, it is the gojq) library
ratt will take the return value of the Lua function and insert it as the data of the feed/item. When a error has occured, just use the error
function.
For more documentation see ratt(5)
Calling another link, parsing it to a goquery.Document and querying the new doc:
item = {
--select the item container html element
container = ".table-list-wrap tbody tr",
--select the title element in the item container
title = "a:nth-child(2)",
--lua script
link = function(sel, _)
--sel is the item container element, find <a/>
a = sel:find("a:nth-child(2)")
--get the href attribute of <a/> and make a item url link from it
itemURL = "https://1337x.to" .. a:attr("href")
--request and parse the document
doc, err = goquery.newDocFromURL(itemURL)
if err ~= nil then
--return error if the request was unsuccesfull
error(err)
end
--find the item link you want
link = doc:find("ul li a[onclick]"):first():attr("href")
--trim space characters
link = link:gsub("%s+", "")
--and finally print the link out so ratt can include it in the item.link
return link
end,
}
You can also parse and query json data, with the help of the awesome gojq) library:
feed = {
title = ".title",
description = function(sel, _)
--find the <script> element where the json data is
script = sel:find("script"):first():text()
index = script::find("var myJsonData =")
--cut of the "var myJsonData =" prefix
jsonData = script:sub(index+16)
--parse a gojq query, that will find the obj["description'] value
query, err = gojq.parse(".description")
if err ~= nil then
error(err)
end
--expecting that the input data is a map/object (otherwise if it's a array use runArray)
desc, err = query.runMap(jsonData)
if err ~= nil then
error(err)
end
return desc[1]["description"]
end,
}
Check the confs dir for other examples.
ratt needs config files for it to run. I really rely on the community to create configs for all the sites!
So please create config files, send them here, then everybody can make the world RSS again!
Anyone can contribute to ratt:
Once you are happy with your work, you can create a commit (or several commits). Follow these general rules:
git log --oneline
.Fixes:
trailer with the commit id and its title.There is a great reference for commit messages in the Linux kernel documentation.
Before sending the patch, you should configure your local clone with sane defaults:
git config format.subjectPrefix "PATCH ratt"
git config sendemail.to "~ghost08/ratt@lists.sr.ht"
And send the patch to the mailing list:
git sendemail --annotate -1
Wait for feedback. Address comments and amend changes to your original commit. Then you should send a v2:
git sendemail --in-reply-to=$first_message_id --annotate -v2 -1