Totally agree! In my system, I run Phoenix Channels separately from other pages/LiveViews, so they don’t affect others.
manhvu
LostKobrakai
json_library is not just use for channels either. It’s used wherever phoenix uses json, including your api responses or whatever else you’d do with json within the web portion of the system.
1
manhvu
I have applied TOON to our API. The backend uses Phoenix Channels.
In our case, most data before being passed to Phoenix and pushed to the client looks like this:
%{
request: "request1",
data: [
%{a: 1, b: "test", c: "other"},
%{a: 2, b: "test 2", c: "other 2"},
%{a: 3, b: "test 3", c: "other 3"}
]
}
Previously, the client would receive JSON:
{"data":[{"c":"other","a":1,"b":"test"},{"c":"other 2","a":2,"b":"test 2"},{"c":"other 3","a":3,"b":"test 3"}],"request":"request1"}
Converted to a pretty format:
{
"data": [
{
"c": "other",
"a": 1,
"b": "test"
},
{
"c": "other 2",
"a": 2,
"b": "test 2"
},
{
"c": "other 3",
"a": 3,
"b": "test 3"
}
],
"request": "request1"
}
With TOON:
data[3]{a,b,c}:
1,test,other
2,test 2,other 2
3,test 3,other 3
request: request1
In my case, arrays can range from a few items to thousands, so TOON significantly reduces network traffic and makes development and debugging easier (JSON is actually hard to track in its compact format).
I have also reworked the TOON library and optimized it to run faster in Phoenix.
Library: ToonEx
1
xu-chris
That’s brilliant! Let’s open a PR on the original repo to push back some great work you did!
0
manhvu
Actually, I renamed and published it under a different name. I want to optimize it and add more supporting functions for use with Phoenix Channels.
0