The fast json sampler
We live in an age where APIs are all around us. Integration with an API means being able to send requests and parse responses. In order to do this, in the statically typed world, we need to create structures that represent each request/response pair that we're handling.
Hand coding the classes becomes very tedious. Enough thanks can't be said to Jon Keith for making the great json2csharp tool, that when given a json, will spit out a set of classes that represent the structure of that json -- Simple and beautiful.
(Aside: It looks as if json2csharp is joining forces with Quicktype and a quick look at Quicktype says that they support a much more wider variety of languages; kudos, Jon!)
So given a json, getting a set of class scaffolding in C#/Java/et al. isn't difficult. So what's the real problem?
The real problem comes when the json is huge. When it is huge, we run into the natural browser limits of pasting text more than a few MB and we are forcing quicktype / json2csharp to process that info for us.
Enter jsam.
jsam is a simple and fast json sampler. The aim of jsam is to give you just enough a json that can be thrown to quicktype or json2csharp to generate the set of classes required to parse it.
- Cross platform support OSX/Linux/Windows
- Simple installation (unzip!)
- Zero dependencies
- Simple CLI
Run jsam on a json file
$ jsam /path/to/file.json
Indent the output json
$ jsam --indent-output /path/to/file.json
Time the whole process
$ jsam --time /path/to/file.json
The way jsam works is by looking into the key value pairs and:
- If a value is an array, it picks only the first one out of this
- If a value is a dictionary, it recurses and repeats the entire process
- If a value is a basic value, it just adds it back
- Support a
--depth <n>flag that, if an array is found would pick outnelements instead of just one - Support a
--url <url>flag that would fetch the json from a given URL instead of a file location