hapijson: A Go repository from busmaker

In most cases, encoding/json is great but if you only want to read or modify a little piece of json it would be an expensive way to do so, this is why we need hapijson comes into play.

Install

go get github.com/LBJ-the-GOAT/hapijson

Usage

some Getters

jsonData := `{"name": "LBJ", "height": 2.04, "title": 3, "teams":["LAL", "CAVS"]}`

hapijson.Get(jsonData, "teams", 0)
// outputs LAL, 0 means the index of an array

hapijson.String(jsonData, "name")
// outputs LBJ

hapijson.Int(jsonData, "title")
// outputs 3

hapijson.StringArray(jsonData, "teams")
// outputs []string{"LAL", "CAVS"}

Set

jsonData := {"teams": ["LAL"] }
jsonData, _ = hapiJSOn.Set(jsonData, "LA Lakers", "teams", 0)
// "teams": ["LAL"] becomes "teams": ["LA Lakers"]

Merge

jsonData := {"name": "LBJ", "height": 2.04}
jsonData, _ = hapijson.Merge(jsonData, true, hapijson.Path(), "mvp", 4, "height": "6 ft. 8 in.")
// jsonData now is {"name": "LBJ", "height": [2.04, "6 ft. 8 in."], "mvp": 4},
// the value of key "height" now became an array, and key "mvp" is appended.

Append

jsonData := {"name": "LBJ", "teams": ["LAL Lakers"]}
jsonData, _ = hapijson.Append(jsonData, hapijson.Path("teams"), "CAVS", "HEAT")
//  "teams":["LA Lakers"] becomes "teams":["LA Lakers", "CAVS", "HEAT"]

Remove

jsonData := {"name": "LBJ", "height": 2.04}
jsonData, _ = hapijson.Remove(jsonData, "height");
// jsonData = {"name": "LBJ"}, key "height" is removed.

Clear

jsonData := {"teams":["LA Lakers", "CAVS", "HEAT"]}
jsonData, _ = hapijson.Clear(jsonData, "teams");
//  jsonData = {"teams":[]}, "teams" array is empty now.

Increase

jsonData := {"name": "LBJ", "title": 3}
jsonData, _ = hapijson.Incr(jsonData, 2, "title");
// jsonData ={"name": "LBJ", "title": 5},  "title": 3 is now increased by 2 to "title": 5

main features have been displayed above, look in /examples for more.

Benchmark

Because the time complexity of getter & setter functions are kind of O(n+m), n is the position of the last path node in the json data, m is the length of value of the path node, so I placed the keys at the very end of the data to get the outcomes of worst situation.

The results are pretty close. allocs/op mostly for keys escaping, means the more keys in path nodes the more allocs/op

json size	500B	2K	20K	280K	2.8M	8.6M
Get	4253 ns/op, 12 allocs/op	7762 ns/op, 39 allocs/op	43011 ns/op, 23 allocs/op	≈0.4 ms/op, 29 allocs/op	≈5.1 ms/op, 27 allocs/op	≈16.2 ms/op, 27 allocs/op
Set	5382 ns/op, 16 allocs/op	8027 ns/op, 40 allocs/op	62824 ns/op, 27 allocs/op	≈0.4 ms/op, 6 allocs/op	≈7 ms/op, 32 allocs/op	≈23 ms/op, 32 allocs/op
Append	4677 ns/op, 14 allocs/op	6919 ns/op, 18 allocs/op	54962 ns/op, 24 allocs/op	≈0.56 ms/op, 9 allocs/op	≈6.86 ms/op, 32 allocs/op	≈22.96 ms/op, 34 allocs/op
Merge	4559 ns/op, 15 allocs/op	8876 ns/op, 40 allocs/op	43238 ns/op, 26 allocs/op	≈0.41 ms/op, 6 allocs/op	≈5.05 ms/op, 28 allocs/op	≈20.74 ms/op, 28 allocs/op
Remove	4286 ns/op, 12 allocs/op	8249 ns/op, 37 allocs/op	45354 ns/op, 23 allocs/op	≈0.4 ms/op, 3 allocs/op	≈5.4 ms/op, 25 allocs/op	≈17.2 ms/op, 25 allocs/op
Clear	4194 ns/op, 12 allocs/op	8207 ns/op, 37 allocs/op	44915 ns/op, 23 allocs/op	≈0.4 ms/op, 3 allocs/op	≈5.3 ms/op, 25 allocs/op	≈16.65 ms/op, 25 allocs/op
Incr	4078 ns/op, 12 allocs/op	5640 ns/op, 16 allocs/op	39334 ns/op, 26 allocs/op	≈0.56 ms/op, 8 allocs/op	≈ 6.97 ms/op, 29 allocs/op	≈15.94 ms/op, 29 allocs/op
Valid	1180 ns/op, 0 allocs/op	3723 ns/op, 0 allocs/op	45410 ns/op, 0 allocs/op	≈0.44 ms/op, 0 allocs/op	≈3.93 ms/op, 0 allocs/op	≈11.71 ms/op, 0 allocs/op
Minify	1364 ns/op, 0 allocs/op	3799 ns/op, 0 allocs/op	38470 ns/op, 0 allocs/op	≈0.39 ms/op, 0 allocs/op	≈ 4.76 ms/op, 0 allocs/op	≈13.26 ms/op, 0 allocs/op
Prettify	5848 ns/op, 2 allocs/op	16282 ns/op, 2 allocs/op	131155 ns/op, 2 allocs/op	≈1.66 ms/op, 3 allocs/op	≈63.41 ms/op, 35 allocs/op	≈39.44 ms/op, 2 allocs/op

busmaker/hapijson