feat(p/json): optimize json

[notJoon]

Description

Optimized the JSON package and simplified JSON node creation using the builder pattern.

• in buffer.gno and escape.gno files are modified the use of map for lookup tables to use slice array instead.
• refactor the Unquote function in escape.gno file
• modified the existing functions that parsed numbers to use strconv package, and deleted related files and functions
  • especially, the eisel_lemire and ryu packages were deleted since they were files that had been added to handle ParseUint and ParseFloat in strconv package.

JSON Generate Example

Plain JSON

node := Builder().
	WithString("name", "Alice").
	WithNumber("age", 30).
	WithBool("is_student", false).
	Node()

value, err := Marshal(node)
if err != nil {
	t.Errorf("unexpected error: %s", err)
}

Output:
{"name":"Alice","age":30,"is_student":false}

Nested Structure

node := Builder().
	WriteString("name", "Alice").
	WriteObject("address", func(b *NodeBuilder) {
		b.WriteString("city", "New York").
		WriteNumber("zipcode", 10001)
	}).
	Node()
// ...

Output:
{"name":"Alice","address":{"city":"New York","zipcode":10001}}

Benchmark Result for Unquote

Before

BenchmarkUnquote-8              	12433488	        98.06 ns/op	     144 B/op	       2 allocs/op
BenchmarkUnquoteWorstCase-8     	24727736	        50.46 ns/op	      48 B/op	       1 allocs/op
BenchmarkUnquoteBestCase-8      	22542354	        52.69 ns/op	      48 B/op	       1 allocs/op
BenchmarkUnquoteEmptyString-8   	394868628	         3.067 ns/op	       0 B/op	       0 allocs/op

After

BenchmarkUnquote-8              	12464704	        96.61 ns/op	     144 B/op	       2 allocs/op
BenchmarkUnquoteWorstCase-8     	25084070	        48.02 ns/op	      48 B/op	       1 allocs/op
BenchmarkUnquoteBestCase-8      	23383227	        52.66 ns/op	      48 B/op	       1 allocs/op
BenchmarkUnquoteEmptyString-8   	400496838	         2.968 ns/op	       0 B/op	       0 allocs/op

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 62.93%. Comparing base (93ece90) to head (f09e339).
Report is 40 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2939      +/-   ##
==========================================
- Coverage   63.09%   62.93%   -0.16%     
==========================================
  Files         563      568       +5     
  Lines       79273    82371    +3098     
==========================================
+ Hits        50015    51843    +1828     
- Misses      25893    27081    +1188     
- Partials     3365     3447      +82     

Flag	Coverage Δ
contribs/gnodev	60.57% <ø> (-0.05%)	⬇️
contribs/gnofaucet	15.77% <ø> (+0.94%)	⬆️
gno.land	67.37% <ø> (-0.09%)	⬇️
gnovm	67.43% <ø> (+0.18%)	⬆️
misc/genstd	79.72% <ø> (ø)
tm2	62.43% <ø> (+0.12%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jefft0]

Hi @notJoon. Maybe it's obvious to people familiar with the code, but in the description can you put a short explanation why you removed the eisel_lemire and ryu folders?

[jefft0]

This is follow-on PR by the same author of the original package in #1415 . On my macOS, all tests pass with make test.components . Removed the review/triage-pending label. Ready for review by core devs.

[notJoon]

Hi @notJoon. Maybe it's obvious to people familiar with the code, but in the description can you put a short explanation why you removed the eisel_lemire and ryu folders?

description updated. thanks!

[thehowl]

BTW the benchmarks are in go and are likely to mean very little in Gno ;)

At most, use -print-runtime-metrics for a comparison, even though it's still going to not be terribly useful to understand real-world CPU usage (yet).