JSON Parser (C)

A lightweight JSON parser written in pure C.
It tokenizes JSON input, builds an Abstract Syntax Tree (AST), and provides a path-based value fetcher.

Features

• Lexer / Tokenizer — converts JSON text into a stream of tokens
• Parser — builds an AST from tokens (supports objects, arrays, and strings)
• Pretty printer — prints the AST in a readable nested format
• Path fetcher — retrieve values using a simple path syntax:
  • Nested objects: nested.key1
  • Array elements: items.[0]
  • Deep access: users.[0].name, deep.level2.level3

Supported JSON Types

Currently supported:

• Objects ({})
• Arrays ([])
• Strings ("...")

Numbers, booleans (true/false), and null are not yet implemented.

Project Structure

src
├── main.c          # Example usage + tests
├── lexer.c         # Parser + AST utilities
├── lexer.h         # AST type definitions
├── fetcher.h       # Path fetcher interface 
├── fetcher.c       # Fetch realization
├── tokenizer.h     # Token definitions
└── tokenizer.c     # Tokenizer realisation

Building

xmake

Usage Example

#include "fetcher.h"

char *data = read_from_file("test/main.json");
Token *tokens = tokenize(data);
Object *ast = tokens_to_ast(tokens);

// Fetch values by path
char *name = (char *)fetch_from_ast("users.[0].name", ast);
char *age  = (char *)fetch_from_ast("users.[1].age", ast);
char *item = (char *)fetch_from_ast("items.[0]", ast);

printf("Name: %s\n", name);
printf("Age:  %s\n", age);
printf("Item: %s\n", item);

// Cleanup
destroy_object(ast);
free(tokens);
free(data);

Path Syntax

Path	Meaning
key	Top-level object key
nested.key	Nested object access
items.[0]	First element of an array
users.[1].name	Field inside an object in array
deep.level2.level3	Deep nested access

Memory Management

The library provides cleanup functions:

• destroy_object(Object *obj)
• destroy_list(List *list)
• destroy_value(ValueTyped *vt)

Always free the AST and the original token buffer after use.

Limitations / TODO

• Support for numbers
• Support for true / false / null
• Better error reporting with line/column information
• Escape sequences in strings (\n, \", etc.)
• Complete ast_to_tokens (serialization)
• More robust memory safety checks

License

MIT (or whatever you prefer)

---

Made with ❤️ in pure C