deccan-toon 0.2.0

Battling JSON bloat with Pythonic elegance. A high-efficiency TOON serializer.

Deccan Toon

"First, we shape our tools, thereafter they shape us."

Deccan Toon is a strict, high-velocity implementation of the TOON v3.0 Specification.

While other parsers focus on flexibility, Deccan Toon focuses on correctness, security, and type fidelity. We built this to be the "Newtonsoft" of the TOON ecosystem—robust enough for enterprise agents and critical infrastructure.

---

📉 Token Economics (The "Why")

Standard JSON repeats keys for every single row (id, name, status), wasting precious context window space in LLMs (GPT-4, Claude). TOON sends the header once.

Payload Size Comparison (1000 items):

████████████████████ 100% - JSON (Standard)
████████████         60%  - CSV (No types)
███████████          55%  - TOON (Typed & Clean)

Stop paying for repeated keys. Save context window space for actual intelligence.

---

📦 Installation

pip install deccan-toon

---

⚡ Quick Start

import deccan_toon

# 1. Define your data (List of Dicts)
data = [
    {"id": 1, "agent": "Alpha", "status": "Active", "score": 0.98},
    {"id": 2, "agent": "Beta",  "status": "Idle",   "score": None},
    {"id": 3, "agent": "Gamma", "status": "Sleep",  "score": 0.45}
]

# 2. Serialize (Python -> TOON)
payload = deccan_toon.dumps(data)

print(payload)
# Output:
# items[3]{id,agent,status,score}:
#   1, Alpha, Active, 0.98
#   2, Beta, Idle, null
#   3, Gamma, Sleep, 0.45

# 3. Deserialize (TOON -> Python)
restored = deccan_toon.loads(payload)

---

🛡️ The Deccan Difference

• Strict Compliance: Fully compliant with TOON v3.0 (Handling nesting, typed arrays, and sparse data).

• Security First: Includes protections against "Billion Laughs" expansion attacks and deep recursion.

• Type Safety: Unlike CSV, we preserve int, float, bool, and null types with 100% fidelity.

• Zero Dependencies: Runs on pure Python standard library.