Implement local storage backend (JSON)

[Fredbcx]

Context

Currently Tracker holds all lineage data in memory and loses it when the process ends.
To be usable in real workflows, MLineage needs to persist the lineage graph to disk.

Task

Implement a JsonStorageBackend in mlineage/storage/ that:

• serializes the LineageGraph to a .json file on tracker.save()
• deserializes it on Tracker.load()
• stores files under the path passed as storage= in the Tracker constructor

Interface to implement

# mlineage/storage/json_backend.py
class JsonStorageBackend:
    def save(self, graph: LineageGraph, path: str) -> None: ...
    def load(self, path: str) -> LineageGraph: ...

Notes

• All core models already have @dataclass — use dataclasses.asdict() as a starting point for serialization
• Datetime fields need special handling (ISO format strings)
• Tests go in tests/unit/test_storage.py

This is a self-contained task with a clear interface — good entry point into the codebase.