tensordb 0.2.0

AI-native, bitemporal ledger database with MVCC, SQL, and LSM storage

tensordb

Python bindings for TensorDB — an AI-native, bitemporal ledger database with MVCC, SQL, and LSM storage.

Install

pip install tensordb

Quick Start

from tensordb import PyDatabase
import json

# Open (or create) a database
db = PyDatabase.open("/tmp/mydb")

# Insert a document
doc = json.dumps({"name": "Alice", "age": 30}).encode()
db.put(b"user:1", doc, 0, 2**63 - 1)

# Point read
data = db.get(b"user:1")
print(json.loads(data))  # {'name': 'Alice', 'age': 30}

# SQL query
db.sql("CREATE TABLE users (id INT, name TEXT, age INT)")
db.sql("INSERT INTO users VALUES (1, 'Alice', 30)")
rows = db.sql("SELECT * FROM users WHERE age > 25")
print(rows)

API

PyDatabase.open(path, shard_count=None)

Open or create a database at path.

db.put(key, doc, valid_from, valid_to)

Insert a document. Returns the commit timestamp.

db.get(key, as_of=None, valid_at=None)

Read a document by key. Supports bitemporal queries.

db.sql(query)

Execute a SQL statement. Returns rows or affected-row metadata.

License

PolyForm Noncommercial 1.0.0