rsn-db 0.2.1

Rust-powered embedded database for Python

RSN DB

A Rust-powered embedded database with a personality.

Features

• Insane Safety & Security: AES-GCM Encryption, Zstd Compression, and SHA-256 Checksums.
• Dynamic Personality: Choose your experience—Professional, Friendly, or Snarky with richer reactions, achievements, and batch summaries.
• Interactive CLI: Just type rsn and get to work.
• Smart Data Handling: Constraints, Aliases, Batch Operations, and Smart Type Coercion.
• Command UX Features: Talk to your data with FIND, inspect schemas with DESCRIBE, and review recent activity via HISTORY.

---

Interactive Session Demo

Snarky Mode Activation

Caption: Setting up RSN DB for the first time and enabling full snark mode.

$ rsn
Select mode:
  [1] Professional (clean, minimal output)
  [2] Friendly     (helpful with personality)
  [3] Snarky       (full commentary enabled)

Choice (default: 1): 3
Remember this choice? (y/n): y

✓ Snarky mode enabled.
  Don't say I didn't warn you.

rsn>

Natural Language Query

Caption: Using the FIND command to translate complex questions into database queries.

rsn> FIND users older than Bob who follow someone
⚙ Translating...
  Interpreted as: READ users WHERE age > (SELECT age FROM users WHERE name = "Bob") AND has_outbound_edge("FOLLOWS")
Is that it?
Y for yes, N or blank for no
r>y
╭── Results ────────────────────╮
│  • Alice (30)                 │
│  • Charlie (35)               │
╰───────────────────────────────╯

Batch Operations & Achievement

Caption: Executing a fast batch of inserts and getting roasted/rewarded by the engine.

rsn> BATCH
batch> INSERT users (name: "User1", age: 20)
batch> INSERT users (name: "User2", age: 21)
batch> INSERT users (name: "User3", age: 22)
batch> COMMIT

✓ Batch executed: 3 operations.
  (Wow! You're fast! Very hard work. Good job. My grandmother can't do that. Not that I have one, since I'm a robot.)

... (after 50 perfect commands) ...

[SYSTEM]: Achievement unlocked: "Actually Competent"
[SYSTEM]: Updating user classification from "Hopeless" to "Occasionally Capable"

---

Quickstart

Installation

pip install rsn_db

CLI Usage

rsn

Library Usage

from rsn_db import Database

# The library is all business. No snark here.
db = Database(storage_path="my_data.rsn", encryption_key="secret-key")
db.create_table("users", {"name": {"type": "string", "required": True}})
db.insert("users", {"name": "Alice"})

---

Security & Safety

RSN DB takes your data seriously (even if it doesn't take you seriously).

• Encryption at Rest: AES-256-GCM for persisted payloads when encryption_key is set.
• Integrity: SHA-256 checksums detect persisted file tampering/corruption before decode.
• Identifier Safety: strict identifier validation for table/field names used in generated SQLite SQL.
• Path Safety: import/export paths must be relative and cannot contain parent traversal segments or absolute/prefixed roots. Database storage paths reject traversal segments and invalid prefixes.
• DoS Guardrails: hard limits for command length, batch operation count, ingest payload size, JSONL import file size, and JSONL line count.

For operational guidance and security limits, see documentation/security.md.

---

License

MIT

GraphRAG (New in v0.2.0)

RSN DB now includes a built-in GraphRAG engine for knowledge retrieval without heavy LLMs. Usage:

db.ingest("Large text document...", source="my_doc")
print(db.graph_query("What are the key entities?"))

Or via SQL:

INGEST "My interesting text..."
GRAPH_QUERY Who is related to what?