entroly-core 0.3.2

Rust core for Entroly — information-theoretic context optimization (knapsack, entropy, dedup, SAST, query analysis)

entroly-core

Rust core for Entroly — information-theoretic context optimization for AI coding agents.

Provides high-performance PyO3 bindings for:

• Knapsack optimizer — 0/1 DP context selection within token budget
• Shannon entropy scorer — boilerplate detection, information density
• SimHash deduplication — near-duplicate fragment detection
• Query analysis — TF-IDF vagueness scoring, heuristic refinement
• SAST scanner — 30+ security rules (XSS, SQL injection, secrets, unsafe memory)
• LSH index — approximate nearest-neighbor semantic recall
• PRISM RL optimizer — online feedback-driven fragment weight learning

Install

pip install entroly-core

Prebuilt wheels for Linux, macOS, Windows (Python 3.10–3.13).

Usage

Usually used via the higher-level entroly package:

pip install entroly
entroly  # starts the MCP server

Or directly:

from entroly_core import ContextFragment, py_knapsack_optimize, py_shannon_entropy

License

MIT