Metadata-Version: 2.4
Name: hdfa-core
Version: 5.0.0
Summary: A non-gradient, cache-native Hyper-Dimensional Fluid Automaton AI core for ultra-low-energy code synthesis.
Author: Sunil Sherikar
Author-email: sunilsv26@gmail.com
License-Expression: Apache-2.0
Classifier: Programming Language :: Python :: 3
Classifier: Operating System :: OS Independent
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: torch>=2.0.0
Requires-Dist: aiohttp>=3.8.0
Requires-Dist: beautifulsoup4>=4.11.0
Requires-Dist: streamlit>=1.30.0
Dynamic: author
Dynamic: author-email
Dynamic: classifier
Dynamic: description
Dynamic: description-content-type
Dynamic: license-file
Dynamic: license-expression
Dynamic: requires-dist
Dynamic: requires-python
Dynamic: summary

# HDFA Core v5.0.0 — Privacy-First Desktop Sandbox AI Chat 🚀

[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.20782928.svg)](https://doi.org/10.5281/zenodo.20782928)

An ultra-lightweight, zero-cloud software intelligence tool driven completely by **Hyperdimensional Computing (HDC)** [0.1.2, 0.2]. Unlike resource-heavy generative models, this framework compresses an entire production software system into an ultra-lean **1MB spatial vector map** executing entirely on local CPU cache lines [0.1.2, 0.2]. Query, audit, and auto-discover repository structures in under **5 milliseconds** with zero token overhead and zero data leaks [0.1.2, 0.2].

---

## 🧱 Key Features

- **Dynamic Workspace Sandbox Isolation:** Fully decoupled multi-window runtime engine executing over randomized Node-to-Python port configurations natively to prevent process context cross-talk.
- **High-Resolution Vector Knowledge Base:** Stream-optimized line extraction training loops that keep laptop RAM overhead flat under 1% while archiving 8,714 precise repository signatures.
- **Low-Latency Fuzzy Search Core:** Localized character n-gram encoding loops computing hyperdimensional vector dot products in under 5ms on standard consumer CPUs [0.1.2, 0.2].
- **Air-Gapped Operational Compliance:** Operates entirely offline with zero network requests, making it completely safe for proprietary, financial, or secure networks.

---

## 📦 Rapid Installation

### 1. Download the Python Core Backend Engine

Install the distributed mathematics package natively via PyPI:

```bash
pip install hdfa-core
```

### 2. Install the VS Code Client App

Search for **"HDFA Core Chat Assistant"** inside the official Microsoft Extensions Marketplace panel and click **Install** [0.1.2, 0.2].

---

## 🏎️ Quick Start Workflow

### 1. Ingest Your Codebase (Run Once Per Project)

Open your terminal window directly inside your target codebase directory (e.g., a React or Python workspace) and trigger the low-RAM streaming aggregator:

```bash
hdfa-train
```

_This scans your local files line-by-line and drops a highly precise text trajectory snapshot file (`codebase_brain.pt`) into your folder root [0.1.2, 0.2]._

### 2. Launch the Ecosystem Sidebar Panel Chat

1. Open your project folder workspace inside **VS Code**.
2. Click on the double speech bubble **HDFA Core AI** icon button located on your far-left Activity Bar.
3. Your interactive side panel will instantly initialize, safely rehydrate your local code snapshot, and open a secure, sandboxed port connection automatically.
4. Type any architectural query or functional keyword block (e.g., `useState` or `handleLogin`) inside the prompt feed container and click **Send** to receive matching structural contexts instantly [0.1.2, 0.2]!

---

## 🛡️ License

Distributed under the **Apache License 2.0**. Permanently locked into open-source scientific history under academic **Zenodo DOI Reference Tracking** [0.2].
