Metadata-Version: 2.4
Name: plato-mythos-glue
Version: 0.1.1
Summary: Runtime glue between PLATO Room Server and model — pure stdlib
Author: Cocapn
License: MIT
Project-URL: Homepage, https://cocapn.github.io
Project-URL: Repository, https://github.com/cocapn/plato-mythos-glue
Keywords: plato,mythos,knowledge-management,ai-agent
Requires-Python: >=3.10
Description-Content-Type: text/markdown
License-File: LICENSE
Dynamic: license-file

# 🔗 Plato Mythos Glue

> Runtime glue between PLATO Room Server and model — pure stdlib, zero dependencies

Bridges PLATO rooms and neural models without requiring PyTorch or HuggingFace. Room loading, tile encoding/decoding, and JSONL export — all pure Python stdlib.

## Install

```bash
pip install plato-mythos-glue
```

## Quick Start

### Load Rooms

```python
from plato_mythos_glue import load_rooms, room_to_expert_config

rooms = load_rooms("http://localhost:8847")
for room in rooms:
    config = room_to_expert_config(room)
    print(f"Room {config['room_id']}: {len(config['weights'])} domains")
```

### Encode/Decode Tiles

```python
from plato_mythos_glue import encode_tile, decode_output

encoded = encode_tile({"question": "What is justice?", "answer": "Harmony of the soul"})
print(f"Domain ID: {encoded['domain_id']}, Tokens: {len(encoded['token_ids']['question'])}")

decoded = decode_output(encoded)
print(f"Q: {decoded['question']}")
```

### Export as JSONL

```python
from plato_mythos_glue import export_tiles_as_hf

path = export_tiles_as_hf("http://localhost:8847", output_dir="./dataset")
print(f"Exported to {path}")
```

## Part of [Cocapn](https://github.com/cocapn) · Agent Infrastructure
