Grand Diomande Research · Full HTML Reader

Packages

``` packages/ ├── ircp/ # Inverse Ring Contextual Propagation ├── tpo/ # Temporal Positional Optimization ├── rcp/ # Ring Contextual Propagation ├── ctsc/ # Computational Topology Search Coordinates └── dlm/ # Dynamic Liquid Motion ```

Agents That Account for Themselves research note backlog reference score 18 .md

Full Public Reader

Packages

This directory contains all core Python packages and libraries for the CC-TPO project.

Packages Overview

packages/
├── ircp/       # Inverse Ring Contextual Propagation
├── tpo/        # Temporal Positional Optimization
├── rcp/        # Ring Contextual Propagation
├── ctsc/       # Computational Topology Search Coordinates
└── dlm/        # Dynamic Liquid Motion

---

IRCP (Inverse Ring Contextual Propagation)

Location: `packages/ircp/`

### Description
Core implementation of the IRCP algorithm for semantic embedding and conversation analysis using ring topology.

### Key Components
- Models: Custom SentenceTransformer implementation
- Algorithms: Ring topology algorithms
- Utilities: Embedding generation and similarity search

Usage

python
import sys
from pathlib import Path

# Add to path
sys.path.append(str(Path.cwd() / "packages" / "ircp"))

from ircp.models.sentence_transformer_icp import SentenceTransformerICP

# Initialize model
config = {
    "model_name": "sentence-transformers/all-MiniLM-L6-v2",
    "embedding_dim": 384,
    "coordinate_dim": 4,
    "hidden_dim": 512,
    "dropout": 0.1
}
model = SentenceTransformerICP(config)

### Used By
- `apps/liquid-chat-backend/` - Main embedding service
- `services/search-api/` - All search services
- Training scripts

---

TPO (Temporal Positional Optimization)

Location: `packages/tpo/`

### Description
Temporal and positional optimization framework for DLM (Dynamic Liquid Motion) coordinates.

### Key Components
- DLM Coordinates: Dynamic coordinate system (x, y, z, t, n)
- Calculators: Coordinate calculation algorithms
- Optimizers: Position optimization

### Coordinate System
- x: Hierarchy depth (conversation tree depth)
- y: Sibling position (lateral position in tree)
- z: Semantic homogeneity (0-1, topic coherence)
- t: Temporal position (normalized timestamp)
- n: Structural complexity (paragraph count/complexity)

Usage

python
from tpo.core.dlm_coordinates import DLMCalculator, DLMCoordinates

calculator = DLMCalculator()
coordinates = calculator.calculate_coordinates(
    message_id,
    messages_dict,
    parent_child_map
)

### Used By
- `apps/liquid-chat-backend/` - Coordinate calculation
- IRCP training pipeline

---

RCP (Ring Contextual Propagation)

Location: `packages/rcp/`

### Description
Ring-based contextual propagation framework for conversation flow analysis.

### Key Features
- Ring topology construction
- Context propagation algorithms
- Message relationship mapping

---

CTSC (Computational Topology Search Coordinates)

Location: `packages/ctsc/`

### Description
Computational topology and search coordinate algorithms.

### Key Features
- Topological search algorithms
- Coordinate space transformations
- Distance metrics

---

DLM (Dynamic Liquid Motion)

Location: `packages/dlm/`

### Description
Dynamic liquid motion visualization and coordinate system.

### Key Features
- Liquid motion algorithms
- Visualization helpers
- Animation frameworks

---

Importing Packages

From Applications

In Python files within `/apps/` or `/services/`:

python
import sys
from pathlib import Path

# Calculate path to project root
PROJECT_ROOT = Path(__file__).parent.parent.parent  # Adjust based on nesting

# Add package paths
sys.path.append(str(PROJECT_ROOT / "packages" / "ircp"))
sys.path.append(str(PROJECT_ROOT / "packages" / "tpo"))

# Import
from ircp.models.sentence_transformer_icp import SentenceTransformerICP
from tpo.core.dlm_coordinates import DLMCalculator

Path Configuration

Directory Structure:

cc-tpo/
├── apps/
│   └── liquid-chat-backend/
│       └── main.py              # Needs: ../../../packages/ircp
├── services/
│   └── search-api/
│       └── search_api.py        # Needs: ../../packages/ircp
└── packages/
    ├── ircp/
    └── tpo/

---

Development

Adding New Packages

1. Create package directory under `packages/`
2. Add `__init__.py` to make it a Python package
3. Update relevant `sys.path` entries in applications
4. Document in this README

Package Structure

Recommended structure for each package:

packages/your_package/
├── __init__.py
├── models/
│   └── __init__.py
├── core/
│   └── __init__.py
├── utils/
│   └── __init__.py
└── tests/
    └── __init__.py

---

Testing

Running Package Tests

bash
# Test IRCP
cd packages/ircp
python -m pytest tests/

# Test TPO
cd packages/tpo
python -m pytest tests/

Integration Tests

Integration tests that use multiple packages should be placed in the top-level `integration/` directory.

---

Dependencies

### IRCP Dependencies
- PyTorch
- sentence-transformers
- numpy
- transformers

### TPO Dependencies
- numpy
- scipy (for optimization)

Installing All Dependencies

bash
pip install -r requirements.txt

Or for development:

bash
pip install -r requirements-dev.txt

---

Migration Notes

> [!IMPORTANT]
> Post-Phase 4 Migration
>
> All packages have been moved from project root to `packages/` directory.
> If you encounter import errors, ensure your `sys.path` is correctly configured.

Before (Phase 1-3):

python
sys.path.append(str(PROJECT_ROOT / "ircp"))
from ircp.models.sentence_transformer_icp import SentenceTransformerICP

After (Phase 4):

python
sys.path.append(str(PROJECT_ROOT / "packages" / "ircp"))
from ircp.models.sentence_transformer_icp import SentenceTransformerICP

---

Documentation

  • IRCP Documentation: `../docs/ircp/`
  • Architecture: `../docs/architecture/`
  • Research Papers: `../docs/research/`

---

Contributing

When modifying packages:

1. Maintain backward compatibility when possible
2. Update tests for any changes
3. Document new features or breaking changes
4. Update this README with new package information

Promotion Decision

Keep in the searchable backlog until it intersects a live paper or system.

Source Anchor

Comp-Core/backend/cc-trajectory/legacy/cc-tpo-original/cc-tpo/packages/README.md

Detected Structure

Method · Figures · Code Anchors · Architecture