Neologe - Neologism Registration API

An API-backed system that allows registered users to submit neologisms, which are then evaluated by multiple LLM providers and reviewed for conflicts.

Features

• ✅ User registration and authentication
• ✅ Neologism submission and management
• ✅ Integration with multiple LLM providers (OpenAI, Anthropic, Google)
• ✅ Automated conflict detection and resolution requests
• ✅ JSON-templated response format for word definitions
• ✅ SQLite database with proper schema
• ✅ Token-based authentication
• ✅ RESTful API design

Quick Start

1. Start the server:

   python neologe_server.py

2. Try the example client:

   python example_client.py

3. Run tests:

   python test_api.py

API Endpoints

• POST /auth/register - Register a new user
• POST /auth/login - User login
• POST /neologisms/ - Submit a new neologism
• GET /neologisms/ - List user's neologisms
• GET /neologisms/{id} - Get neologism details
• POST /neologisms/{id}/resolve - Resolve conflicts for a neologism

How It Works

1. User submits a neologism with their definition and optional context
2. Three LLM providers (OpenAI, Anthropic, Google) analyze the word independently
3. Each provider responds with a standardized JSON format containing:
   • Definition, part of speech, etymology
   • Word variations (plural, adjective forms, etc.)
   • Usage examples and confidence score
4. A single LLM evaluates the three responses for conflicts
5. If conflicts exist, the user is notified and can resolve them
6. Final status is updated based on evaluation results

LLM Response Format

{
  "word": "technophilic",
  "definition": "Having a strong affinity for technology",
  "part_of_speech": "adjective",
  "etymology": "From Greek 'techno-' (art, skill) + 'philic' (loving)",
  "variations": {
    "noun": "technophile",
    "adverb": "technophilically"
  },
  "usage_examples": ["She has a technophilic approach to problem-solving."],
  "confidence": 0.92
}

Example Usage

from neologe_client import NeologeClient

# Initialize client
client = NeologeClient("http://localhost:8000")

# Register and login
client.register("wordsmith", "user@example.com", "password")
client.login("wordsmith", "password")

# Submit a neologism
result = client.submit_neologism(
    word="digitality",
    definition="The quality or state of being digital",
    context="Used in discussions about digital transformation"
)

# Check status and resolve conflicts if needed
if result['status'] == 'conflict':
    client.resolve_conflict(result['id'], "accept_consensus")

Documentation

• See API_DOCUMENTATION.md for detailed API documentation
• Run the server and visit http://localhost:8000/ for endpoint reference

Architecture

• Backend: Python with built-in libraries (http.server, sqlite3, json)
• Database: SQLite for development (easily replaceable with PostgreSQL/MySQL)
• Authentication: HMAC-based tokens with password hashing
• LLM Integration: HTTP clients for OpenAI, Anthropic, and Google APIs
• Conflict Resolution: Automated analysis with user override capabilities

Status Values

• pending - Initial submission
• evaluated - LLM analysis complete, no conflicts
• conflict - Conflicts detected, user resolution required
• resolved - User has resolved conflicts
• llm_error - Error during LLM processing

Production Notes

This implementation uses Python's built-in libraries for maximum compatibility. For production use, consider:

• Replace http.server with a proper WSGI server (Gunicorn, uvicorn)
• Use production database (PostgreSQL, MySQL)
• Configure real LLM API keys in .env file
• Add rate limiting, logging, and monitoring
• Implement proper CORS and security headers