weycop 2.0.0

Python client for WeyCP API - Multi-provider AI chat completions (Local/OpenAI/Anthropic)

WeyCP Python Client

The official Python client for WeyCP API chat completions created by Weycop.

---

🚀 Features

• Multi-Provider Support – Use local models (Ollama), OpenAI, and Anthropic models.
• Familiar API Interface – Easy-to-use chat completions API.
• Auto-Detection – Automatically detects provider based on model name.
• Sync & Async support – Use synchronous or asynchronous clients.
• Built-in error handling – Comprehensive exception handling.
• Usage tracking – Monitor token usage and costs across providers.
• Type hints – Full type annotation support.
• High performance – Optimized for concurrent requests.

---

📦 Installation

pip install weycop

---

⚡ Quick Start

Synchronous Client

from weycop import WeycopClient

# Initialize client
client = WeycopClient(api_key="your-api-key")

# Simple chat with local model
response = client.chat(
    message="Hello, how are you?",
    model="qwen3:4b-instruct",
    system_prompt="You are a helpful assistant"
)
print(response)

Multi-Provider Usage

from weycop import WeycopClient, Message

client = WeycopClient(api_key="your-api-key")

# Local models (free)
local_response = client.chat_local(
    model="qwen3:4b-instruct",
    messages=[Message("user", "Hello from local model!")]
)

# OpenAI models (requires OPENAI_API_KEY in server)
openai_response = client.chat_openai(
    model="gpt-4o",
    messages=[Message("user", "Hello from OpenAI!")]
)

# Anthropic models (requires ANTHROPIC_API_KEY in server)
anthropic_response = client.chat_anthropic(
    model="claude-3-5-sonnet-20241022",
    messages=[Message("user", "Hello from Anthropic!")]
)

# Auto-detection (provider detected from model name)
auto_response = client.chat_completions_create(
    model="gpt-4o",  # Automatically detected as "openai"
    messages=[Message("user", "Auto-detected provider!")]
)

Asynchronous Client

import asyncio
from weycop import AsyncWeycopClient

async def main():
    async with AsyncWeycopClient(api_key="your-api-key") as client:
        # Use any provider asynchronously
        response = await client.chat_local(
            model="qwen3:8b",
            messages=[Message("user", "What is machine learning?")]
        )
        print(response.choices[0].message.content)

asyncio.run(main())

---

🧠 Supported Models

Local Models (Free)

• qwen3:4b-instruct – Fast, efficient model for general conversations (4GB VRAM)
• qwen3:8b – Advanced model with extended context (8.5GB VRAM)

OpenAI Models (Pay-per-token)

• gpt-3.5-turbo – Fast and affordable ($0.0005/$0.0015 per 1K tokens)
• gpt-4o – Advanced reasoning and analysis ($0.0025/$0.01 per 1K tokens)
• gpt-4o-mini – Lightweight and fast ($0.00015/$0.0006 per 1K tokens)
• gpt-4 – Premium model for complex tasks ($0.03/$0.06 per 1K tokens)

Anthropic Models (Pay-per-token)

• claude-3-5-sonnet-20241022 – Most intelligent model ($0.003/$0.015 per 1K tokens)
• claude-3-5-haiku-20241022 – Fastest model ($0.00025/$0.00125 per 1K tokens)
• claude-3-opus-20240229 – Most powerful model ($0.015/$0.075 per 1K tokens)

---

🔍 Model Information

# Get available models by provider
models = client.get_available_models()
print(models)
# Output: {
#   "local": ["qwen3:4b-instruct", "qwen3:8b"],
#   "openai": ["gpt-3.5-turbo", "gpt-4o", "gpt-4o-mini"],
#   "anthropic": ["claude-3-5-sonnet-20241022", ...]
# }

# Get detailed model information
model_info = client.get_model_info("gpt-4o")
print(f"Provider: {model_info.provider}")
print(f"Context: {model_info.context_length} tokens")
print(f"Cost: ${model_info.cost_per_1k_input_tokens}/1K input tokens")

---

🛡️ Error Handling

from weycop import WeycopClient, AuthenticationError, RateLimitError, APIError

client = WeycopClient(api_key="your-key")

try:
    # Try different providers
    response = client.chat_local("Hello", model="qwen3:4b-instruct")
    print(response.choices[0].message.content)
except AuthenticationError:
    print("Invalid API key")
except RateLimitError:
    print("Rate limit exceeded")
except APIError as e:
    print(f"API error: {e}")

---

🔄 Examples

Provider-Specific Usage

from weycop import Message

# Explicit provider specification
completion = client.chat_completions_create(
    model="gpt-4o",
    provider="openai",  # Optional - auto-detected if omitted
    messages=[
        Message("system", "You are a helpful assistant"),
        Message("user", "Explain quantum computing")
    ],
    max_tokens=200
)

System Prompts

from weycop import Message

completion = client.chat_completions_create(
    model="qwen3:8b",
    messages=[
        Message("system", "You are a SQL expert. Only respond with SQL code."),
        Message("user", "Get all users created in the last 30 days")
    ]
)

Multi-turn Conversation

messages = [
    Message("system", "You are a helpful math tutor."),
    Message("user", "What is 15 + 27?"),
]

completion = client.chat_completions_create(model="qwen3:4b-instruct", messages=messages)
assistant_response = completion.choices[0].message.content
messages.append(Message("assistant", assistant_response))

messages.append(Message("user", "Now multiply that by 3"))
completion = client.chat_completions_create(model="qwen3:4b-instruct", messages=messages)
print(completion.choices[0].message.content)

Cost Optimization

# Free local models for development/testing
dev_response = client.chat_local(
    model="qwen3:4b-instruct", 
    messages=[Message("user", "Test message")]
)

# Cost-effective cloud models for production
prod_response = client.chat_openai(
    model="gpt-4o-mini",  # Most affordable OpenAI model
    messages=[Message("user", "Production query")]
)

# Premium models for complex tasks
complex_response = client.chat_anthropic(
    model="claude-3-5-sonnet-20241022",
    messages=[Message("user", "Complex analysis task")]
)

Context Management

with WeycopClient(api_key="your-key") as client:
    response = client.chat("Hello!")
    print(response)
# Client is automatically closed

---

🚀 Migration from v1.x

If you're upgrading from v1.x, your existing code continues to work:

# v1.x code (still works)
response = client.chat_completions_create(
    model="qwen3:4b-instruct",
    messages=[{"role": "user", "content": "Hello"}]
)

# v2.x additions (new capabilities)
response = client.chat_local(  # Provider-specific methods
    model="qwen3:4b-instruct",
    messages=[{"role": "user", "content": "Hello"}]
)

---

---

Support

• Email: [e-mail]

---

📜 License

MIT License - see LICENSE file for details.