Skip to content

Tools API

Utilities and helpers for EvoToolkit, including LLM API clients.

HttpsApi

See the dedicated page: HttpsApi.

Purpose:

  • Connect to LLM chat APIs
  • Send prompts and receive responses
  • Handle authentication and retries automatically

Usage Example:

from evotoolkit.tools import HttpsApi
import os

# Create LLM API client
llm_api = HttpsApi(
    api_url="https://api.openai.com/v1/chat/completions",
    key=os.environ["OPENAI_API_KEY"],
    model="gpt-4o",
    temperature=1.0
)

# Send a prompt
messages = [{"role": "user", "content": "Write a Python function to compute fibonacci"}]
response, usage = llm_api.get_response(messages)

print(response)
print(f"Tokens used: {usage['total_tokens']}")

Using with evotoolkit.solve():

import evotoolkit
from evotoolkit.task.python_task import ScientificRegressionTask, EvoEngineerPythonInterface
from evotoolkit.tools import HttpsApi

# Configure LLM
llm_api = HttpsApi(
    api_url="https://api.openai.com/v1/chat/completions",
    key=os.environ["OPENAI_API_KEY"],
    model="gpt-4o"
)

# Create task and interface
task = ScientificRegressionTask(dataset_name="bactgrow")
interface = EvoEngineerPythonInterface(task)

# Solve with the LLM
result = evotoolkit.solve(
    interface=interface,
    output_path='./results',
    running_llm=llm_api,
    max_generations=5
)

Constructor & Examples

See the HttpsApi page for the full constructor, parameters, and usage examples.

Methods

Environment Variables

OPENAI_API_KEY

Store your OpenAI API key:

export OPENAI_API_KEY="sk-..."

Then use in Python:

import os
from evotoolkit.tools import HttpsApi

llm_api = HttpsApi(
    api_url="https://api.openai.com/v1/chat/completions",
    key=os.environ["OPENAI_API_KEY"],
    model="gpt-4o"
)

LLM_API_URL / LLM_API_KEY

For custom configurations:

export LLM_API_URL="https://your-api.com/v1/chat/completions"
export LLM_API_KEY="your-key"
export LLM_MODEL="gpt-4o"

Best Practices

Do's ✅

  • Store API keys in environment variables (never hardcode)
  • Use appropriate timeout values for your use case
  • Handle rate limits gracefully
  • Monitor token usage to control costs
  • Use lower temperatures (0.0-0.5) for deterministic outputs
  • Use higher temperatures (0.7-1.5) for creative outputs

Don'ts ❌

  • Don't commit API keys to git
  • Don't use excessively low timeouts (< 30s)
  • Don't ignore token usage metrics
  • Don't disable retries for production code

Provider-Specific Notes

OpenAI

llm_api = HttpsApi(
    api_url="https://api.openai.com/v1/chat/completions",
    key=os.environ["OPENAI_API_KEY"],
    model="gpt-4o"
)
  • Models: gpt-4o, gpt-4o-mini, gpt-3.5-turbo
  • Rate limits: Check your account tier
  • Docs: https://platform.openai.com/docs/api-reference

Anthropic Claude

Requires a compatible proxy or API gateway:

llm_api = HttpsApi(
    api_url="https://your-gateway.com/v1/chat/completions",
    key=os.environ["ANTHROPIC_API_KEY"],
    model="claude-3-5-sonnet-20241022"
)
  • Models: claude-3-5-sonnet-20241022, claude-3-opus-20240229
  • Note: Requires OpenAI-compatible API format
  • Docs: https://docs.anthropic.com/

Custom Providers

Many LLM providers offer OpenAI-compatible APIs:

llm_api = HttpsApi(
    api_url="api.custom-provider.com",  # Hostname only
    key="your-key",
    model="provider-model-name"
)

Check your provider's documentation for: - API endpoint URL - Authentication method - Supported models - Request/response format

Troubleshooting

Connection Errors

Problem: RuntimeError: Model Response Error!

Solutions: - Check your API URL is correct - Verify your API key is valid - Ensure network connectivity - Check provider status page

Timeout Errors

Problem: Requests timing out

Solutions: - Increase timeout parameter - Check network latency - Try a smaller model - Reduce prompt complexity

Rate Limiting

Problem: Too many requests

Solutions: - Add delays between requests - Reduce parallelism - Upgrade API tier - Implement exponential backoff

Next Steps