fastmlapi 0.1.0

A FastAPI-based ML worker node library for easy model deployment

FastMLAPI

A FastAPI-based ML worker node library for easy model deployment. Create production-ready ML APIs with minimal boilerplate.

Features

• 🚀 Simple API: Extend MLController and implement a few methods
• 🔄 Automatic /predict endpoint: Generated automatically with proper request/response handling
• 🎯 Preprocessing/Postprocessing: Decorators for clean data pipeline
• 📊 Health checks: Built-in /health endpoint
• 📝 Auto documentation: Swagger/OpenAPI docs out of the box
• 🔧 Customizable: Custom request/response models supported

Installation

pip install fastmlapi

Or install from source:

pip install -e .

Quick Start

from fastmlapi import MLController, preprocessing, postprocessing
import numpy as np

class MyClassifier(MLController):
    model_name = "my-classifier"
    model_version = "1.0.0"
    
    def load_model(self):
        # Load your model here (sklearn, pytorch, tensorflow, etc.)
        return lambda x: np.array([1 if sum(x[0]) > 0 else 0])
    
    @preprocessing
    def preprocess(self, data: dict) -> np.ndarray:
        """Transform input data for the model."""
        features = data.get("features", [])
        return np.array(features).reshape(1, -1)
    
    @postprocessing  
    def postprocess(self, prediction: np.ndarray) -> dict:
        """Format model output for API response."""
        return {
            "class": int(prediction[0]),
            "label": "positive" if prediction[0] == 1 else "negative"
        }

# Run the server
if __name__ == "__main__":
    MyClassifier().run()

Or with uvicorn:

# main.py
classifier = MyClassifier()
app = classifier.app

uvicorn main:app --reload

API Endpoints

Once running, your API will have:

• POST /predict - Run predictions
• GET /health - Health check
• GET / - API info
• GET /docs - Swagger documentation

Example Request

curl -X POST http://localhost:8000/predict \
  -H "Content-Type: application/json" \
  -d '{"data": {"features": [1.0, 2.0, 3.0]}}'

Example Response

{
  "success": true,
  "prediction": {
    "class": 1,
    "label": "positive"
  },
  "metadata": {
    "model_name": "my-classifier",
    "model_version": "1.0.0"
  }
}

Advanced Usage

Custom Request/Response Models

from pydantic import BaseModel
from typing import List

class ImageRequest(BaseModel):
    image_url: str
    threshold: float = 0.5

class ImageResponse(BaseModel):
    objects: List[dict]
    count: int

class ObjectDetector(MLController):
    model_name = "object-detector"
    request_model = ImageRequest
    response_model = ImageResponse
    
    def load_model(self):
        return load_yolo_model()
    
    @preprocessing
    def preprocess(self, data: dict):
        image = download_image(data["image_url"])
        return image
    
    @postprocessing
    def postprocess(self, detections) -> dict:
        return {
            "objects": detections,
            "count": len(detections)
        }

Custom Prediction Logic

Override predict_raw for models without a standard predict() method:

class PyTorchController(MLController):
    def load_model(self):
        model = torch.load("model.pt")
        model.eval()
        return model
    
    def predict_raw(self, preprocessed_data):
        with torch.no_grad():
            return self.model(preprocessed_data)

Configuration

MLController Options

Attribute	Type	Default	Description
model_name	str	"ml-model"	Name of the model
model_version	str	"1.0.0"	Model version
title	str	"FastMLAPI"	API title for docs
description	str	"ML Model Serving API"	API description
api_version	str	"1.0.0"	API version
request_model	BaseModel	PredictionRequest	Custom request schema
response_model	BaseModel	PredictionResponse	Custom response schema
enable_health	bool	True	Enable /health endpoint
enable_docs	bool	True	Enable Swagger/OpenAPI docs

run() Options

Parameter	Type	Default	Description
host	str	"0.0.0.0"	Host to bind to
port	int	8000	Port to bind to
reload	bool	False	Enable auto-reload

Development

# Install dev dependencies
pip install -e ".[dev]"

# Run tests
pytest

# Run example
python examples/simple_classifier.py

License

MIT License