Deploying & Hosting A Machine Learning Model with FastAPI and Heroku

Assume that you're a data scientist. Following a typical machine learning workflow, you'll define the problem statement along with objectives based on business needs. You'll then start finding and cleaning data followed by analyzing the collected data and building and training your model. Once trained, you'll evaluate the results. This process of finding and cleansing data, training the model, and evaluating the results will continue until you're satisfied with the results. You'll then refactor the code and package it up in a module, along with its dependencies, in preparation for testing and deployment.

What happens next? Do you hand the model off to another team to test and deploy the model? Or do you have to handle this yourself? Either way, it's important to understand what happens when a model gets deployed. You may have to deploy the model yourself one day. Or maybe you have a side project that you'd just like to stand up in production and make available to end users.

In this tutorial, we'll look at how to deploy a machine learning model, for predicting stock prices, into production on Heroku as a RESTful API using FastAPI.

Objectives

By the end of this post you should be able to:

Develop a RESTful API with Python and FastAPI
Build a basic machine learning model to predict stock prices
Deploy a FastAPI app to Heroku
Use the Heroku Container Registry for deploying Docker to Heroku

FastAPI

FastAPI is a modern, high-performance, batteries-included Python web framework that's perfect for building RESTful APIs. It can handle both synchronous and asynchronous requests and has built-in support for data validation, JSON serialization, authentication and authorization, and OpenAPI.

Highlights:

Heavily inspired by Flask, it has a lightweight microframework feel with support for Flask-like route decorators.
It takes advantage of Python type hints for parameter declaration which enables data validation (via pydantic) and OpenAPI/Swagger documentation.
Built on top of Starlette, it supports the development of asynchronous APIs.
It's fast. Since async is much more efficient than the traditional synchronous threading model, it can compete with Node and Go with regards to performance.

Review the Features guide from the official docs for more info. It's also encouraged to review Alternatives, Inspiration, and Comparisons, which details how FastAPI compares to other web frameworks and technologies, for context.

Project Setup

Create a project folder called "fastapi-ml":

$ mkdir fastapi-ml
$ cd fastapi-ml

Then, create and activate a new virtual environment:

$ python3.8 -m venv env
$ source env/bin/activate
(env)$

Add a two new files: requirements.txt and main.py.

Unlike Django or Flask, FastAPI does not have a built-in development server. So, we'll use Uvicorn, an ASGI server, to serve up FastAPI.

New to ASGI? Read through the excellent Introduction to ASGI: Emergence of an Async Python Web Ecosystem blog post.

Add FastAPI and Uvicorn to the requirements file:

fastapi==0.68.0
uvicorn==0.14.0

Install the dependencies:

(env)$ pip install -r requirements.txt

Then, within main.py, create a new instance of FastAPI and set up a quick test route:

from fastapi import FastAPI


app = FastAPI()


@app.get("/ping")
def pong():
    return {"ping": "pong!"}

Start the app:

(env)$ uvicorn main:app --reload --workers 1 --host 0.0.0.0 --port 8008

So, we defined the following settings for Uvicorn:

--reload enables auto-reload so the server will restart after changes are made to the code base.
--workers 1 provides a single worker process.
--host 0.0.0.0 defines the address to host the server on.
--port 8008 defines the port to host the server on.

main:app tells Uvicorn where it can find the FastAPI ASGI application -- e.g., "within the the 'main.py' file, you'll find the ASGI app, app = FastAPI().

Navigate to http://localhost:8008/ping. You should see:

{
  "ping": "pong!"
}

ML Model

The model that we'll deploy uses Prophet to predict stock market prices.

Add the following functions to train the model and generate a prediction to a new file called model.py:

import datetime
from pathlib import Path

import joblib
import pandas as pd
import yfinance as yf
from fbprophet import Prophet

BASE_DIR = Path(__file__).resolve(strict=True).parent
TODAY = datetime.date.today()


def train(ticker="MSFT"):
    # data = yf.download("^GSPC", "2008-01-01", TODAY.strftime("%Y-%m-%d"))
    data = yf.download(ticker, "2020-01-01", TODAY.strftime("%Y-%m-%d"))
    data.head()
    data["Adj Close"].plot(title=f"{ticker} Stock Adjusted Closing Price")

    df_forecast = data.copy()
    df_forecast.reset_index(inplace=True)
    df_forecast["ds"] = df_forecast["Date"]
    df_forecast["y"] = df_forecast["Adj Close"]
    df_forecast = df_forecast[["ds", "y"]]
    df_forecast

    model = Prophet()
    model.fit(df_forecast)

    joblib.dump(model, Path(BASE_DIR).joinpath(f"{ticker}.joblib"))


def predict(ticker="MSFT", days=7):
    model_file = Path(BASE_DIR).joinpath(f"{ticker}.joblib")
    if not model_file.exists():
        return False

    model = joblib.load(model_file)

    future = TODAY + datetime.timedelta(days=days)

    dates = pd.date_range(start="2020-01-01", end=future.strftime("%m/%d/%Y"),)
    df = pd.DataFrame({"ds": dates})

    forecast = model.predict(df)

    model.plot(forecast).savefig(f"{ticker}_plot.png")
    model.plot_components(forecast).savefig(f"{ticker}_plot_components.png")

    return forecast.tail(days).to_dict("records")


def convert(prediction_list):
    output = {}
    for data in prediction_list:
        date = data["ds"].strftime("%m/%d/%Y")
        output[date] = data["trend"]
    return output

Here, we defined three functions:

train downloads historical stock data with yfinance, creates a new Prophet model, fits the model to the stock data, and then serializes and saves the model as a Joblib file.
predict loads and deserializes the saved model, generates a new forecast, creates images of the forecast plot and forecast components, and returns the days included in the forecast as a list of dicts.
convert takes the list of dicts from predict and outputs a dict of dates and forecasted values (i.e., {"07/02/2020": 200}).

This model was developed by Andrew Clark.

Update the requirements file:

# pystan must be installed before prophet
# you may need to pip install it on it's own
# before installing the remaining requirements
# pip install pystan==2.19.1.1

pystan==2.19.1.1

fastapi==0.68.0
uvicorn==0.14.0

fbprophet==0.7.1
joblib==1.0.1
pandas==1.3.1
plotly==5.1.0
yfinance==0.1.63

Install the new dependencies:

(env)$ pip install -r requirements.txt

If you have problems installing the dependencies on your machine, you may want to use Docker instead. For instructions on how to run the application with Docker, review the README on the fastapi-ml repo on GitHub.

To test, open a new Python shell and run the following commands:

(env)$ python

>>> from model import train, predict, convert
>>> train()
>>> prediction_list = predict()
>>> convert(prediction_list)

You should see something similar to:

{
    '08/12/2021': 282.99012951691776,
    '08/13/2021': 283.31354121099446,
    '08/14/2021': 283.63695290507127,
    '08/15/2021': 283.960364599148,
    '08/16/2021': 284.2837762932248,
    '08/17/2021': 284.6071879873016,
    '08/18/2021': 284.93059968137834
}

This is the predicted prices for the next seven days for Microsoft Corporation (MSFT). Take note of the saved MSFT.joblib model along with the two images:

plot

components

Go ahead and train a few more models to work with. For example:

>>> train("GOOG")
>>> train("AAPL")
>>> train("^GSPC")

Exit the shell.

With that, let's wire up our API.

Routes

Add a /predict endpoint by updating main.py like so:

from fastapi import FastAPI, HTTPException
from pydantic import BaseModel

from model import convert, predict

app = FastAPI()


# pydantic models


class StockIn(BaseModel):
    ticker: str


class StockOut(StockIn):
    forecast: dict


# routes


@app.get("/ping")
async def pong():
    return {"ping": "pong!"}


@app.post("/predict", response_model=StockOut, status_code=200)
def get_prediction(payload: StockIn):
    ticker = payload.ticker

    prediction_list = predict(ticker)

    if not prediction_list:
        raise HTTPException(status_code=400, detail="Model not found.")

    response_object = {"ticker": ticker, "forecast": convert(prediction_list)}
    return response_object

So, in the new get_prediction view function, we passed in a ticker to our model's predict function and then used the convert function to create the output for the response object. We also took advantage of a pydantic schema to covert the JSON payload to a StockIn object schema. This provides automatic type validation. The response object uses the StockOut schema object to convert the Python dict -- {"ticker": ticker, "forecast": convert(prediction_list)} -- to JSON, which, again, is validated.

For the web app, let's just output the forecast in JSON. Comment out the following lines in predict:

# model.plot(forecast).savefig(f"{ticker}_plot.png")
# model.plot_components(forecast).savefig(f"{ticker}_plot_components.png")

Full function:

def predict(ticker="MSFT", days=7):
    model_file = Path(BASE_DIR).joinpath(f"{ticker}.joblib")
    if not model_file.exists():
        return False

    model = joblib.load(model_file)

    future = TODAY + datetime.timedelta(days=days)

    dates = pd.date_range(start="2020-01-01", end=future.strftime("%m/%d/%Y"),)
    df = pd.DataFrame({"ds": dates})

    forecast = model.predict(df)

    # model.plot(forecast).savefig(f"{ticker}_plot.png")
    # model.plot_components(forecast).savefig(f"{ticker}_plot_components.png")

    return forecast.tail(days).to_dict("records")

Run the app:

(env)$ uvicorn main:app --reload --workers 1 --host 0.0.0.0 --port 8008

Then, in a new terminal window, use curl to test the endpoint:

$ curl \
  --header "Content-Type: application/json" \
  --request POST \
  --data '{"ticker":"MSFT"}' \
  http://localhost:8008/predict

You should see something like:

{
  "ticker":"MSFT",
  "forecast":{
    "08/12/2021": 282.99012951691776,
    "08/13/2021": 283.31354121099446,
    "08/14/2021": 283.63695290507127,
    "08/15/2021": 283.960364599148,
    "08/16/2021": 284.2837762932248,
    "08/17/2021": 284.6071879873016,
    "08/18/2021": 284.93059968137834
  }
}

What happens if the ticker model doesn't exist?

$ curl \
  --header "Content-Type: application/json" \
  --request POST \
  --data '{"ticker":"NONE"}' \
  http://localhost:8008/predict

{
  "detail": "Model not found."
}

Heroku Deployment

Heroku is a Platform as a Service (PaaS) that provides hosting for web applications. They offer abstracted environments where you don't have to manage the underlying infrastructure, making it easy to manage, deploy, and scale web applications. With just a few clicks you can have your app up and running, ready to receive traffic.

Sign up for a Heroku account (if you don’t already have one), and then install the Heroku CLI (if you haven't already done so).

Next, log in to your Heroku account via the CLI:

$ heroku login

You'll be prompted to press any key to open your web browser to complete login.

Create a new app on Heroku:

$ heroku create

You should see something similar to:

Creating app... done, ⬢ tranquil-cliffs-74287
https://tranquil-cliffs-74287.herokuapp.com/ | https://git.heroku.com/tranquil-cliffs-74287.git

Next, we'll use Heroku's Container Registry to deploy the application with Docker. Put simply, with the Container Registry, you can deploy pre-built Docker images to Heroku.

Why Docker? We want to minimize the differences between the production and development environments. This is especially important with this project, since it relies on a number of data science dependencies that have very specific system requirements.

Log in to the Heroku Container Registry, to indicate to Heroku that we want to use the Container Runtime:

$ heroku container:login

Add a Dockerfile file to the project root:

FROM python:3.8

WORKDIR /app

RUN apt-get -y update  && apt-get install -y \
  python3-dev \
  apt-utils \
  python-dev \
  build-essential \
&& rm -rf /var/lib/apt/lists/*

RUN pip install --upgrade setuptools
RUN pip install \
    cython==0.29.24 \
    numpy==1.21.1 \
    pandas==1.3.1 \
    pystan==2.19.1.1

COPY requirements.txt .
RUN pip install -r requirements.txt

COPY . .

CMD gunicorn -w 3 -k uvicorn.workers.UvicornWorker main:app --bind 0.0.0.0:$PORT

Here, after pulling the Python 3.8 base image, we installed the appropriate dependencies, copied over the app, and ran Gunicorn, a production-grade WSGI application server, to manage Uvicorn with three worker processes. This config takes advantage of both concurrency (via Uvicorn) and parallelism (via Gunicorn workers).

Add Gunicorn to the requirements.txt file:

# pystan must be installed before prophet
# you may need to pip install it on it's own
# before installing the remaining requirements
# pip install pystan==2.19.1.1

pystan==2.19.1.1

fastapi==0.68.0
gunicorn==20.1.0
uvicorn==0.14.0

fbprophet==0.7.1
joblib==1.0.1
pandas==1.3.1
plotly==5.1.0
yfinance==0.1.63

Add a .dockerignore file as well:

__pycache__
env

Build the Docker image and tag it with the following format:

registry.heroku.com/<app>/<process-type>

Make sure to replace <app> with the name of the Heroku app that you just created and <process-type> with web since this will be for a web process.

For example:

$ docker build -t registry.heroku.com/tranquil-cliffs-74287/web .

It will take several minutes to install fbprophet. Be patient. You should see it hang here for some time:

Running setup.py install for fbprophet: started

Once done, you can run the image like so:

$ docker run --name fastapi-ml -e PORT=8008 -p 8008:8008 -d registry.heroku.com/tranquil-cliffs-74287/web:latest

Ensure http://localhost:8008/ping works as expected. Once done, stop and remove the container:

$ docker stop fastapi-ml
$ docker rm fastapi-ml

Push the image to the registry:

$ docker push registry.heroku.com/tranquil-cliffs-74287/web

Release the image:

$ heroku container:release -a tranquil-cliffs-74287 web

This will run the container. You should now be able to view your app. Make sure to test the /predict endpoint:

$ curl \
  --header "Content-Type: application/json" \
  --request POST \
  --data '{"ticker":"MSFT"}' \
  https://<YOUR_HEROKU_APP_NAME>.herokuapp.com/predict

Finally, check out the interactive API documentation that FastAPI automatically generates at https://<YOUR_HEROKU_APP_NAME>.herokuapp.com/docs:

swagger ui

Conclusion

This tutorial looked at how to deploy a machine learning model, for predicting stock prices, into production on Heroku as a RESTful API using FastAPI.

What's next?

Set up a database to save prediction results
Create a production Dockerfile that uses multistage Docker builds to reduce the size of the production image
Add logging and monitoring
Convert your view functions and the model prediction function into asynchronous functions
Run the prediction as a background task to prevent blocking
Add tests
Store trained models to AWS S3, outside of Heroku's ephemeral filesystem

Check out the following resources for help with the above pieces:

If you're deploying a non-trivial model, I recommend adding model versioning and support for counterfactual analysis along with model monitoring (model and feature drift, bias detection). Check out the Monitaur platform for help in these areas.

You can find the code in the fastapi-ml repo.

Original article source at: https://testdriven.io/

#fastapi #heroku #machinelearning

What is GEEK

Buddha Community

Deploying & Hosting A Machine Learning Model with FastAPI and Heroku

Michael Hamill

1617331277

Workshop Alert! Deep Learning Model Deployment & Management

The Association of Data Scientists (AdaSci), the premier global professional body of data science and ML practitioners, has announced a hands-on workshop on deep learning model deployment on February 6, Saturday.

Over the last few years, the applications of deep learning models have increased exponentially, with use cases ranging from automated driving, fraud detection, healthcare, voice assistants, machine translation and text generation.

Typically, when data scientists start machine learning model development, they mostly focus on the algorithms to use, feature engineering process, and hyperparameters to make the model more accurate. However, model deployment is the most critical step in the machine learning pipeline. As a matter of fact, models can only be beneficial to a business if deployed and managed correctly. Model deployment or management is probably the most under discussed topic.

In this workshop, the attendees get to learn about ML lifecycle, from gathering data to the deployment of models. Researchers and data scientists can build a pipeline to log and deploy machine learning models. Alongside, they will be able to learn about the challenges associated with machine learning models in production and handling different toolkits to track and monitor these models once deployed.

#hands on deep learning #machine learning model deployment #machine learning models #model deployment #model deployment workshop

sophia tondon

1620898103

5 Latest Technology Trends of Machine Learning for 2021

Check out the 5 latest technologies of machine learning trends to boost business growth in 2021 by considering the best version of digital development tools. It is the right time to accelerate user experience by bringing advancement in their lifestyle.

#machinelearningapps #machinelearningdevelopers #machinelearningexpert #machinelearningexperts #expertmachinelearningservices #topmachinelearningcompanies #machinelearningdevelopmentcompany

Visit Blog- https://www.xplace.com/article/8743

#machine learning companies #top machine learning companies #machine learning development company #expert machine learning services #machine learning experts #machine learning expert

Nora Joy

1604154094

Hire Machine Learning Developers in India

Hire machine learning developers in India ,DxMinds Technologies is the best product engineering company in India making innovative solutions using Machine learning and deep learning. We are among the best to hire machine learning experts in India work in different industry domains like Healthcare retail, banking and finance ,oil and gas, ecommerce, telecommunication ,FMCG, fashion etc.
**
Services**
Product Engineering & Development
Re-engineering
Maintenance / Support / Sustenance
Integration / Data Management
QA & Automation
Reach us 917483546629

Hire machine learning developers in India ,DxMinds Technologies is the best product engineering company in India making innovative solutions using Machine learning and deep learning. We are among the best to hire machine learning experts in India work in different industry domains like Healthcare retail, banking and finance ,oil and gas, ecommerce, telecommunication ,FMCG, fashion etc.

Services

Product Engineering & Development

Re-engineering

Maintenance / Support / Sustenance

Integration / Data Management

QA & Automation

Reach us 917483546629

#hire machine learning developers in india #hire dedicated machine learning developers in india #hire machine learning programmers in india #hire machine learning programmers #hire dedicated machine learning developers #hire machine learning developers

Nora Joy

1607006620

Applications of machine learning in different industry domains

Machine learning applications are a staple of modern business in this digital age as they allow them to perform tasks on a scale and scope previously impossible to accomplish.Businesses from different domains realize the importance of incorporating machine learning in business processes.Today this trending technology transforming almost every single industry ,business from different industry domains hire dedicated machine learning developers for skyrocket the business growth.Following are the applications of machine learning in different industry domains.

Transportation industry

Machine learning is one of the technologies that have already begun their promising marks in the transportation industry.Autonomous Vehicles,Smartphone Apps,Traffic Management Solutions,Law Enforcement,Passenger Transportation etc are the applications of AI and ML in the transportation industry.Following challenges in the transportation industry can be solved by machine learning and Artificial Intelligence.

ML and AI can offer high security in the transportation industry.
It offers high reliability of their services or vehicles.
The adoption of this technology in the transportation industry can increase the efficiency of the service.
In the transportation industry ML helps scientists and engineers come up with far more environmentally sustainable methods for powering and operating vehicles and machinery for travel and transport.

Healthcare industry

Technology-enabled smart healthcare is the latest trend in the healthcare industry. Different areas of healthcare, such as patient care, medical records, billing, alternative models of staffing, IP capitalization, smart healthcare, and administrative and supply cost reduction. Hire dedicated machine learning developers for any of the following applications.

Identifying Diseases and Diagnosis
Drug Discovery and Manufacturing
Medical Imaging Diagnosis
Personalized Medicine
Machine Learning-based Behavioral Modification
Smart Health Records
Clinical Trial and Research
Better Radiotherapy
Crowdsourced Data Collection
Outbreak Prediction

**
Finance industry**

In financial industries organizations like banks, fintech, regulators and insurance are Adopting machine learning to improve their facilities.Following are the use cases of machine learning in finance.

Fraud prevention
Risk management
Investment predictions
Customer service
Digital assistants
Marketing
Network security
Loan underwriting
Algorithmic trading
Process automation
Document interpretation
Content creation
Trade settlements
Money-laundering prevention
Custom machine learning solutions

Education industry

Education industry is one of the industries which is investing in machine learning as it offers more efficient and easierlearning.AdaptiveLearning,IncreasingEfficiency,Learning Analytics,Predictive Analytics,Personalized Learning,Evaluating Assessments etc are the applications of machine learning in the education industry.

Outsource your machine learning solution to India,India is the best outsourcing destination offering best in class high performing tasks at an affordable price.Business** hire dedicated machine learning developers in India for making your machine learning app idea into reality.
**
Future of machine learning
Continuous technological advances are bound to hit the field of machine learning, which will shape the future of machine learning as an intensively evolving language.

Improved Unsupervised Algorithms
Increased Adoption of Quantum Computing
Enhanced Personalization
Improved Cognitive Services
Rise of Robots

**Conclusion
**
Today most of the business from different industries are hire machine learning developers in India and achieve their business goals. This technology may have multiple applications, and, interestingly, it hasn’t even started yet but having taken such a massive leap, it also opens up so many possibilities in the existing business models in such a short period of time. There is no question that the increase of machine learning also brings the demand for mobile apps, so most companies and agencies employ Android developers and hire iOS developers to incorporate machine learning features into them.

#hire machine learning developers in india #hire dedicated machine learning developers in india #hire machine learning programmers in india #hire machine learning programmers #hire dedicated machine learning developers #hire machine learning developers

Nora Joy

1607006620

Hire Machine Learning Developer | Hire ML Experts in India

Machine learning applications are a staple of modern business in this digital age as they allow them to perform tasks on a scale and scope previously impossible to accomplish.Businesses from different domains realize the importance of incorporating machine learning in business processes.Today this trending technology transforming almost every single industry ,business from different industry domains hire dedicated machine learning developers for skyrocket the business growth.Following are the applications of machine learning in different industry domains.

Transportation industry

Machine learning is one of the technologies that have already begun their promising marks in the transportation industry.Autonomous Vehicles,Smartphone Apps,Traffic Management Solutions,Law Enforcement,Passenger Transportation etc are the applications of AI and ML in the transportation industry.Following challenges in the transportation industry can be solved by machine learning and Artificial Intelligence.

ML and AI can offer high security in the transportation industry.
It offers high reliability of their services or vehicles.
The adoption of this technology in the transportation industry can increase the efficiency of the service.
In the transportation industry ML helps scientists and engineers come up with far more environmentally sustainable methods for powering and operating vehicles and machinery for travel and transport.

Healthcare industry

Technology-enabled smart healthcare is the latest trend in the healthcare industry. Different areas of healthcare, such as patient care, medical records, billing, alternative models of staffing, IP capitalization, smart healthcare, and administrative and supply cost reduction. Hire dedicated machine learning developers for any of the following applications.

Identifying Diseases and Diagnosis
Drug Discovery and Manufacturing
Medical Imaging Diagnosis
Personalized Medicine
Machine Learning-based Behavioral Modification
Smart Health Records
Clinical Trial and Research
Better Radiotherapy
Crowdsourced Data Collection
Outbreak Prediction

**
Finance industry**

In financial industries organizations like banks, fintech, regulators and insurance are Adopting machine learning to improve their facilities.Following are the use cases of machine learning in finance.

Fraud prevention
Risk management
Investment predictions
Customer service
Digital assistants
Marketing
Network security
Loan underwriting
Algorithmic trading
Process automation
Document interpretation
Content creation
Trade settlements
Money-laundering prevention
Custom machine learning solutions

Education industry

Education industry is one of the industries which is investing in machine learning as it offers more efficient and easierlearning.AdaptiveLearning,IncreasingEfficiency,Learning Analytics,Predictive Analytics,Personalized Learning,Evaluating Assessments etc are the applications of machine learning in the education industry.

Outsource your machine learning solution to India,India is the best outsourcing destination offering best in class high performing tasks at an affordable price.Business** hire dedicated machine learning developers in India for making your machine learning app idea into reality.
**
Future of machine learning
Continuous technological advances are bound to hit the field of machine learning, which will shape the future of machine learning as an intensively evolving language.

Improved Unsupervised Algorithms
Increased Adoption of Quantum Computing
Enhanced Personalization
Improved Cognitive Services
Rise of Robots

**Conclusion
**
Today most of the business from different industries are hire machine learning developers in India and achieve their business goals. This technology may have multiple applications, and, interestingly, it hasn’t even started yet but having taken such a massive leap, it also opens up so many possibilities in the existing business models in such a short period of time. There is no question that the increase of machine learning also brings the demand for mobile apps, so most companies and agencies employ Android developers and hire iOS developers to incorporate machine learning features into them.

#hire machine learning developers in india #hire dedicated machine learning developers in india #hire machine learning programmers in india #hire machine learning programmers #hire dedicated machine learning developers #hire machine learning developers