https://github.com/spyker77/fastapi-tdd-docker
Transformers with test-driven development
https://github.com/spyker77/fastapi-tdd-docker
asyncio celery docker fastapi heroku mysql oauth2 pytest python rabbitmq redis sqlalchemy transformers
Last synced: 12 months ago
JSON representation
Transformers with test-driven development
- Host: GitHub
- URL: https://github.com/spyker77/fastapi-tdd-docker
- Owner: spyker77
- License: mit
- Created: 2021-01-07T15:38:57.000Z (over 5 years ago)
- Default Branch: main
- Last Pushed: 2023-11-03T15:22:17.000Z (over 2 years ago)
- Last Synced: 2025-06-20T22:00:19.435Z (about 1 year ago)
- Topics: asyncio, celery, docker, fastapi, heroku, mysql, oauth2, pytest, python, rabbitmq, redis, sqlalchemy, transformers
- Language: Python
- Homepage:
- Size: 598 KB
- Stars: 8
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Test-Driven Development with FastAPI and Docker
## Table of Contents
- [Introduction](#introduction)
- [Prerequisites](#prerequisites)
- [Installation](#installation)
- [Usage](#usage)
- [Testing](#testing)
- [Deployment](#deployment)
- [License](#license)
## Introduction
This project is an implementation of [the course](https://testdriven.io/courses/tdd-fastapi/) on Test-Driven Development to build the API for summarizing articles, with additional features and optimizations. Key enhancements include:
- Dependencies updated to the latest version at the moment
- CORSMiddleware used to manually control allowed origins
- Venv replaced with PDM
- Added versions of the API
- Refactored the code
- Gunicorn added to manage the uvicorn
- BackgroundTasks replaced with Celery, Redis and RabbitMQ
- Optimized CI/CD pipeline in GitHub Actions
- Migrated to the Container registry from the Docker registry
- Implemented authentication and authorization using OAuth2
- Tortoise-ORM has been replaced by SQLAlchemy
- Transformers models are used instead of the NLP from Newspaper3k
## Prerequisites
- Docker and Docker Compose installed on your machine.
- Basic knowledge of Python, FastAPI, and Docker.
## Installation
1. Clone the repository:
```bash
git clone https://github.com/spyker77/fastapi-tdd-docker.git
cd fastapi-tdd-docker
```
2. Build and start the containers:
```bash
docker compose up -d --build
```
3. Generate the database schema:
```bash
docker compose exec web alembic upgrade head
```
## Usage
1. Access the API documentation at:
2. Create a test user at:
Example of the payload:
```json
{
"full_name": "Cute Koala",
"username": "cute",
"email": "cute@example.com",
"password": "supersecret"
}
```
3. Use the **Authorize** button (simplest way) at the top and enter the username and password you've just created, then click **Authorize**.
4. At this point you're authorized. Now use the endpoint to send the article you want to summarize, for example like so:
```json
{
"url": "https://dev.to/spyker77/how-to-connect-godaddy-domain-with-heroku-and-cloudflare-mdh"
}
```
5. This triggers the ML models to download, which may take a few minutes for the first run (in the current implementation). After that, reach the endpoint and in the response you should see something like:
```json
[
{
"id": 1,
"url": "https://dev.to/spyker77/how-to-connect-godaddy-domain-with-heroku-and-cloudflare-mdh",
"summary": "If you struggle to connect newly registered domain from GoDaddy with your app at Heroku, and in addition would like to use advantages of Cloudflare – this article is for you. Hope it will help you and without many words, let's jump in!Sections: Heroku settings, Cloudflare settings and GoDaddy settings.",
"user_id": 1
}
]
```
## Testing
Run the tests using the following command:
```bash
docker compose exec web pytest -n auto --cov
```
## Deployment
For production deployment, don't forget to change the ENVIRONMENT variables. The default CI/CD pipeline is set up to build images with GitHub Actions, store them in GitHub Packages, and deploy the application to Heroku. But the deployment part is currently disabled/commented out.
### Note
The current implementation of text summarization using the transformer models is not ideal for production due to the following reasons:
- The requirement to install a heavy transformers library along with its dependencies.
- The necessity to download several gigabytes of the model.
- The need for powerful hardware to run the model.
Typically, in a production environment, the models would be provided via an API using services like AWS SageMaker or Paperspace Gradient.
## License
This project is licensed under the MIT License. See the [MIT License](LICENSE) file for details.