Summarizer

Description

Small web app exposing a Machine Learning model from HuggingFace. The model creates summaries of text.

Requirements

This web app server was built with python 3.6 and depends on the following packages:

flask==1.1.2
flask-restful==0.3.8
transformers==4.5.0
torch==1.8.1+cpu
torchvision==0.9.1+cpu
torchaudio==0.8.1
sentencepiece==0.1.95
protobuf==3.15.8
pytest==6.2.3
flake8==3.9.0
flake8-docstrings==1.6.0
pep8-naming==0.11.1
mypy==0.812

In order to install all the dependencies execute the following command in your terminal:

$ pip install -r requirements.txt --find-links https://download.pytorch.org/whl/torch_stable.html

It's recommended that you use a virtual environment, ie. virtualenv, venv, etc.

What to expect

This web app is only intended for demo purposes. It uses Flask for the web app and transformers for the Machine Learning model.

The model is a summarization transformer from HuggingFace, and its exposed at POST http://127.0.0.1:5000/api/predict.

How to use it

First clone the project to your local computer:

$ git clone https://github.com/jsgilberto/summarizer.git

To start the web app, execute the following command:

$ python3 src/app.py

By default, the app is going to run on port 5000.

To start making predictions, make a POST request to the following endpoint:

$ curl --location --request POST 'http://127.0.0.1:5000/api/predict' \
--header 'Content-Type: application/json' \
--data-raw '{
    "text": "As they rounded a bend in the path that ran beside the river..."
}'

Make sure the text is long enough for the model so it is be able to make a summary. If its not long enough, it will return as a response the same text you initially passed.

Type consistency

I used mypy for type consistency with the configuration provided in mypy.ini

If you want to check type consistency, execute the following command:

$ mypy src

Code validation (PEP-8)

flake8 is the tool used in this module. For validating the code, the following command is used:

$ flake8

It uses the .flake8 configuration file placed in the root of the project.

Tests

For testing pytest was the choice. The tests are located in the tests folder placed in the root of this project.

$ python3 -m pytest -W ignore::DeprecationWarning

or:

$ pytest -W ignore::DeprecationWarning

The tests are only an example of what it would look like in a real world project.

Deployment

For this project I decided to use a WSGI application server: Gunicorn. This helps us to keep multiple processes of the flask application running.

The following command was used to run gunicorn:

gunicorn -w 4 -b 127.0.0.1:8000 --chdir $(pwd)/src wsgi:app

Suggestions

Please feel free to make any suggestions on the project.

jsgilberto / summarizer